UNIVERSIDAD POLITÉCNICA DE MADRID

DEPARTAMENTO DE SILVOPASCICULTURA

ESCUELA TÉCNICA SUPERIOR DE INGENIEROS DE MONTES

GESTIÓN DE LA FERTILIDAD DE SUELOS Y LA NUTRICIÓN

DE PLANTACIONES DE TECA (Tectona grandis L.f.) EN

AMÉRICA CENTRAL

JESÚS FERNÁNDEZ MOYA

Ingeniero de Montes

DIRECTORES

MIGUEL MARCHAMALO SACRISTÁN ALFONSO SAN MIGUEL AYANZ

Dr. Ingeniero de Montes Dr. Ingeniero de Montes

2014

Tribunal nombrado por el Mgfco. Y Excmo. Sr. Rector de la Universidad Politécnica

de Madrid, el día …….. de …………………………… de 2013

Presidente D. ………………………………………………………………………………………………………..

Vocal D. ………………………………………………………………………………………………………………..

Vocal D. ………………………………………………………………………………………………………………..

Vocal D. ………………………………………………………………………………………………………………..

Secretario D. ………………………………………………………………………………………………………..

Realizado el acto de defensa y lectura de la Tesis el día ……..

de ………………………………………. de 2014 en Madrid.

Calificación ……………………………………………………..

EL PRESIDENTE LOS VOCALES

EL SECRETARIO

MENCIÓN DE DOCTORADO INTERNACIONAL

INTERNATIONAL DOCTORATE MENTION

Esta Tesis ha sido informada positivamente para su defensa en exposición

pública por los siguientes investigadores:

This Ph.D. Thesis has been positively evaluated for its public defense by

the following external reviewers:

Dr. Lynn Carpenter

Dept. of Ecology and Evolutionary Biology

University of California (UNITED STATES OF AMERICA)

Dr. Raúl Jaramillo

Regional office for Northern Latin America

International Plant Nutrition Institute (ECUADOR)

Al campo,

a sus maravillas naturales

a la belleza de los pueblos

a toda la gente que vive en él , de él y por él

a todos los que esperamos hacerlo

Dice el árbol:

Yo soy la cama donde naciste

y la cuna donde te mecieron

La silla y la mesa donde aprendiste a comer,

a leer y a escribir

El suelo que pisas

y el techo que te cubre

Yo soy la fruta que te comes

y el oxígeno que respiras

Yo soy la sombra que refresca tu calor

y el fuego que caliente tus inviernos

Yo soy el adorno de tus calles,

el aroma de tus campos y la belleza de tus paisajes

Yo soy la alegría de tus pájaros y el freno del mal tiempo

Por último, yo te acogeré amorosamente al final de tus días

Caminante que pasas, párate un momento.

Mírame bien

¡y no me hagas daño!

Rabindranath Tagore

Versionado por Jesús Ribagorda Robles

(Chapinería, Octubre 2009)

AGRADECIMIENTOS

Quiero empezar esta Tesis agradeciendo, de manera muy especial, a “Don Alfredo”. Co-

director de ésta junto con Miguel y Alfonso, Alfredo no aparece como tal por cuestiones

burocráticas pero es, de hecho, uno de los pilares fundamentales de este trabajo. Fue él el quien

me introdujo en estos temas de la nutrición forestal “y esas varas” y me acogió con los brazos

abiertos cuando tenía 4 años de beca pero no sabíamos qué hacer con ellos. Gran parte de lo

aprendido durante estos años se lo debo a él, pero hay otros Profesores que también me han

enseñado muchas, y para mí muy importantes lecciones: Gilberto, Rafa y Floria. Los cuatro,

además de regalarme su amistad, me contagiaron el amor que ellos sienten por los suelos, que

ahora me parece inimaginable no sentir yo mismo. Como he dicho, aparte de tener toda esta

ayuda a un lado del charco, en Madrid he tenido siempre el apoyo de Miguel y de Alfonso. Les

quedo profundamente agradecido por este apoyo y las enseñanzas que me han ido aportando.

No me cabía ninguna duda de que esto iba a ser así, sabiendo cómo son los dos. En mi caso, por

mi forma de ser, hay poca gente en la que confíe tanto como lo hago en ellos dos. Siempre un

buen consejo apropiado a la ocasión y siempre dándolo con una generosidad y una preocupación

por los demás abrumadora.

Después de esto, es necesario también agradecer la participación fundamental que han tenido

otras personas, empresas e instituciones para la elaboración de la tesis, bien sea financiándola

parcialmente o realizando trabajo de campo. Los datos, que siempre dan la sensación de ser

escasos, resultan caros y duros de conseguir en campo, en situaciones que no siempre son tan

cómodas como nos gustaría o como estamos acostumbrados. En ese sentido, mi más sincero

agradecimiento a todos aquellos que han participado en las diversas campañas para conseguir

unos datos que luego he usado parcialmente para este trabajo. La AECID, la Universidad

Politécnica de Madrid y las empresas Panamerican Woods, Ecoforest, Inversiones

Agroforestales, Green Millenium y Hermanos Cabalceta han financiado parte de las muestras

tomadas, han aportado algunas que tenían registrados en sus bases de datos, han permitido que

usásemos sus fincas para establecer ensayos y han autorizado que varios de sus técnicos y

capataces nos ayudasen en las tareas de campo. En especial cabe mencionar en ese sentido a:

Folkert Kottman, Vinicio Ríos, Manuel Morales, Jean Marc Verjans, Edwin Vaides, Jose

Manuel Segura, Edwin Cabalceta, Yurien Gutiérrez, Sabas Elijio y Randall González. Además,

otros profesionales que realizaron sus tesis en temas similares han compartido sus datos: Rafael

Murillo, Edward Portuguez, Juan Luis Fallas y Helga Thiele. Adam Collins, Paul Robertson y

Richard Anderson me han ayudado mucho con la revisión del inglés al escribir.

El Centro de Investigaciones Agronómicas de la Universidad de Costa Rica (especialmente

el Laboratorio de Recursos Naturales) ha sido nuestra casa durante cuatro años. Hemos

compartido risas y lamentos durante las muchas horas que pasábamos allí, los cafés … y por

supuesto el Fito’s … Es intangible la ayuda que se recibe de los compañeros, que muchas veces

no se traduce en algo concreto más allá del estar ahí “y comentar la jugada”…aunque otras

veces se materializa en ayudas y favores varios de muy distinta índole. En ese sentido, además

de los ya comentados anteriormente, tengo que agradecer a Oldemar, Warren, Mariel, Manuel,

DAngelo, Arabela, Bryan, Keneth, Aileen, Juanjo, Juanca, Roger, Rafa Salas, Eloy, Jorge,

Mario, Carlos, Gloria, Luis Gómez, María Gabriela, Carmen, Karol, Patri y Diana. Igual que

antes, los compañeros no siempre están cerca. César, aparte de ser un grandísimo amigo, es un

compañero excelente que nos ha ayudado muchísimo con todos los trámites y papeleos que

hemos tenido que hacer en la distancia…con lo que le gustan. Además, Elena, Rubén y el

personal del Laboratorio de Topografía de Caminos han sido siempre un gran apoyo. En ese

sentido, Aída y Ramón han resultado unos compañeros magníficos en la recta final y Juan y

Sonia dos “grandes fichajes” que me han recibido con los brazos abiertos.

Por último, aunque seguro que me dejo a algunas personas inmerecidamente fuera de estas

líneas, quiero agradecer a mi familia…por todo. A Nur, además de por hacerme feliz todos los

días, por toda la ayuda que he recibido de ella en lo relativo a esta tesis, moral y materialmente.

Al resto de mi familia, a mis madres, hermanas/os y abuelos/as. Po su apoyo incondicional y su

cariño. A todos los que han venido antes de mí, me han hecho en parte ser como soy y, como

tal, se reflejan en mi persona. Encarnando el pasado (Celeya 1955), reinstalándose en el

presente (Cortázar1984). Muchas gracias a todos/as.

i

ÍNDICE

RESUMEN ..................................................................................................................................... iii

ABSTRACT ..................................................................................................................................... v

PRÓLOGO ..................................................................................................................................... vii

1. Introducción .............................................................................................................................. 1

1.1. Justificación ........................................................................................................................ 3

1.2. Objetivos ............................................................................................................................ 5

1.3. Esquema de trabajo ........................................................................................................... 6

2. Marco téorico ......................................................................................................................... 11

2.1. Bosques plantados para satisfacer demandas humanas ............................................... 133

2.1.1. Aspectos generales de los bosques plantados ........................................................ 133

2.1.2. Aspectos ambientales de los bosques plantados .................................................... 166

2.1.3. Aspectos socioeconómicos de los bosques plantados ............................................ 177

2.1.4. Plantaciones forestales como alternativa de desarrollo ........................................... 18

2.2. Fertilidad de suelos y nutrición forestal ........................................................................... 20

2.2.1. Nutrición y fertilización forestal: una perspectiva histórica ..................................... 20

2.2.2. Fertilidad del suelo en sistemas forestales ............................................................. 211

2.2.3. Concentración foliar de nutrientes ........................................................................... 26

2.2.4. Empobrecimiento de los suelos por salida de nutrientes en la madera ................... 28

2.2.5. Aplicación de enmiendas y fertilizantes .................................................................... 32

2.3. Plantaciones de teca (Tectona grandis L.f.) ..................................................................... 40

2.4. Suelo y nutrición de plantaciones de teca ....................................................................... 45

3. Soil fertility characterization of teak plantations in Central America .................................. 49

3.1. Introduction ..................................................................................................................... 51

3.2. Material and Methods...................................................................................................... 52

3.3. Results and discussion ...................................................................................................... 55

4. Nutrient concentration age dynamics of teak plantations in Central America .................... 65

4.1. Introduction ..................................................................................................................... 67

4.2. Materials and methods .................................................................................................... 68

4.5. Results .............................................................................................................................. 71

4.6. Discussion ......................................................................................................................... 75

5. Nutrient accumulation and export in teak plantations of Central America ......................... 81

5.1. Introduction ..................................................................................................................... 83

5.2. Material and methods ...................................................................................................... 85

ii

5.3. Results .............................................................................................................................. 90

5.4. Discussion ......................................................................................................................... 95

6. Modifying harvesting time as a tool to reduce nutrient export by timber extraction ...... 103

6.1. Introduction ................................................................................................................... 105

6.2. Material and methods .................................................................................................... 105

6.3. Results and discussion .................................................................................................... 106

7. Relationships between nutrient soil availability, foliar concentrationand tree growth in

teak plantations .................................................................................................................... 11919

7.1. Introduction ............................................................................................................... 12121

7.2. Material and methods .................................................................................................... 122

7.3. Results ............................................................................................................................ 125

7.4. Discussion ................................................................................................................... 13131

8. Using multivariate analysis of soil fertility as a tool for forest fertilization planning .... 13737

8.1. Introduction ............................................................................................................... 13939

8.2. Materials and methods .................................................................................................. 141

8.3. Results and discussion ................................................................................................ 14646

9. Is N-P-K fertilization of teak plantations always a good choice? .................................... 15555

9.1. Introduction ................................................................................................................... 157

9.2. Materials and methods .................................................................................................. 158

9.3. Results ............................................................................................................................ 162

9.4. Discussion ....................................................................................................................... 166

10. Discusión general ................................................................................................................ 171

10.1. Caracterización de la fertilidad del suelo y la nutrición de las plantaciones ............... 173

10.2. Implicaciones para la elección de sitio ......................................................................... 175

10.3. Implicaciones para la sostenibilidad ............................................................................ 176

10.4. Implicaciones para la interpretación de análisis foliares ............................................. 178

10.5. Implicaciones para el diseño de planes de fertilización ............................................... 179

11. Conclusiones [en español] .............................................................................................. 18383

12. Conclusions [in English] .................................................................................................. 18387

13. Bibliografía ...................................................................................................................... 19191

Anexo I

Anexo II

iii

RESUMEN

La teca (Tectona grandis L.f.) ha sido tradicionalmente considerada como una madera

preciosa en los países del SE Asiático, de donde es originaria, pero durante las últimas décadas

ha alcanzado especial relevancia en el sector internacional de las maderas tropicales duras de

buena calidad. La especie ha sido ampliamente establecida en América Central, donde tiene una

gran importancia socioeconómica, tanto por el impacto de las grandes empresas multinacionales

que gestionan grandes plantaciones en la región, como por el gran número de pequeños y

medianos propietarios que han elegido esta especie para reforestar sus tierras. Pese a la gran

importancia de esta especie, se ha desarrollado relativamente poca investigación acerca de su

nutrición y de la gestión del suelo necesaria para su establecimiento y mantenimiento en

condiciones sostenibles y productivas. En la presente Tesis Doctoral, tras realizar una amplia

revisión bibliográfica, se caracterizan los suelos y la nutrición de las plantaciones de teca en

América Central y se proponen varias herramientas para la mejora de su gestión.

Las plantaciones de teca de América Central presentan habitualmente deficiencias de K y P,

además de algunos problemas de acidez ocasionales. Estos se originan, principalmente, por la

mala selección de sitio que se realizó en las últimas dos décadas del siglo XX y por el

establecimiento de plantaciones de teca por pequeños propietarios en terrenos que no tienen

características propicias para la especie. Además, estos problemas comunes relativos a la baja

disponibilidad de P y de K en el suelo son causantes de las relativamente bajas concentraciones

foliares de estos elementos (0,88±0,07% K y 0,16±0,04% P) encontradas en plantaciones de

teca características de la región.

Se presentan varios modelos estadísticos que permiten a los gestores: (a) usarlos como

referencia para la interpretación de análisis foliares, ya que ofrecen valores que se consideran

característicos de plantaciones de teca con un buen estado nutricional; (b) estimar la cantidad de

nutrientes acumulados en la biomasa aérea de sus plantaciones y, sobre todo, su extracción a

través de la madera en un aprovechamiento forestal, bien sea una clara o la corta final.

iv

La gran acumulación de N, P y K en plantaciones de teca ha de ser considerada como un

factor fundamental en su gestión. Además, P y K adquieren mayor relevancia aún ya que su

extracción del sistema a través de la madera y su escasa disponibilidad en los suelos hacen que

se presente un importante desequilibrio que pone en riesgo la sostenibilidad del sistema. En ese

sentido, cambiar la época de cosecha, de la actual (en Enero-Mayo) a Septiembre o Diciembre,

puede reducir entre un 24 y un 28% la salida de N asociada a la extracción de madera, un 29%

la de P y entre un 14 y un 43% la de K.

Se estima que la concentración foliar de P es un factor limitante de la productividad de

plantaciones de teca en América Central, proponiéndose un nivel crítico de 0,125%. Además, la

teca presenta una tolerancia muy baja a suelos salinos, tendencia que no había sido señalada

hasta el momento, siendo muy alta la probabilidad de que la plantación tenga un crecimiento

lento o muy lento cuando la Saturación de Na es mayor de 1,1%. Por otro lado, se confirma que

K es uno de los elementos clave en la nutrición de las plantaciones de teca en la región

centroamericana, proponiéndose un nivel crítico provisional de 3,09% para la Saturación de K,

por encima del cual es muy probable que la plantación tenga un crecimiento muy alto.

Se ha comprobado que las técnicas estadísticas de análisis multivariante pueden ser usadas

como herramientas para agrupar los rodales en base a sus similitudes en cuanto a la fertilidad

del suelo y mejorar así el diseño de planes de fertilización en plantaciones con una superficie

relativamente grande. De esta manera, se pueden ajustar planes de fertilización más eficientes a

escala de grupos de rodales, como un primer paso hacia la selvicultura de precisión,

intensificando y diversificando la gestión en función de las diferencias edáficas.

Finalmente, aunque los análisis foliares y de suelos indiquen la existencia de deficiencias

nutricionales, la fertilización de las plantaciones no siempre va a producir efectos positivos

sobre su crecimiento si no se diseña adecuadamente teniendo en cuenta varios factores que

pueden estar influyendo negativamente en dicha respuesta, como la densidad de las plantaciones

(sinergias con la programación de los clareos y claras) y la elección de la dosis y del producto a

aplicar (habitualmente dosis bajas de N-P-K en lugar de incluir otros nutrientes como Mg, B y

Zn o usar otros productos como micorrizas, biofertilizantes etc…).

v

ABSTRACT

Teak (Tectona grandis L.f.) has been traditionally considered as a precious wood in SE Asia,

where it is indigenous. However, during recent decades the species has reached worldwide

relevance in the tropical high quality hardwood sector. Teak has been widely established in

Central America, where it has become a key species in the forest sector due to its

socioeconomic impact, either because of the big-scale plantations of transnational companies

and the abundant small-scale plantations established by many farmers. Despite the relevance of

the species, little research has been carried out regarding its soil fertility and nutrition

management, a key issue both for sustainability and productivity. The present Thesis performs a

literature review to this respect, characterize the soil fertility and the nutrition of teak plantations

of Central America and propose several management tools.

Soil deficiencies of K and P are usually found in teak plantations in Central America, in

addition to occasional acidity problems. These problems are mainly derived of (a) a poor site

selection performed during 80s and 90s; and (b) small-scale plantations by farmers in sites

which are not adequate for the species. These common soil fertility problems related with P and

K soil availability are probably the cause of the relatively low P and K foliar concentration

(0,88±0,07% K y 0,16±0,04% P) found in representative teak plantations of the region.

Several statistical models are proposed, which allow forest managers to: (a) use them as a

reference for foliar analysis interpretation, as they show values considered as representative for

teak plantations with an adequate nutritional status in the region; (b) estimate the amount of

nutrients accumulated in the aerial biomass of the plantations and, especially, the amount of

them which are extracted from the systems as wood is harvested in thinning or final clearcuts.

The accumulation of N, P and K result in a key factor for teak management in the region.

This turns out to be especially relevant for the P and K because their high output rate by timber

extraction and the low soil availability result in an important unbalance which constitutes a risk

regarding the sustainability of the system. To this respect, modifying the harvesting time from

the usual right now (January-May, business as usual scenario) to September or December

vi

(proposed alternatives) can reduce between 24 and 28% the N output associated to timber

extraction, 29% the P output and between 14 and 43% the K.

Foliar P concentration is a main limiting factor for teak plantations productivity in Central

America and a 0.125% critical level is proposed. In addition, the results show a very low

tolerance for soil salinity, tendency which was not previously reported. Hence, the probability

of teak plantations to have low or very low Site Index is high when Na Saturation is higher than

1.1%. On the other hand, K is confirmed as one of the key nutrients regarding teak nutrition in

Central America and a 3.09% provisional critical level is proposed for K Saturation; when

values are above this level the probability of having very high Site Index is high.

Multivariate statistical analyses have been successfully tested to be used as tools to group

forest stands according to their soil fertility similarities. Hence, more efficient fertilization plans

can be designed for each group of stands, intensifying and diversifying nutritional management

according to soil fertility differences. This methodology, which is considered as a first step

towards precision forestry, is regarded as helpful tool to design fertilization plans in big scale

plantations.

Finally, even though foliar and soil analysis would point out some nutritional deficiencies in

a forest stand, the results show how the fertilization is not always going to have a positive effect

over forest growth if it is not adequately designed. Some factors have been identified as

determinants of tree response to fertilization: density (synergisms between fertilization and

thinning scheduling) and the appropriate selection of dosages and product (usually low dosages

are applied and N-P-K is preferred instead of applying other nutrients such as Mg, B or Zn or

using other alternatives such as mycorrhizas or biofertilizers).

vii

PRÓLOGO

La presente tesis doctoral ha sido realizada gracias a una beca de formación predoctoral de la

Universidad Politécnica de Madrid [ayuda del programa propio de la UPM del personal

investigador en formación para la realización del Doctorado en sus Departamentos, Centros

I+D e Institutos (RR01/2009)] en su modalidad de Cooperación al Desarrollo. Además de esta

beca, varios proyectos de I+D+i y de Cooperación al Desarrollo han financiado parcialmente las

investigaciones que componen esta tesis, todos ellos enmarcados dentro del Programa CAB

(www2.caminos.upm.es/Departamentos/imt/Topografia/Cab/cab.html, Comunidad Agua y

Bosque) y financiados por la Agencia Española de Cooperación Internacional para el Desarrollo

(AECID), la Universidad Politécnica de Madrid (UPM-Cooperación) y la Conferencia Nacional

de Rectores de Costa Rica (CONARE): Manejo Integral de Agua y Suelo (MAIAS),

Fortalecimiento de la Red para el Manejo Integral de Agua y Suelo (FORMAIAS), Manejo

comunitario de Suelo y Agua en Centroamérica (MACOSACEN), Red Universitaria para el

manejo de Agua y Suelo en Centroamérica (REUNAS), Comunidad, Agua y Bosques (CAB),

Vulnerabilidad, impactos y adaptación al cambio climático sobre recursos hídricos en

Iberoamérica (VIAGUA), y Manejo agroforestal participativo como inicio de encadenamientos

productivos en fincas integrales.

Por otro lado se ha contado con la financiación directa e indirecta y la colaboración de

empresas del sector forestal en América Central (Panamerican Woods, Green Millenium y

Ecoforest), de los Laboratorios de Recursos Naturales y de Suelos y Foliares del Centro de

Investigaciones Agronómicas de la Universidad de Costa Rica (CIA-UCR), del Laboratorio de

Topografía del Departamento de Ingeniería y Morfología del Terreno de la Universidad

Politécnica de Madrid y del Departamento de Silvopascicultura de la Universidad Politécnica de

Madrid.

viii

El hecho de que la tesis se haya realizado en un marco de Cooperación al Desarrollo ha

condicionado que la investigación realizada haya sido eminentemente aplicada para que sea

fácilmente transferida a los gestores y propietarios y que éstos puedan llevar a la práctica los

resultados de investigación obtenidos. Estos resultados se han plasmado en siete artículos

científicos de los cuales tres ya están publicados (o aceptados) y otros cuatro se encuentran en

fase de revisión en revistas científicas incluidas en el Science Citation Index (SCI). Estos siete

trabajos componen la mayoría de capítulos de la tesis y, además, se incluyen como Anexos los

que han sido publicados en su formato definitivo de publicación:

1) Fernández-Moya J, Alvarado A, Mata R, Thiele H, Segura JM, Vaides E, San Miguel-Ayanz A,

Marchamalo-Sacristán M. Soil fertility characterization of teak (Tectona grandis L.f.) plantations

in Central America. En revisión en European Journal of Forest Research.

2) Fernández-Moya J, Murillo R, Portuguez E, Fallas JL, Ríos V, Kottman F, Verjans JM, Mata R,

Alvarado A. 2013 Nutrient concentration age dynamics of teak (Tectona grandis L.f.) plantations

in Central America. Forest Systems 22 (1): 123-133. [Anexo I].

3) Fernández-Moya J, Murillo R, Portuguez E, Fallas JL, Ríos V, Kottman F, Verjans JM, Mata R,

Alvarado A. Nutrient accumulation and export in teak (Tectona grandis L.f.) plantations in Central

America. Aceptado en iForest - Biogeosciences and Forestry.

4) Fernández-Moya J, Alvarado A, Morales M, San Miguel-Ayanz A, Marchamalo-Sacristán M.

2014. Using multivariate analysis of soil fertility as a tool for forest fertilization planning. Nutrient

Cycling in Agroecosystems 98 (2): 155-167. [Anexo II].

5) Fernández-Moya J, Alvarado A, Verjans JM, San Miguel-Ayanz A, Marchamalo-Sacristán M.

Preliminary soil and foliar critical levels for teak (Tectona grandis L.f.) plantations in Central

America: a study case in Panama. En revisión en European Journal of Forest Research.

6) Fernández-Moya J, Algeet-Abarquero N, Cabalceta G, Alvarado A, San Miguel-Ayanz A,

Marchamalo-Sacristán M. Modifying harvesting time as a tool to reduce nutrient export by timber

extraction: a case study in teak (Tectona grandis L.f.) planted forests in Costa Rica. Enviado a

Forest Ecology and Management.

7) Fernández-Moya J, Alvarado A, Fallas JL, San Miguel-Ayanz A, Marchamalo-Sacristán M. Is N-

P-K fertilization of teak (Tectona grandis L.f.) plantations always a good choice? Some lessons

learned from a case study in Costa Rica. En revisión en European Journal of Forest Research.

ix

Además de estos 7 artículos publicados, o que pretenden ser publicados, en revistas

indexadas en el SCI, los resultados de la tesis se han expuesto en 8 comunicaciones a congresos,

talleres y seminarios nacionales e internacionales y se incluyen en tres publicaciones de carácter

divulgativo:

1) Fernández-Moya J, Alvarado A, San Miguel-Ayanz A, Marchamalo-Sacristán M. Forest nutrition

and fertilization in teak (Tectona grandis L.f.) plantations in Central America. En revisión en un

número especial de la New Zealand Journal of Forestry Science correspondiente al III

International Congress on Planted Forests

2) Fernández-Moya J. Las plantaciones forestales como alternativa de desarrollo sostenible. En:

Manejo Integral de Agua y Suelo en Centroamérica. Bases científicas para el desarrollo rural

comunitario. Algeet N, Fernández J, Lianes E, Marchamalo M, Martínez R, Rejas JG (eds).

Universidad Politécnica de Madrid, Programa de Cooperación Comunidad Agua y Bosque en

Centroamérica. Madrid (España). ISBN: 84-7493-467-2. pp: 164 – 168.

3) Fernández-Moya J, Alvarado A, San Miguel-Ayanz A, Marchamalo-Sacristán M. Guía de

recomendaciones para la gestión de la fertilidad del suelo y la nutrición de plantaciones de teca

(Tectona grandis L.f.) en América Central. En preparación.

Precisamente la divulgación y transmisión de la investigación realizada es imprescindible

para cumplir con el principal objetivo que se busca, en esta área temática, desde el Programa

CAB: colaborar para mejorar la gestión sostenible de las plantaciones forestales de la región. En

ese sentido, una vez realizada y publicada la investigación, desde el Programa CAB se plantea

una segunda fase complementaria con visitas de campo con varios usuarios potenciales para

asegurar la transferencia de la investigación realizada.

x

CAPÍTULO 1

INTRODUCCIÓN

3

1.1. Justificación

La superficie ocupada por plantaciones forestales ha crecido exponencialmente en las

últimas décadas y lo mismo ha sucedido con su relevancia en el sector forestal internacional.

Además, se prevé que esta tendencia se mantenga en el futuro próximo. Esto se debe a que el

sector debe hacer frente a la creciente demanda social de madera y otros productos forestales no

maderables, además de a las necesidades de los servicios ambientales que proveen los sistemas

forestales: hidrológicos, paisaje y recreo (incluyendo a veces religión y cultura), conservación

de la biodiversidad y fijación y reserva de carbono. Aunque algunos de estos servicios

ambientales sean provistos por las plantaciones forestales en menor medida que por los bosques

naturales, las plantaciones son contempladas como una alternativa de desarrollo sostenible que

permite obtener una productividad relativamente alta a la vez que mantiene parte de estos

servicios ambientales. Además, su alta productividad supone que las demandas del mercado son

satisfechas principalmente por los productos que se obtienen de ellas, lo que supone liberar a los

bosques naturales de una parte sustancial de la acción antrópica, con las ventajas que esto

conlleva para su conservación, especialmente en regiones tropicales.

Históricamente, las especies forestales se veían relegadas a los terrenos que no servían para

la agricultura por lo que los suelos forestales se consideran poco fértiles y con numerosos

problemas para su gestión. Desde hace unos años se ha producido un cambio de paradigma en

ese aspecto, buscándose algunos suelos fértiles (marginales para la agricultura o incluso buenos

para ella) para el establecimiento de plantaciones forestales de gestión intensiva. Así, en la

actualidad, los estudios de fertilidad de suelos en sistemas forestales se hacen fundamentalmente

con dos motivos:

1. Para evaluar la disponibilidad actual de nutrientes de un suelo, analizar si existen

deficiencias o toxicidades y elaborar un plan de gestión nutricional acorde. Esta

evaluación se podría hacer de forma periódica en los sistemas forestales gestionados

intensamente de manera que se pueden detectar problemas nutricionales antes de ser

demasiado severos, aunque en la práctica se suele hacer en aquellas zonas en las que se ha

detectado un problema y se piensa que puede ser debido a problemas edáficos

4

2. Para evaluar la potencialidad del suelo antes de establecer una nueva plantación forestal

con la finalidad de:

a. elegir sitios que cumplan con unas condiciones de fertilidad adecuadas para la

especie, descartando sitios que presenten problemas edáficos o que resulte

antieconómico enmendarlos, cuando se están buscando nuevas tierras que comprar

b. elegir una especie cuyos requerimientos edáficos se adapten a la fertilidad del

terreno cuando éste ya ha sido elegido (comprado previamente, heredado …)

En ese sentido, la Sociedad Española de Ciencias Forestales (2005) define edafología como

la ciencia que estudia el suelo y su influencia sobre los seres vivos, particularmente sobre las

plantas, y la utilización del suelo por el hombre como medio de cultivo de éstas; y selvicultura

como la teoría y práctica sobre el establecimiento, desarrollo, composición, sanidad, calidad,

aprovechamiento y regeneración de las masas forestales, para satisfacer las diversas demandas

de la sociedad, de forma contínua o sostenible. De igual manera, se definen los tratamientos

selvícolas como las intervenciones a las que se somete una masa forestal con el fin de que pueda

cumplir mejor los objetivos a que esté destinada, asegurando su mejora y regeneración. Así, la

gestión de la fertilidad del suelo y la nutrición de sistemas forestales no dejan de ser una serie de

tratamientos selvícolas planificados tomando en consideración la edafología.

La teca (Tectona grandis L.f.) es una especie tropical que proporciona una madera noble, de

calidad similar a la de otras como el cedro (Cedrela odorata L.) y la caoba (Swietenia

macrophylla King). Las plantaciones de teca han aumentado mucho desde la década de 1980.

En América Central son muy abundantes ya que representan una actividad económica con una

mayor rentabilidad que la ganadería o que algunos cultivos agrícolas. En ese sentido, la especie

ha sido ampliamente establecida tanto por empresas multinacionales, en grandes plantaciones

forestales con una gestión relativamente intensiva, como por pequeños propietarios, que han

establecido pequeñas plantaciones o han plantado la teca en sistemas agroforestales.

5

Pese a no ser una especie de crecimiento rápido per se, la teca ha demostrado ser una especie

que puede presentar crecimientos bastantes altos (5–15 m3 ha

-1 año

-1) si se gestiona

adecuadamente y la calidad de sitio es buena. No obstante, si la gestión es mala y la elección de

sitio no ha sido adecuada, las expectativas de crecimiento no se cumplen y la productividad

obtenida es relativamente baja ( < 5 m3 ha

-1 año

-1). Se ha observado, además, que la

productividad ha disminuido en plantaciones de teca que se encuentran en su segundo o tercer

turno (o periodo de rotación), lo que supone un problema no sólo de producción sino de

sostenibilidad del sistema. Así, se considera que la edafología y la gestión de la fertilidad del

suelo y la nutrición forestal son clave para la obtención de altas tasas de productividad en este

tipo de sistemas, así como para asegurar su sostenibilidad.

1.2. Objetivos

En base a lo expuesto en el punto anterior se plantea este trabajo, que tiene como objetivo

general analizar las relaciones suelo-planta en plantaciones de teca de América Central,

evaluando su sostenibilidad y desarrollando herramientas para mejorar su productividad. Para

conseguirlo, se han marcado los siguientes objetivos específicos:

1. Caracterizar los suelos de las plantaciones de teca en América Central (OE-1)

2. Analizar la dinámica de la concentración de nutrientes en plantaciones de teca en función

de la edad de las plantaciones y creación de unas modelos que sirvan de referencia para

evaluar el estado nutricional de otras plantaciones en la región (OE-2)

3. Analizar la dinámica de la acumulación de nutrientes en plantaciones de teca en función

de la edad de las plantaciones y evaluar sostenibilidad de las mismas en base a la

extracción de nutrientes del sistema (OE-3)

4. Diseñar alternativas que minimicen esta extracción de nutrientes del sistema para mejorar

la sostenibilidad del mismo (OE-4)

5. Analizar la influencia de variaciones nutricionales sobre el crecimiento de plantaciones de

teca y estimar los niveles críticos preliminares para algunos nutrientes (OE-5)

6

6. Proponer alternativas que mejoren las prácticas de fertilización que se llevan a cabo en

estos sistemas (OE-6)

1.3. Esquema de trabajo

Para cumplir con los objetivos propuestos se han empleados dos escalas de trabajo. En el

caso del objetivo específico 1 (OE-1) se ha trabajado con datos a escala de país en Guatemala,

Costa Rica y Panamá (Figura 1.1). Simultáneamente, para alcanzar el resto de objetivos

específicos (OE-2, OE-3, OE-4, OE-5 y OE-6), se han llevado a cabo estudios más detallados en

algunas regiones concretas: Guanacaste y San Carlos (o Zona Norte), en Costa Rica, y la

Cuenca del Canal de Panamá, en Panamá (Figura 1.1). Se considera que las zonas de estudio

elegidas son representativas de la gran superficie potencial para el establecimiento de

plantaciones de teca en América central (Figura 1.2).

Figura 1.1. Zonas de estudio en América Central. En verde los países en los que se ha trabajado a escala

nacional con datos de localidades dispersas (Guatemala, Costa Rica y Panamá) y marcadas en rojo

algunas regiones específicas donde se han llevado a cabo estudios de caso a nivel local (Guanacaste y San

Carlos en Costa Rica y el Canal de Panamá)

7

Figura 1.2. Área potencial (coloreada) para el establecimiento de plantaciones de teca (Tectona grandis

L.f.) en América Central en función de las regiones bioclimáticas propuestas por Holdridge (1967)

La tesis está dividida en 4 bloques (Figura 1.3), a saber:

BLOQUE I --- Marco teórico

En él se exponen conceptos básicos sobre el papel de las plantaciones forestales en el

mundo, la gestión de la fertilidad del suelo, la nutrición forestal y su relación con varios

tratamientos selvícolas (p.ej. fertilización, desbroces, claras y clareos). Además, se aborda el

caso concreto de las plantaciones de teca y, específicamente, de su nutrición y su relación con el

suelo. Se consideran las ideas expuestas en este bloque como fundamentales a la hora de

interpretar los resultados expuestos en los siguientes.

BLOQUE II --- Caracterización de las plantaciones de teca en América Central

En este bloque (correspondiente con los capítulos 3, 4 y 5 y a los OE-1, OE-2 y OE-3) se

analiza la fertilidad de los suelos sobre los que se han establecido plantaciones de teca en la

región, para lo que se usa una base de datos de 684 sitios diseminados entre Panamá, Costa Rica

8

y Guatemala. Además, se exponen las curvas de absorción de nutrientes de la especie, cuyo

objetivo es analizar la variación con la edad de la concentración y la acumulación de nutrientes

en diferentes tejidos de la planta, para lo que se usaron tres estudios de caso en el Canal de

Panamá, Guanacaste y San Carlos (Figura 1.1). Estos trabajos no sólo permiten caracterizar las

plantaciones de teca en la región sino que, además, establecen referencias que sirven como guías

a la hora de interpretar análisis foliares y evaluar el estado nutricional de las plantaciones. Por

otra parte, permiten evaluar los problemas de sostenibilidad con respecto a la salida de

nutrientes del sistema por la extracción de la madera.

Figura 1.3. Esquema de los contenidos de la presente tesis y su relación con los capítulos de la misma,

distribuidos en bloques temáticos

9

BLOQUE III --- Herramientas para la gestión del suelo y la nutrición de teca

En base al análisis realizado en el bloque II, en este bloque (correspondiente a los capítulos

6, 7, 8 y 9 y a los OE-4, OE-5 y OE-6) se abordan cuatro líneas de investigación de carácter

aplicado. Su objetivo es proporcionar herramientas útiles para que por los técnicos de las

plantaciones puedan mejorar su gestión: (a) Se exploran posibilidades para reducir la cantidad

de nutrientes que salen del sistema como consecuencia de la extracción de madera; (b) se

delimitan niveles críticos para algunos nutrientes que permiten una mejor interpretación de los

análisis foliares y de suelos; (c) se muestra cómo se pueden usar los análisis estadísticos

multivariantes para una mejor interpretación de las bases de datos de análisis de suelo y para

mejorar la planificación de la fertilización en plantaciones grandes; y (d) se analizan posibles

prácticas para mejorar la eficiencia de la fertilización.

BLOQUE IV --- Discusión general, implicaciones para la gestión y conclusiones

Este último bloque, de carácter integrador, ofrece una discusión general de los resultados

obtenidos en los bloques anteriores de la que, finalmente, se deducen las conclusiones de la

Tesis Doctoral. Aparte de integrar líneas de investigación, en este bloque se trata de

compatibilizar aspectos de carácter científico con otros más prácticos, intentando generar

recomendaciones para la gestión. Estos últimos puntos están enfocados, además, a poder

publicarlos como un manual que pueda ser distribuido a gestores y propietarios para que puedan

tener en cuenta estos resultados en la gestión de sus plantaciones.

10

CAPÍTULO 2

MARCO TÉORICO

13

2.1. Bosques plantados para satisfacer demandas humanas

2.1.1. Aspectos generales de los bosques plantados

Durante las últimas décadas la demanda mundial de madera ha seguido una tendencia

creciente que ha obligado al sector forestal a hacer frente a la necesidad de mayor producción.

De hecho, se prevé que siga incrementándose a un ritmo incluso más acelerado en los próximos

años (Evans 2009; FAO 2009). Para poder satisfacer esta creciente demanda de madera a la vez

que se conservan los bosques naturales y se cumplen otros objetivos de conservación de la

naturaleza, durante las últimas décadas ha aumentado en gran medida el área de bosques

plantados (FRA 2010). Este crecimiento ha ocurrido especialmente en Asia y en América del

Norte y Central (FRA 2010), llegando hasta las 264 ·106 ha (7% del área total mundial

forestada). El aumento de los bosques plantados ha incrementado significativamente el área

forestada mundial y ha compensado parcialmente la deforestación sufrida en otras zonas (FRA

2010).

La nomenclatura propuesta por FAO (Evans 2009) considera bosques plantados todos

aquellos que son establecidos por plantación o semillado, en contraposición a aquellos bosques

naturales cuyo método de regeneración es natural. A pesar de la diversidad de tipos de bosque

que se engloban en esta categoría, éstos presentan numerosas similitudes desde un punto de

vista ambiental ya que tienen en común estar condicionados por la regeneración artificial de las

masas. No obstante, se establecen tres grandes grupos en función de los objetivos principales

que éstos busquen y su gestión selvícola: (a) bosques semi-naturales de especies nativas con

diferentes intensidades en la gestión y con regeneración artificial que mantiene el mismo

sistema previo a la plantación; (b) plantaciones forestales “productoras” para la producción de

madera y otros productos forestales no maderables, comúnmente denominadas plantaciones

forestales; (c) plantaciones forestales “protectoras” para la provisión de servicios ambientales

(Evans 2009). Así, a lo largo del presente documento se mantendrá esta nomenclatura teniendo

en cuenta que las llamadas plantaciones forestales no son sino un grupo de bosques plantados.

Con tan sólo el 7% del área forestal total, los bosques plantados satisfacen una gran parte de

la demanda mundial de madera: aproximadamente 65-70% según Evans (2009) o 33-66% según

14

la discusión abierta que se produjo en el III Congreso Internacional de Bosques Plantados

(EFIATLANTIC 2013). Esta tendencia supone a su vez una liberación de presión antrópica

sobre los bosques naturales, los cuáles han visto reducir su aprovechamiento en un 25-30%, con

los beneficios que esto supone para la conservación de la naturaleza (Evans 2009;

EFIATLANTIC 2013). No obstante, hay autores que también advierten de un riesgo de pérdida

de valor del bosque natural como consecuencia de esta tendencia (EFIATLANTIC 2013). Así,

al aumentar la oferta mundial de madera como consecuencia de la alta producción de las

plantaciones forestales intensivas, este bien de consumo bajará su precio internacional y por lo

tanto la madera existente en los bosques naturales tendrá menor valor comercial, lo que podría

traducirse en una eliminación del bosque natural para convertirlo a otro uso con una mayor

productividad (Buongiorno 2013).

A menudo se describe a las plantaciones forestales como carentes de sostenibilidad social y

ambiental, causantes de deforestación de bosques naturales y en manos de grandes empresas

multinacionales, aunque la realidad parece ser otra (Carle 2013). Así, aproximadamente el 50%

de los bosques plantados son de propiedad pública, el 33% está mamos de pequeños

propietarios y tan sólo el 15-17% pertenece a grandes empresas (Carle 2013). Además, el 75%

de los bosques plantados se establecen con especies nativas (FRA 2010). Asimismo, la

certificación forestal es frecuente en los bosques plantados lo que permite asegurar un cierto

nivel de sostenibilidad y, sobre todo, ha disminuído el problema de la deforestación de bosques

naturales para el posterior establecimiento de bosques plantados, ya que está prohibido en los

estándares de las certificadoras, aunque el problema sigue existiendo en algunas zonas

(EFIATLANTIC 2013). Se estima que el 76% de los bosques plantados tenían en 2005 la

producción de madera u otros productos forestales no maderables como objetivo principal,

porcentaje que seguramente haya descendido ligeramente en estos últimos años al aumentar en

gran medida las plantaciones protectores a escala mundial (FRA 2010).

Aunque las zonas tropicales geográficamente se delimitan entre los trópicos de Cáncer y

Capricornio (23º 26’ N y S, respectivamente), habitualmente se consideran como plantaciones

tropicales aquellas establecidas en un rango más amplio que llega hasta los 27-28 N y S de

15

manera que se incluyen las grandes plantaciones de zonas subtropicales como, por ejemplo,

Queensland, Swaziland, Sao Paolo, México, India o China (Evans y Turnbull 2004). Las

plantaciones forestales en estas zonas se incrementan en gran medida a partir de la década de

1960 y adquieren una gran importancia gracias a los incentivos fiscales y los pagos por servicios

ambientales de los que fueron objeto en algunos países (Evans y Turnbull 2004). Durante la

década de 1990 se plantaban aproximadamente 230.000 ha cada año en América tropical y en el

año 2000 el 40% de las plantaciones forestales mundiales eran tropicales, aproximadamente 90

·106 ha (Evans y Turnbull 2004). En la actualidad se estima que hay 125-150 ·10

6 ha de bosques

plantados en los trópicos (Evans y Turnbull 2004).

Las plantaciones forestales consideradas como tropicales tienen algunos aspectos, tanto

ambientales como socio-económicos, que las diferencian de las de otras regiones y que se

exponen a continuación. En primer lugar cabe destacar las altas productividades de los sistemas

forestales en estas regiones que llegan hasta los 70 – 90 m3 ha

-1 año

-1 de incremento medio anual

(Evans y Turnbull 2004). Otra característica reseñable es la situación económica y socio-política

de los países de estas regiones, muchos de los cuales se consideran como en vías de desarrollo

(Evans y Turnbull 2004). El bajo índice de desarrollo y los problemas económicos de estos

países, junto con la falta de oportunidades de empleo en las zonas rurales, hacen que las

inversiones de capital exterior en grandes plantaciones sean interesantes como fuentes de trabajo

y riqueza. Por otro lado, en muchas ocasiones los procedimientos mediante los cuáles se

establecen estas inversiones pueden generar problemas, por ejemplo: (a) las concesiones de

terrenos estatales a empresas extranjeras generan conflictos con los habitantes locales, como ha

sucedido en numerosas plantaciones forestales en Asia; (b) la compra de grandes terrenos para

el establecimiento de plantaciones forestales se ha visto ligada en algunas ocasiones a

operaciones de lavado de dinero, especulación e incluso fraudes financieros (Evans y Turnbull

2004; De Camino y Morales 2013; EFIATLANTIC 2013). Además, la situación sociopolítica

de algunos países hace que en algunos casos no se realicen inversiones de gran calibre por

riesgos asociados a lo que algunos inversores consideran inestabilidad política (EFIATLANTIC

2013). Con respecto a estos problemas, la gran cantidad de plantaciones forestales certificada

16

(sobre todo con sello FSC) garantiza en parte el cumplimiento de estándaers básicos desde el

punto de vista socio-económico y laboral, además de buscar la sostenibilidad ambiental

(EFIATLANTIC 2013).

2.1.2. Aspectos ambientales de los bosques plantados

Independientemente de si su objetivo principal es la oferta de productos forestales o la

restauración ambiental, los bosques plantados generan beneficios ambientales: hidrológicos,

biodiversidad, fijación de carbono, paisaje, recreo, cultura y religión. De hecho, estos servicios

se recompensan económicamente en algunos países mediante los Pagos por Servcios

Ambientales (PSA) (p.ej. Arias 2004; Brauman et al. 2007; FONAFIFO 2010).

Desde el punto de vista de la biodiversidad, Brockerhoff et al. (2008) señalan que las

plantaciones forestales pueden, aunque en menor grado que los bosques naturales, constituir

valiosos habitats para algunas especies amenzadas y contribir a la conservación de la

biodiversidad mediante varios mecanismos. En regiones en las que el bosque fuese la cobertura

vegetal del suelo natural, las plantaciones forestales representan un uso del suelo relativamente

similar y la reforestación de tierras puede ayudar a la conservación creando habitats

complementarios, aprovechando áreas limítrofes de bosques naturales con menor impeacto que

otros usos y, sobre todo, incrementando la conectividad (Brockerhoff et al. 2008). Sin embargo,

la deforestación de bosques naturales y su posterior conversión a plantaciones forestales supone

una gran pérdida de biodiversidad y, en menor grado, la forestación de tierras que no eran

forestales de forma natural puede también tener efectos negativos (Brockerhoff et al. 2008).

Finalmente, Brockerhoff et al. (2008) resumen el concepto considerando a las plantaciones

forestales con una gestión adecuada como un mal menor comparado con la presión

deforestadora que ejercen habitualmente la agricultura y la ganadería.

Por otro lado, dentro de las actividades que se plantean para la reducción del incremento de

la concentración de CO2 atmosférico, en el marco de las políticas asociadas al cambio climático,

están (a) la posibilidad de reforestar tierras degradads para fijar C y (b) la gestión sostenible de

los recursos forestales para evitar su degradación y reducir el consumo de combustibles fósiles

17

(Marland y Schlamadinger 1997). El secuestro y la fijación de C de las plantaciones forestales

es ampliamente reconocido y valorado a veces con pagos por servicios ambientales,

estimándose que se acumulan entre 0,4 y 8 t C ha-1

año-1

en plantaciones forestales en regiones

tropicales (Evans y Turnbull 2004; Canadell y Raupach 2008). Sin embargo, el establecimiento

masivo de plantaciones forestales con el objetivo de fijar C ha sido a su vez criticado, ya que se

considera que podría tener consecuencias negativas, sobre todo teniendo en cuenta el consumo

de agua de estos sistemas (Jackson et al. 2005; Whitehead 2011).

A los bosques plantados también se les atribuye la prestación de servicios hidrológicos (p.ej.

Bruijnzeel 2004; Brauman et al. 2007; van Dijk y Keenan 2007). La reforestación a veces

consigue recuperar parcial o totalmente las propiedades hidrológicas de los suelos degradados

(Brauman et al. 2007; Ilstedt et al. 2007). Por otro lado, el gran consumo de agua de los bosques

y las plantaciones forestales ha generado una gran polémica durante el siglo XX, que continúa

en los comienzos de este siglo XXI como contraposición a sus beneficios hidrológicos

asociados (para una revisión exhaustiva ver Bruijnzeel 2004). Cuando estas funciones

hidrológicas de los ecosistemas se recuperan, se reduce la escorrentía superficial, los daños por

tormentas y las inundaciones a pequeña escala, e incluso se pueden producir aumentos en la

recarga de agua subterránea si esa mejora de la infiltración es mayor que el consumo de agua de

la propia plantación (van Dijk y Keenan 2007).

2.1.3. Aspectos socioeconómicos de los bosques plantados

En Costa Rica, según Arias (2004), las plantaciones forestales:

1) “además de producir importantes servicios ambientales (fijación de carbono,

protección de suelo y agua, mejora del paisaje), generan mucho empleo

(principalmente mano de obra no calificada) y desarrollo económico en las áreas

rurales de mayor pobreza”

2) “han contribuido significativamente a un desarrollo territorial equitativo y sostenible

porque, entre otras cosas, promueven el acceso de pequeños y medianos propietarios”

18

3) “las empresas individuales que han alcanzado una masa crítica mínima (Maderas

Cultivadas de Costa Rica, Pan American Wood, Flora y Fauna, entre otras), han

invertido en la industria, han desarrollado tecnología y han sido exitosas introduciendo

y posicionando productos, tanto en el mercado nacional como en el internacional. Esto

demuestra claramente que, en el ámbito del país, también se puede lograr el desarrollo

de plantaciones competitivas, siempre que se tome la decisión política de apoyar el

establecimiento de la masa crítica mínima”.

Según el análisis realizado por Arias (2004) para el caso concreto de Pan American Woods,

empresa en cuyas plantaciones se llevó a cabo una parte de la presente investigación (ver zona

de Guanacaste en la Figura 1.1), se observa que la empresa había plantado hasta el año 2004

alrededor de 3.000 ha de teca, distribuidas en las municipalidades de Nicoya y Nandayure

(Costa Rica), ambas con un alto índice de pobreza, y había invertido alrededor de 50 ·106 US $

en compra de tierras, plantación, caminos, puentes, viviendas, talleres e industria. En el año

2004 la empresa generaba 230 empleos directos entre la plantación y la industria. La empresa ha

resultado ser un fuerte apoyo a pequeños productores independientes al actuar de intermediaria

y permitir que estos puedan exportar al mercado internacional.

Evans y Turnbull (2004), desde un punto de vista global, también reconocen el importante

papel de los bosques plantados en el desarrollo rural de regiones tropicales incluyendo los

sistemas agro-silvo-pastorales, las pequeñas plantaciones forestales para leña o madera y otros

usos como la extracción de aceites y esencias, la apicultura, religioso-culturales, etc.

2.1.4. Plantaciones forestales como alternativa de desarrollo

Basándose en los puntos expuestos anteriormente, las plantaciones forestales se han

identificado como un uso de suelo que supone una alternativa potencialmente sostenible para el

desarrollo rural en zonas en las las que la agricultura y la ganadería extensivas tienen serios

problemas ambientales o no son económicamente rentables. Además, las plantaciones forestales

también suponen una alternativa para zonas en las que se quiere mantener una cobertura

forestal, con una apariencia y una estética paisajistica más parecida a la natural (p. ej. para el

19

turismo) y a la vez obtener beneficios en forma de madera u otros productos forestales no

maderables.

El sector de las plantaciones forestales en América Central puede dividirse en dos grupos: (a)

empresas internacionales con inversión de gran capital en plantaciones de especies con mercado

internacional consolidado; y (b) pequeños productores con pequeña inversión de capital y

plantaciones enfocadas a un mercado local (nacional o regional) y/o internacional (vía

intermediarios). Desde un punto de vista de desarrollo rural, las empresas suponen actividades

económicas en zonas rurales comúnmente deprimidas donde las actividades productivas

escasean y, así, suponen una importante fuente de trabajo en zonas rurales, como se ha

comentado anteriormente. Los pequeños productores generalmente establecen las plantaciones

como un complemento del aprovechamiento principal de sus fincas agro-ganaderas, en muchos

casos en las zonas menos productivas. Las empresas ocupan sitios buenos para asegurar altas

productividades que aseguren el rendimiento para el pago de intereses a sus inversores. En ese

sentido, es común el uso de especies con mercado internacional como, por ejemplo, teca

(Tectona grandis L.f.) y melina (Gmelina arborea Roxb.). Por otro lado, las plantaciones de

pequeños productores normalmente se encuentran en sitios malos o moderados (ácidos, poco

fértiles y/o en laderas) con productividades bajas, dado el alto precio de la tierra en sitios de

mayor calidad. Las especies plantadas en estos casos varían entre exóticas, como teca y melina,

y otras especies nativas, adaptadas a estos suelos ácidos poco fértiles y condiciones de la ladera,

donde alcanzan buenas productividades, llegando incluso a considerarse más rentables

económicamente que las plantaciones de teca (Griess y Knoke 2011).

En este marco, el Programa de Cooperación Comunidad Agua y Bosques (con los diversos

proyectos que se encuentran bajo su paraguas) bajo el cuál se desarrolla la presente Tesis

Doctoral, trabaja para desarrollar herramientas que ayuden en la gestión de plantaciones de dos

especies forestales, teca y amarillón o roble coral (Terminalia amazonia (J.F.Gmel.) Exell), que

mejoren la productividad y posibiliten la sostenibilidad en el uso del recurso suelo a dos escalas

socioeconómicas: grandes empresas y pequeños propietarios.

20

2.2. Fertilidad de suelos y nutrición forestal

2.2.1. Nutrición y fertilización forestal: una perspectiva histórica

Los estudios de nutrición forestal se remontan a los de Jean Baptiste van Helmont a

mediados del S. XVIII analizando cómo las plantas absorbían agua y nutrientes en función de su

crecimiento (Binkley 1986). Sin embargo, los edafólogos no se interesan por tema hasta

mediados del S. XIX. En esa época, Chevalier de Valdrome establece un ensayo y obtiene un

40% de incremento en el crecimiento como respuesta a la fertilización (Binkley 1986). Unas

décadas después (en 1906), la fertilización forestal se analizó en profundidad en el VI Congreso

Internacional de IUFRO, donde se muestran varios ensayos en los que se observa respuesta

positiva a la fertilización con N, P y/o K (Binkley 1986). Los investigadores que trabajaron en

estás líneas de conocimiento desde el principio estuvieron preocupados por dos líneas

fundamentales: (1) la posible relación de la productividad forestal con la nutrición, los

nutrientes disponibles en el suelo y sus ciclos biogeoquímicos (especialmente aquellos procesos

que involucraban la materia orgánica); y (2) los nutrientes acumulados en la biomasa que se

extrae y los efectos que esto acarrea sobre el agotamiento de los nutrientes del suelo, con la

consiguiente preocupación por la productividad futura (Rennie 1955; Binkley 1986).

Un siglo después de los primeros ensayos de fertilización forestal, a mediados del S. XX, la

aplicación de fertilizantes químicos a sistemas forestales había adquirido cierta relevancia y

comenzaba a realizarse de manera comercial, aunque seguía en una fase más investigadora

(Binkley 1986). Esta aplicación comercial tardó en producirse porque (1) los fertilizantes no

eran tan abundantes en aquellos momentos; (2) el precio de la madera no permitía la inversión

financiera necesaria y, sobre todo, (3) muchos forestales de la época consideraban que los suelos

forestales eran sistemas “naturales” que no debían someterse a manipulaciones “artificiales”

(Binkley 1986).

El origen de la silvicultura moderna se encuentra en Europa Central durante los siglos XVIII

y XIX (Evans 2009) y está basado, en su mayoría, en rotaciones largas en muchos casos de

coníferas. La escasa concentración de nutrientes de la madera extraída, comparada con la de

ramas y hojas, que quedaban en el sitio y se reciclaban, se compensaba por las adiciones por

21

meteorización de las rocas del suelo y, sobre todo, de nutrientes disueltos en agua de lluvia.

Estas adiciones, que anualmente no suponen mucho, garantizaban la sustentabilidad nutricional

de estos sistemas que eran gestionados con rotaciones de 80-100 o incluso más años. En las

décadas de 1980 y 1990 hubo otro grupo de trabajos que volvieron a llamar la atención acerca

del problema del empobrecimiento de los suelos a causa de la salida de nutrientes del sistema

asociada a la extracción de la madera (Worrel y Hampson 1997), como se detalla

posteriormente en el apartado 2.2.4.

En la actualidad, la aplicación de enmiendas y fertilizantes químicos es habitual en

numerosos sistemas forestales cumpliendo ese doble objetivo de aumentar la productividad y

asegurar la sostenibilidad (p.ej. Ballard 1984; Gonçalvez 1997; Fox et al. 2000). No obstante, se

considera que hay muchos aspectos técnicos aún por resolver en este campo, que

tradicionalmente tiene poca relevancia en el conjunto del sector forestal. En ese sentido, durante

el III Congreso Internacional de Bosques plantados sólo se presentó un trabajo relacionado con

la nutrición y/o fertilización forestal pese a las continuas referencias a la necesidad de aumentar

la productividad y asegurar la sostenibilidad (EFIATLANTIC 2013).

2.2.2. Fertilidad del suelo en sistemas forestales

La productividad de los sistemas forestales depende, en términos generales y sin contar con

la gestión antrópica, de una serie de factores ambientales como son la radiación, la temperatura

y la disponibilidad o el exceso de agua y nutrientes (Binkley 1986; Nambiar y Brown 1997). La

gestión forestal no puede modificar los factores netamente climáticos, así que los únicos

factores que puede controlar son los relativos a los suelos: el agua y los nutrientes. La

modificación de las propiedades hídricas de un sistema (riegos y drenajes) son operaciones

caras, difíciles de justificar en muchos sistemas forestales. Así, la fertilidad de los suelos es casi

el único de esos grandes factores ambientales susceptible de ser modificado por medio de

prácticas selvícolas directas (p.ej. fertilización) o indirectas (Binkley 1986).

La fertilidad del suelo suele dividirse en física, biológica y química, en función de si hace

referencia a esas propiedades del suelo. No obstante, cuando se menciona la fertilidad del suelo

22

en términos generales se suele hacer referencia a la fertilidad química, al ser ésta dependiente a

su vez de las propiedades físicas y biológicas del suelo y ser la parte de la edafología estudiada

en mayor profundidad (p.ej. Bertsch 1998). Las propiedades físicas y biológicas tienen una gran

influencia en la productividad de los sistemas agrícolas y forestales (p.ej. Binkley 1986;

Nambiar y Brown 1997); sin embargo, en esta Tesis se profundiza principalmente en la

fertilidad química del suelo.

Para analizar la fertilidad del suelo es primero fundamental entender los ciclos de los

nutrientes esenciales (macronutrientes –N, P, K, Ca, Mg y S– y micronutrientes –Fe, Mn, Cu,

Cl, Mo, Zn y B–) para entender (a) el reciclaje de los nutrientes en el sistema y (b) la

disponibilidad de los mismos para su absorción por parte de las plantas (p. ej. Binkley 1986;

Bertsch 1998; Gandullo 2000). La evaluación de la fertilidad del suelo se realiza en base a

análisis de los mismos en laboratorio con metodologías variadas pero que se pueden dividir en

dos grandes grupos: (1) los que miden la cantidad total de elemento en el suelo, y (2) los que

miden la cantidad del elemento en la solución del suelo (Bertsch 1998). Este segundo grupo es

el más común en América Central al considerarse que estima directamente la disponibilidad de

los nutrientes para las plantas (Bertsch 1998) y, por ambos motivos, siempre se va a tratar de

esta disponibilidad de nutrientes a lo largo del presente trabajo.

Se conocen como niveles críticos los umbrales establecidos para la interpretación de la

nutrición de plantas. Estos valores se estiman como el valor de la disponibilidad de un nutriente

que se corresponde con el 90% del rendimiento productivo máximo de un sistema dado (Figura

2.1). Estos umbrales son diferentes para cada elemento y varían según la metodología usada en

el análisis de laboratorio. Además, pueden ser distintos para las diferentes especies en función

de los requerimientos edáficos de cada una, aunque suelen existir algunos establecidos de forma

genérica para su uso cuando no hay unos específicos para una especie en concreto (p.ej. Bertsch

1998). Así, se considera que un nutriente es deficiente cuando la disponibilidad de éste es menor

que ese nivel crítico considerado como umbral y, por lo tanto, ese elemento podría estar

limitando la productividad del sitio. De esta manera, cuando el análisis del suelo indica que la

disponibilidad de ciertos nutrientes está por debajo de sus niveles críticos se puede asumir que si

23

se consiguiese aumentar ésta disponibilidad aumentaríamos de manera considerable el

rendimiento productivo del sistema con el que se esté trabajando (i.e. habría en teoría un efecto

positivo de una hipotética fertilización). De manera contraria, cuando la disponibilidad de los

nutrientes esté por encima de sus niveles críticos, no conseguiríamos aumentar

significativamente el rendimiento productivo aunque aumentásemos más esa disponibilidad (i.e.

no habría efecto de la fertilización). En otros casos, cuando lo que existe es una toxicidad en

lugar de una deficiencia, los valores por debajo del nivel crítico se considerarían adecuados

mientras que los valores en exceso se considerarían como posibles limitantes a la productividad

(p.ej. Bertsch 1998).

Otro principio fundamental que rige los estudios de fertilidad de suelos es la conocida como

ley del mínimo de Liebig, que establece que la productividad de un sitio está limitada en primer

lugar por el factor (puede ser la disponibilidad de un nutriente o no) más desfavorable (p.ej.

Bertsch 1998). Esta ley implica que aunque se detecten déficits de nutrientes en algunos suelos

puede que esto no esté afectando negativamente a la productividad del sitio porque haya otros

factores (p.ej. climáticos, selvícolas…) que sean los que estén limitando en mayor grado el

crecimiento. De igual modo, puede que se observen deficiencias de algunos elementos pero no

se mejore la productividad si sólo se actúa para aumentar la disponibilidad de algunos de ellos

mientras otros quedan deficitarios.

Figura 2.1. Esquema de

la estimación de niveles

críticos mediante el

análisis de la relación

entre la disponibilidad

de nutrientes y el

rendimiento productivo

(basado en Bertsch

1998).

24

Los suelos tropicales son con frecuencia considerados como lateríticos, i.e. ácidos y poco

fértiles debido al intenso y prolongado proceso de meteorización y lixiviación que han sufrido.

Sin embargo, la única característica que tienen en común los suelos de zonas tropicales es el

régimen de temperatura relativamente constante a lo largo del año. Por lo demás, las zonas

tropicales presentan mayor o igual diversidad de suelos que las templadas o boreales,

encontrándose todos los órdenes de suelos descritos en la Taxonomía de Suelos (p.ej. Lathwell

y Grove 1986; Sánchez y Logan 1992; Hartemink 2004). No obstante, aunque los suelos

tropicales no sean sólo lateríticos, Ultisoles y Oxisoles son muy abundantes y hay algunas

consideraciones generales que, pese a los problemas que conlleva la generalización, se detallan

a continuación (p.ej. Lathwell y Grove 1986; Sánchez y Logan 1992).

Aunque no todos los suelos tropicales sean lateríticos, el problema de la acidez del suelo es

relativamente común en zonas tropicales y está causado por la toxicidad causada por el Al3+

que

queda disponible para las plantas (p.ej. Molina y Alvarado 2012). Se considera que una tercera

parte de los suelos tropicales tienen problemas severos de toxicidad de Al (Saturación de Al >

60%) mientras que una cuarta parte presenta problemas moderados de acidez (pH < 5,5 pero

Saturación de Al < 60%) (Sánchez y Logan 1992). Al lavarse las bases y bajar el pH por debajo

de 5,5 (i.e. acidificarse el suelo), el Al del suelo se solubiliza y pasa a comportarse como un

catión trivalente en el suelo, acaparando los espacios intercambiables e incluso desplazando a

otros cationes (Ca2+

, Mg2+

, K+ …) que son fácilmente lixiviados a su vez, por lo que sigue

bajando aún más el pH del suelo en un proceso circular que se va intensificando (p.ej. Molina y

Alvarado 2012). Hay tres parámetros para evaluar la acidez del suelo: el pH, la acidez total

[cmol(+) L-1

] y la saturación de acidez [%], siendo ésta última la medida más recomendable al

ser la proporción de la Capacidad de Intercambio de Cationes Efectiva (CICE) que está ocupada

por Al3+

(p.ej. Molina y Alvarado 2012). Este problema puede ser tanto de origen natural como

antrópico, especialmente por el mal uso de fertilizantes (p.ej. Molina y Alvarado 2012). La

práctica más común para solucionar los problemas de acidez del suelo ha sido el encalado, como

se expone posteriormente en el apartado 2.2.5.

25

Otra característica general común a muchos suelos tropicales, con excepción de los que se

encuentran en zonas altas o áridas, es la intensidad de los procesos de mineralización de la

materia orgánica (M.O.). Hasta los años 1970-1980 se pensaba que los suelos tropicales tenían

menores valores de M.O. comprados con los de zonas templadas. Posteriormente se ha

demostrado que la variabilidad del contenido de M.O. es igualmente alta en los suelos tropicales

que en los de zonas templadas sin que haya grandes diferencias en los contenidos medios entre

una y otra región o, de existir, serían ligeramente mayores en zonas tropicales (Sánchez y Logan

1992). Sin embargo, sí que puede generalizarse que las tasas de descomposición y

mineralización de la M.O. en bosques tropicales húmedos (y de zonas bajas) son del orden de 5

veces las medidas en bosques de zonas templadas debido a que temperatura y humedad son

constantemente altas. Esta intensa dinámica de la M.O. se ve compensada en los bosques

tropicales por una igualmente alta tasa de aporte de M.O. al suelo por parte de la vegetación

(Sánchez y Logan 1992).

Por otro lado, aunque el N es comúnmente considerado el nutriente más limitante para la

productividad de muchos ecosistemas terrestres, parece que no es tan crítico para los tropicales

como otros, por ejemplo, el P (Vitousek 1984; Hedin et al. 2009). El problema del P en muchos

suelos es que precipita como fosfato de Fe o de Al en suelos ácidos con abundancia de esos

metales (comunes en suelos tropicales como se ha comentado anteriormente) y como fosfato de

Ca en suelos calcáreos. Además, la fracción orgánica también puede acumular una gran

proporción del P total, la cual no estaría disponible para las plantas (p. ej. Gyaneshwar et al.

2002; Khan et al. 2007). Movilizar este gran porcentaje de P inmovilizado en el suelo sin ser

disponible para las plantas es uno de los retos que se afrontan desde las ciencias agrarias en este

momento, para lo cual se están desarrollando biofertilizantes que consisten en organismos que

consiguen solubilizar y mineralizar el P ya presente en el suelo en lugar de aplicar fertilizantes

químicos que a continuación se precipitarán y serán poco usados por las plantas (p. ej.

Gyaneshwar et al. 2002; Khan et al. 2007). Un efecto parecido al de estos biofertilizantes es el

que producen las micorrizas (p.ej. Alvarado et al. 2004; Corryanti et al. 2007).

26

En definitiva, aunque una gran parte de los suelos tropicales presentan algunos problemas de

fertilidad, ésta no es en general tan baja como se pensaba tradicionalmente (p.ej. Sánchez y

Logan 1992). En ese sentido, Sánchez y Logan (1992) estiman que sólo el 5% de los suelos

tropicales presentan una Capacidad de Intercambio de Cationes Efectiva (CICE) menor a 4 cmol

(+) kg-1

en el horizonte superficial, umbral por debajo del cual se considera que un suelo tiene

una capacidad muy limitada para retener cationes, evitar su lixiviación y permitir que éstos sean

absorbidos por las plantas. Esto se explica en parte porque los valores de M.O. en los suelos

tropicales son en general más altos de lo que se estimaba históricamente, lo cual redunda en una

mayor Capacidad de Intercambio de Cationes (CIC) comparada con lo que se pensaba (p.ej.

Sánchez y Logan 1992). Además, muchos horizontes subsuperficiales de suelos comunes en

regiones tropicales (p.ej. Oxisoles, Ultisoles y Andisoles) tienen una relativamente elevada

Capacidad de Intercambio Aniónico (CIA) que hace que las pérdidas por lixiviación de algunos

aniones como nitratos y sulfatos sean relativamente pequeñas (Sánchez y Logan 1992).

La mineralogía de los suelos tropicales ha sido dividida en cuatro grandes grupos

(kaoliníticos, oxídicos, alofánicos y esmectíticos), de forma similar a la de las regiones

templadas (p.ej. Hartemink 2004). La caracterización de la mineralogía es importante a la hora

de evaluar y gestionar la fertilidad de un suelo, especialmente teniendo en cuenta la carga

variable que predomina en algunos de esos grupos mineralógicos (Bertsch 1998). Sin embargo,

a efectos de gestión práctica, normalmente se estima la mineralogía del suelo en función de su

clasificación taxonómica y de otras de sus propiedades más fácilmente medibles y observables

(p.ej. Bertsch 1998; Hartemink 2004).

2.2.3. Concentración foliar de nutrientes

Aunque la realización de análisis de suelo para evaluar su fertilidad es una práctica muy

común en la edafología, los muestreos para estimar la concentración foliar de nutrientes son

también frecuentes y muchas veces considerados como más efectivos. Éstos reflejan más

fielmente la cantidad de nutrientes que está aprovechando la planta, teniendo en cuenta a la vez

la disponibilidad de ellos en el suelo y la capacidad de la planta para absorberlos en función de

27

otros factores (ambientales, selvícolas …) (p.ej. Mead 1984; Dreschel y Zech 1991; West 2006).

El diagnóstico nutricional de sistemas forestales se ha hecho a menudo en base a la observación

de síntomas visuales de algunas deficiencias; sin embargo, el análisis en laboratorio de muestras

foliares presenta ciertas ventajas a tener en cuenta: (a) los síntomas visuales de deficiencia foliar

se pueden confundir con relativa facilidad entre ellos y con síntomas de otros problemas; (b) el

análisis de laboratorio nos permite detectar valores bajos de algunos nutrientes que puede que

no sean tan bajos como para producir síntomas visuales, lo cual hace que el gestor sea capaz de

solucionar el problema antes de que sea grave.

La teoría de los niveles críticos descrita anteriormente para la interpretación de los resultados

de los análisis de disponibilidad de nutrientes en el suelo se aplica de la misma manera para

evaluar si la concentración foliar de nutrientes es adecuada o muestra deficiencias (Richards y

Bevege 1972). Así, se delimitan cuatro rangos (Figura 2.2) en función de la relación entre el

incremento en la concentración foliar de un elemento y el incremento correspondiente en el

crecimiento y/o en la cosecha (deficiencia, crítico, consumo de lujo y toxicidad), siendo el punto

crítico el punto de inflexión de la curva que marca el límite entre el rango de deficiencia y el

rango crítico (p.ej. Richards y Bevege 1972; Blinn y Buckmer 1989; Bertsch 1998).

Figura 2.2. Esquema

de la relación teórica

entre el contenido

foliar de nutrientes

de una planta y su

crecimiento

(modificado de Blinn

y Buckmer 1989)

28

De forma similar a lo comentado anteriormente sobre la fertilidad del suelo, se define el

rango de deficiencia como aquel en el que un pequeño aumento en la concentración del

nutriente se traduce en un gran incremento en el rendimiento (Figura 2.2). El rango crítico, es

aquel en el que un incremento en la disponibilidad del elemento se traduce en un incremento en

el rendimiento pero poco intenso (<10%) (Figura 2.2). Por el contrario, el rango de consumo de

lujo es aquel en el que aunque se aumente la disponibilidad del elemento no aumenta el

rendimiento y el de toxicidad, aquel en el que un aumento en la concentración del elemento

provoca una disminución en el rendimiento (Figura 2.2) (p.ej. Richards y Bevege 1972; Blinn y

Buckmer 1989; Bertsch 1998).

Dreschel y Zech (1991) muestran algunos valores de concentración foliar de ciertos

nutrientes (N, P, S, K, Ca, Mg, Al, Fe, Mn, Zn, Cu, Mo, B) que son considerados como

deficientes, adecuados y excesivos para alrededor de 40 de las principales especies usadas para

plantaciones forestales en los trópicos y subtrópicos.

2.2.4. Empobrecimiento de los suelos por salida de nutrientes en la madera

El empobrecimiento de los suelos es un proceso que ocurre de forma natural en los

ecosistemas, especialmente en los tropicales lluviosos, como parte del proceso natural de

pedogénesis (p.ej. Buol et al. 1989; Schaetzl y Anderson 2005). Bruijnzeel (1991) sintetiza un

balance de nutrientes en ecosistemas forestales tropicales de zonas bajas basado en las entradas

por deposición atmosférica (incluyendo el agua de lluvia) y las pérdidas hidrológicas por

lixiviación y drenaje, sin tener en cuenta la extracción de madera ni otros factores como la

meteorización, la fijación biológica, la erosión, la sedimentación o la volatilización (Figura 2.3).

Las entradas Ea consideradas por Bruijnzeel (1991) son elevadas en el caso de N y Ca y

normales en el caso de Mg y K, si se comparan con los valores proporcionados por otros autores

(p.ej. Fölster y Khanna 1997 Alvarado 2012 c) (Tabla 2.1, Figura 2.3).

29

Tabla 2.1. Entrada de nutrientes por deposición atmosférica en ecosistemas tropicales según revisión de

literatura (Fölster y Khanna 1997; Alvarado 2012 c)

Fölster y Khanna (1997) Alvarado (2012 c) Agrupados

kg ha-1 yr-1 kg ha-1 yr-1 kg ha-1 yr-1 kg ha-1 (20 años)*

N 2-10 5-21 2-21 230 (40-420)

Ca 2-8 1.4-34 1.4-34 354 (28-680)

K 2-12 2.5-24 2-24 260 (40-480)

Mg 1-5 1.6-26 1-26 270 (20-520)

P ---- 0.2-1.1 0.2-1.1 13 (4-22)

* Estimación de la entrada de nutrientes al sistema a lo largo de un período de 20 años, considerado como

un turno o período de rotación común en plantaciones forestales en regiones tropicales

Figura 2.3. Balance natural de nutrientes (Bnat) entre las entradas por deposición atmosférica (Ea) y las

salidas por drenaje y lixiviación (Sd) en bosques tropicales con suelos de diferente fertilidad según la

revisión de literatura de Bruijnzeel (1991).

30

La mayor concentración de nutrientes en el árbol se da generalmente en las hojas, donde se

considera que se acumula entre el 20 y el 40% del total; mientras que la concentración en el

tronco suele ser baja (p.ej. Miller 1984, 1995). Sin embargo, la gran cantidad de biomasa que se

acumula en los troncos los convierte en un gran depósito de nutrientes, por lo que la extracción

de madera supone una salida de éstos del sistema, lo que está considerado como una de las

principales causas del empobrecimiento en suelos forestales (p.ej. Miller 1984; Fölster y

Khanna 1997; Worrel y Hampson 1997).

Como se ha comentado anteriormente, este proceso ha preocupado a científicos y gestores

desde hace varios siglos, ya que se considera que la salida periódica de nutrientes del sistema

podría poner en riesgo la productividad de futuras rotaciones si no se reponen suficientemente

(p.ej. Fölster y Khanna 1997; Merino et al. 2005). Evans (2009) señala que a largo de la historia

de los bosques centroeuropeos algunos gestores han visto como se reducía la productividad

entre sucesivas rotaciones, lo cual podría asociarse con este fenómeno, que de hecho ha sido

observado por otros autores en la actualidad (p. ej. Fölster y Khanna 1997). Aunque muchos

autores recomiendan la aplicación de fertilizantes y enmiendas para compensar ésta salida de

nutrientes del sistema (p. ej. Rennie 1955; Worrel y Hampson 1997; Merino et al. 2005), una

gran mayoría de los gestores forestales han ignorado este problema (Fölster y Khanna 1997).

Por otro lado, el uso continuado de fertilizante para compensar la extracción de madera puede

también ser juzgado como insostenible, dados los impactos que tiene: contaminación, uso de

energía, minería, etc. (Worrel y Hampson 1997).

La corteza presenta concentraciones moderadamente altas y numerosos autores han

propuesto el descortezado en campo como una práctica eficiente para evitar la degradación de

los suelos (p.ej. Fölster y Khanna 1997; Yamada et al. 2004; Ma et al. 2007) aunque, dados los

altos costes que supone y la dependencia de maquinaría especializada, son prácticas que no

están generalizadas en plantaciones del mundo. Pese a estas altas concentraciones de nutrientes

en la corteza, su biomasa es muy pequeña comparada con la de la madera en plantaciones

maduras y por lo tanto su importancia relativa en el total de nutrientes que se extrae queda

diluida. Sin embargo, la proporción de corteza en los troncos jóvenes es mayor y, por lo tanto, la

31

extracción de árboles individuales en las primeras claras puede suponer también un problema;

aunque la biomasa de estos árboles suele ser más pequeña y por lo tanto el problema es de

menor magnitud que en los aprovechamientos al final del turno.

El empobrecimiento de los suelos se produce de forma gradual tras varias rotaciones según

van saliendo nutrientes del sistema. La velocidad de este proceso depende de la intensidad del

aprovechamiento que se realice. Así, cosechar árboles enteros (incluyendo ramas y follaje)

supone una mayor salida de nutrientes del sistema que la extracción de árboles con corteza y, al

contrario, los árboles descortezados en campo suponen una mucho menor salida de nutrientes

que los dos casos anteriores, como ya se ha mencionado (p.ej. Fölster y Khanna 1997; Worrel y

Hampson 1997; Merino et al. 2005). Además de la cosecha de madera, la extracción de la capa

superficial orgánica del suelo para usarla como combustible fue identificada como una

importante salida de nutrientes del sistema (Binkley 1986). En ese sentido, la extracción de

frutos, corteza y otros productos forestales no maderables podría tener también un impacto

sobre la reserva de nutrientes del suelo que podría llegar a empobrecerlo.

Worrel y Hampson (1997) distinguen dos escuelas de pensamiento a este respecto: (a) la

escuela “optimista”, que interpreta que la entrada de nutrientes al sistema (por precipitación y

deposición atmosférica) es mayor que la salida por extracción de madera, concluyendo que no

existe peligro de empobrecimiento de la reserva de nutrientes del suelo; y (b) la escuela

“pesimista”, que expone que la salida de nutrientes del sistema por lixiviación y drenaje de agua

en el suelo unida a la extracción de nutrientes en la madera es mayor que la entrada natural,

advirtiendo del riesgo potencial de empobrecimiento del suelo. Los mismos autores advierten

que, pese a que la escuela “optimista” admita que haya pérdida neta de nutrientes del sistema,

ésta argumenta que el empobrecimiento de los suelos tardaría muchas rotaciones en

manifestarse; mientras que ellos encuentran evidencias que indican que esto no siempre se

cumple y se posicionan en la línea de la considerada escuela “pesimista” (para una revisión

consultar Worrel y Hampson 1997).

Teniendo en cuenta turnos de corta de 70-100 años comunes en aprovechamientos forestales

en zonas templadas, el impacto de la extracción de nutrientes a través de la madera podía no ser

32

considerado un problema de sostenibilidad (Hornbeck et al. 1986). Sin embargo, la cosecha de

árboles enteros incluyendo lo que antes se consideraban residuos en los aprovechamientos

forestales (ramas, troncos con pequeño diámetro, follaje y corteza) para su aprovechamiento

bioenergético, está produciendo problemas ambientales y de sostenibilidad económica

especialmente en zonas templadas y boreales donde el uso de la biomasa como energía

renovable está adquiriendo un papel importante en los últimos años (Jacobson et al. 2000;

Lattimore et al. 2009; Egnell 2011; Helmisaari et al. 2011). De igual manera, las plantaciones

forestales establecidas en áreas tropicales presentan generalmente crecimientos mucho más

elevados (hasta 90 m3 ha

-1 año

-1), lo que lleva a que se gestionen en muchos casos con turnos

muy cortos, de incluso 6-8 años para algunas frondosas (Evans y Turnbull 2004), suponiendo

así un problema de sostenibilidad si no se toman medidas para compensar las salidas de

nutrientes del sistema (p. ej. Fölster y Khanna 1997).

2.2.5. Aplicación de enmiendas y fertilizantes

Aunque frecuentemente se usan indistintamente, los conceptos de “enmienda” y

“fertilizante” están relativamente bien diferenciados. Así, las enmiendas (en general calizas u

orgánicas) se aplican para mejorar las propiedades generales del suelo, mientras que los

fertilizantes se orientan a cubrir las necesidades o los problemas nutricionales de la planta y

pueden aplicarse a través del suelo, el agua y/o las aspersiones foliares. Como ejemplo, es

común encontrar recomendaciones de enmiendas en suelos forestales con unidades del orden de

t ha-1

, mientras que las de fertilización suelen hacerse en unidades de g árbol-1

(Bertsch 1998;

Alvarado 2012 d).

Las enmiendas orgánicas son una práctica milenaria en agricultura y presentan una serie de

ventajas para el suelo y el crecimiento de las plantas: (a) mejoran la estructura y propiedades

físicas del suelo; (b) mejoran la fertilidad del suelo contrarrestando en parte la acidez y

aumentando su capacidad de intercambio catiónico y aniónico; y (c) aunque la concentración de

nutrientes que se incorporan sea habitualmente baja, aportan una gran variedad de ellos que no

son aplicados comúnmente de otra manera (p.ej. micronutrientes).

33

Las enmiendas calizas (encalado) constituyen, junto con la elección de especies tolerantes, la

práctica más apropiada y económica para corregir los problemas de acidez del suelo, muy

comunes en suelos tropicales como se comentaba anteriormente (Molina y Alvarado 2012).

Molina y Alvarado (2012) exponen una buena revisión de la gestión de los problemas de acidez

y las prácticas de encalado en suelos forestales en el trópico.

Las enmiendas y la fertilización pueden incrementar el crecimiento y la productividad de los

sistemas forestales, mejorar la sanidad de las masas (i.e. resistencia a plagas y enfermedades) e

incluso mejorar la calidad de la madera, además de compensar en parte la salida de nutrientes

del sistema por la extracción de madera mencionada con anterioridad (p.ej. Molina 2012). La

aplicación de enmiendas y fertilizantes a sistemas forestales ha sido analizada en profundidad

por varios autores como, por ejemplo, Miller (1981, 1984), Ballard (1984), Binkley (1986),

Gonçalves et al. (1997), Alvarado (2012 d), Molina (2012) y Molina y Alvarado (2012).

Miller (1981) desarrolló una teoría respecto a la fertilización forestal que ya se ha convertido

en clásica y que se basa en tres conceptos: (1) La fertilización beneficia a los árboles, no al

sitio. Tras la aplicación del fertilizante, los nutrientes se distribuyen de forma relativamente

rápida dentro de los distintos componentes del ecosistema excepto una pequeña parte que se

pierde por lixiviación y otras pérdidas (p.ej. volatilización). Así, los árboles aumentan la

cantidad de nutrientes acumulados y puede ser que los nutrientes aplicados sean suficientes para

que éstos mantengan un crecimiento mayor incluso después de cesar en la aplicación de

fertilizante. No obstante, se si compara la cantidad de nutrientes aplicada siguiendo una

recomendación habitual con los nutrientes que componen la reserva capital en el suelo, ésta

primera es tan baja que impide que se pueda considerar como una mejora sustancial del sitio. (2)

La respuesta a la fertilización es principalmente una disminución del turno. Como se muestra

en la Figura 2.4, los árboles fertilizados crecen más inicialmente y luego vuelven normalmente a

seguir la curva de crecimiento correspondiente pero cumpliendo con ella en función de su

estado de desarrollo (i.e. tamaño) en lugar de ser en función de su edad; i.e. los árboles

fertilizados estarán más avanzados en que los árboles no fertilizados. Así, en la Figura 2.4, los

puntos fertilizados x o y volverán a la curva de crecimiento en los puntos x’ e y’,

34

respectivamente, mientras que los árboles no fertilizados habrían alcanzado los puntos x’’ e y’’,

correspondientes a la edad de los rodales según la curva de crecimiento normal. Esto supone que

en edades maduras, una hipotética respuesta a la fertilización podría incluso disminuir el

crecimiento de los árboles. (3) Los requerimientos de fertilizantes varían en función del estado

de desarrollo. En la Figura 2.5 se pueden distinguir tres etapas (joven, intermedia y madura) en

función de su estado de desarrollo y sus requerimientos nutricionales. En la primera etapa, hasta

el cierre de copas, los árboles tienen altos requerimientos nutricionales para mantener tasas de

crecimiento elevadas y llegar a establecerse con éxito. Es en esta primera etapa donde la

nutrición es más crítica y por lo tanto la fertilización forestal tiene grandes probabilidades de

éxito. En la segunda etapa, siempre y cuando las necesidades nutricionales se hayan cumplido

suficientemente durante la primera, los árboles cubrirán sus requerimientos nutricionales con el

reciclaje de nutrientes (interno dentro del árbol o externo a través de la hojarasca y el suelo) y

probablemente no haya respuesta a la fertilización. No obstante, la aplicación de claras en ésta

época hace retroceder a los rodales a un estado equivalente a la etapa joven y se produciría una

respuesta a la fertilización con una rápida recuperación de la biomasa foliar del rodal y una

mejora en el crecimiento del mismo. En una hipotética tercera etapa, cuando ya los rodales son

muy maduros, la ralentización de los procesos de mineralización y la consecuente

inmovilización de N en la materia orgánica del suelo podrían causar deficiencias de N que se

podrían traducir en una respuesta a la fertilización (Miller 1981).

Figura 2.4. Curva de

crecimiento que ilustra como los

árboles fertilizados en los puntos

x e y vuelven a comportarse

según la curva correspondiente

incorporándose a la misma en

los puntos x’ e y’, más

avanzados a los

correspondientes x’’ e y’’ que

alcanzan los árboles de la misma

edad que no han sido fertilizados

(modificado de Miller 1981)

35

Figura 2.5. Las tres etapas

marcadas a lo largo del turno en

un rodal según su estado

nutricional según la teoría de

Miller (modificado de Miller

1981)

Siguiendo los dos primeros conceptos propuestos por Miller (1981), especialmente el

segundo, la fertilización forestal no podría hacer aumentar el índice de sitio de un rodal ni la

productividad en términos absolutos (p.ej. volumen maderable total). El efecto de la

fertilización sólo sería entonces un aumento del rendimiento económico que supone tener la

misma producción en menor tiempo. No obstante, Binkley (1986) muestra como en algunos

casos la aplicación de fertilizantes logra solucionar los problemas de fertilidad de un rodal de

forma más eficiente y conseguir un incremento en la producción más estable temporalmente,

equivalente a un incremento en el índice de sitio (Figura 2.6). En ese sentido, Alvarado y

Herrera (2012) consideran que cuando se evalúa la productividad de una parcela para el

establecimiento de una nueva plantación, se puede estimar razonablemente que ésta tiene por

ejemplo una productividad de media pero es mejorable a clase buena si se solucionan ciertos

problemas.

Además de los efectos positivos que la fertilización tiene a menudo sobre los sistemas

forestales comentados anteriormente, Binkley (1986) hace hincapié en que la respuesta

económica a la fertilización suele ser incluso mejor que la respuesta productiva. Esto se explica

porque, además de conseguir un aumento en la productividad, normalmente se consiguen

árboles con mayor diámetro los cuales se encuentran en categorías de mayor precio en el

mercado (Binkley 1986).

36

Figura 2.6. Curvas de crecimiento que ilustran como la fertilización puede suponer un adelanto temporal

de la dinámica de crecimiento en un rodal [según lo también propuesto por Miller (1991) y expuesto en la

Figura 2.4] y también un incremento más estable de la calidad de estación y la productividad de un rodal,

equivalente a un incremento del índice de sitio (tomado de Binkley 1986)

En cuanto a las dosis de fertilizantes y enmiendas aplicadas en sistemas forestales, cabe

destacar que aunque son parecidas a las usadas en agricultura, sólo se realizan una o unas pocas

veces a lo largo del turno del sistema lo que resulta en dosis medias de aplicación mucho

menores (p.ej. Smethurst 2010). Serrada (2008) recomienda en general dosis de cal común

(CO3Ca) entre 5 y 7 t ha-1

para subir el pH de un suelo aproximadamente de 5 a 6 en sistemas

forestales españoles, mientras que Molina y Alvarado (2012), para sistemas forestales

tropicales, consideran la aplicación de un rango más amplio (entre 1 y 10 t ha-1

de cal común o

de dolomita) teniendo en cuenta la tolerancia de la especie y las propiedades del suelo, teniendo

especial cuidado de no usar dosis excesivas para evitar problemas de sobre encalado. Además,

Molina y Alvarado (2012) también contemplan el uso de otras fuentes como, por ejemplo el

yeso, del que recomiendan dosis de hasta 2 t ha-1

.

Las dosis y las fuentes de fertilizantes usadas comúnmente son muy variables aunque, en

términos generales, varían entre 50-300 g árbol-1

de N-P-K (o una fórmula completa) (p.ej.

37

Alvarado 2012 d), y o 50-500 kg ha-1

de N-P-K (Evans y Turnbull 2004; Smethurst 2010).

Smethurst (2010) resume esta gran variabilidad mediante el empleo de tres sistemas que sirven

como ejemplo de algunos de los programas de fertilización que se usan de manera habitual en

algunos sistemas forestales alrededor del mundo: (a) En coníferas de crecimiento lento (p.ej. en

Europa) se fertiliza durante el establecimiento con fórmulas variadas (normalmente N-P-K) y

posteriormente se aplican dosis de 150 kg N ha-1

cada 5 años (o nunca en mucho casos); (b) en

coníferas de crecimiento medio (p.ej. en Australia y SE de EEUU) se aplican 10-50 kg ha-1

de N

y de P mediante fórmulas variadas tipo N-P-K durante el establecimiento y, posteriormente,

208-324 kg N ha-1

y 50-112 kg P ha-1

cada 5-15 años; y (c) en eucaliptos de crecimiento rápido

(p.ej. en Brasil), con turnos de unos 6 años, reciben 2 ó 3 aplicaciones de 20 kg N ha-1

y 53 kg P

ha-1

durante los dos primeros años además de dosis variables de K y de B (Smethurst 2010).

La aplicación de enmiendas y fertilizantes son prácticas selvícolas clasificadas como de

cuidados culturales o tratamientos parciales sobre el suelo según Serrada (2008). Sin embargo,

con frecuencia se planifican estas prácticas sin tener en cuenta las interacciones que tienen con

otros cuidados culturales como pueden ser las claras, los clareos, las limpias etc. Estas otras

prácticas afectan a la respuesta a la fertilización como se detalla en el apartado siguiente (2.2.6).

Por otro lado, la respuesta a la fertilización también está condicionada a que los rodales tratados

tengan problemas nutricionales que sean los que realmente estén condicionando su

productividad, según la Ley del mínimo de Liebig. En ese sentido, el déficit hídrico puede ser el

mayor condicionante de la productividad en algunos ambientes mediterráneos (p.ej. Serrada

2008; Montes et al. 2012), mientras que otros aspectos, como la profundidad efectiva del suelo,

la genética de las plantas, la temperatura y/o la radiación solar, entre otros, pueden estar

condicionando la respuesta a la fertilización de un sistema.

2.2.6. Efectos de otras prácticas selvícolas sobre la nutrición y fertilización forestal

Aunque la aplicación de enmiendas y fertilizantes son las prácticas selvícolas que tienen un

efecto más directo sobre la nutrición forestal, hay otras prácticas que también la afectan

indirectamente y además condicionan la respuesta de los sistemas forestales a la fertilización.

38

El control de la competencia herbácea (también llamado control de malezas) juega un doble

papel en la dinámica nutricional de los bosques plantados. Por un lado, la competencia de esta

vegetación puede causar deficiencias nutricionales (entre otros problemas) en las plantas de la

especie principal objeto del aprovechamiento forestal. Además, si la competencia que ejerce el

sotobosque es muy fuerte el crecimiento de la especie principal se verá perjudicado y se

impedirá la respuesta a la fertilización incluso cuando la nutrición sea deficiente (p.ej. Kumar

2011). Por otro lado, tener en cuenta la nutrición de la vegetación acompañante puede presentar

ciertas ventajas. Binkley (1986) señala que el aumento de la producción de biomasa del

sotobosque puede tomarse en cuenta como una mayor producción de pasto si se complementa la

producción forestal con la ganadera. Además, el sotobosque de los bosques plantados absorbe

grandes cantidades de N (y quizás otros nutrientes) del suelo en las etapas iniciales de las

plantaciones, cuando este elemento es más abundante y corre el riesgo de lixiviarse en

abundancia. Así, las altas concentraciones de N que se encuentran a menudo en esta vegetación

acompañante (que de hecho a veces está compuesta por leguminosas fijadoras de N) pueden

tomarse en cuenta como una reserva de N en el sistema que pueden ser gestionadas durante los

siguientes años para favorecer a la especie principal (Smethurst y Nambiar 1989; Lugo 1992;

Woods et al. 1992; Fölster y Khanna 1997; Turner y Lambert 2008).

Además de lo que se ha comentado con anterioridad acerca del empobrecimiento de los

suelos por la salida de nutrientes causada por la extracción de madera, las actividades de

aprovechamiento forestal y la extracción de la madera del campo generan varios problemas de

pérdida de fertilidad del suelo, entre otros: (1) pérdida de nutrientes por la erosión causada por

el aumento de la escorrentía al perder la vegetación protectora; (2) aumento de la lixiviación de

nutrientes al disminuir la absorción por la vegetación; y (3) remoción en algunos casos de parte

del horizonte superficial del suelo (a veces el orgánico), en general rico en nutrientes (p.ej.

Binkley 1986; Worrel y Hampson 1997; Alvarado 2012 c).

Finalmente, la densidad del rodal es un factor fundamental a la hora de analizar la nutrición

forestal y planificar actividades de enmiendas y fertilización. Por un lado, como ocurre en

general en selvicultura, el análisis de la densidad del rodal permite relacionar los valores de

39

árboles individuales con los de la masa. Así, las mediciones de nutrientes acumulados o

absorbidos por varios árboles o sus requerimientos nutricionales pueden ser extrapoladas a

valores medios del rodal. La densidad de individuos influye directamente en la nutrición de los

árboles mediante la competencia que se establece entre árboles adyacentes por la absorción de

los nutrientes disponibles en el suelo, teniendo en cuenta que los árboles dominantes tendrán en

general sistemas radicales más desarrollados y también competirán con ventaja en ese sentido

con los dominados.

En ese sentido, los tratamientos selvícolas que modifican la densidad del rodal (clareos y

claras en España, raleos en América Central) tienen varios efectos sobre la nutrición y/o sobre la

respuesta a la fertilización de los rodales. Por un lado, al eliminar algunos individuos, los que

quedan en pie tienen menos competencia por los nutrientes del suelo y, además, los residuos de

corta que se dejan en el sitio suponen un aporte de nutrientes importante con la ventaja de ser

más o menos prolongado en el tiempo en función de la velocidad de los procesos de

descomposición y mineralización (p.ej. Alvarado 2012 c). Por otro lado, la combinación de la

programación de claras y clareos con la aplicación de fertilizantes ha sido observada por

numerosos autores (p.ej. Binkley 1986; Folster y Khanna 1997). Esto se atribuye a que si el

dosel de copas está totalmente desarrollado y se abre mediante una clara, los árboles pueden

aprovecharse de los nutrientes aportados por la fertilización para cerrar rápidamente el dosel

aumentando su índice de área foliar y por lo tanto creciendo más que si no hubiesen recibido la

fertilización (p.ej. Binkley 1986). Esto es equivalente a pasar de la etapa 2 descrita por Miller

(1981) como intermedia y sin respuesta a la fertilización a la etapa 1 (Figura 2.5) en la que no se

ha producido todavía el cierre de copas y es probable la respuesta a la fertilización. Al contrario,

la aplicación de fertilizantes cuando el dosel de copas está totalmente desarrollado y no se

realizan claras ni clareos difícilmente puede tener un efecto positivo sobre el crecimiento de los

árboles ya que éstos no tienen forma de expandirse y poder usar esos nutrientes que aporta la

fertilización, los cuáles acaban generalmente lixiviándose.

40

2.3. Plantaciones de teca (Tectona grandis L.f.)

Originaria del sudeste asiático (India, Laos, Myanmar y Tailandia), la teca (Tectona grandis

L.f.) ha sido ampliamente establecida en regiones tropicales alrededor del mundo, especialmente

desde la década de 1980, cuando la superficie de bosques naturales de teca empezó a decaer.

Aunque tradicionalmente fue considerada como una madera preciosa en los países donde es

originaria, durante las últimas décadas se ha convertido en una importante especie en el sector

internacional de las maderas tropicales duras de buena calidad (e.g. Pandey y Brown 2000;

Kollert y Cherubini 2012). Basándose en la revisión de varios trabajos anteriores, Kollert y

Cherubini (2012) estiman que entre 1975 y 2000 la superficie de bosques naturales de teca se

mantuvo en torno a los 20 ·106 ha, mientras que el área mundial de plantaciones de la especie

aumentó desde 1,3 a 5,7 ·106 ha. Myanmar, India e Indonesia eran los principales países

productores de teca, aunque durante este período creció la importancia de varios países

africanos (Nigeria, Costa de Marfil y Ghana) y centroamericanos (Costa Rica, Panamá, El

Salvador y Trinidad y Tobago).

En un estudio más detallado, Kollert y Cherubini (2012) reportan la existencia en 2010 de 29

·106 ha de bosques naturales de teca, lo que coincide con lo expuesto por Tewari (1992, citado

por Kollert y Cherubini 2012) para el período 1976-1979 y 4,3 ·106 ha de bosques plantados de

teca, de los cuales el 83% están en Asia, el 11% en África, el 6% en América y menos del 1%

en Oceanía. India, Indonesia y Myanmar siguen siendo los principales países en superficie

ocupada por bosques de teca, con un 76% del total entre los tres. Los países tropicales de

América han crecido en su importancia relativa en el sector de la teca, con una superficie total

estimada de 270.000 ha, entre los que dominan Brasil (con 65.000 ha, siendo el octavo país del

mundo en superficie de plantaciones de teca, Figura 2.7), Panamá (55.000 ha), Ecuador (45.000

ha), Costa Rica (31.500 ha) y Guatemala (28.000 ha). Cabe destacar que en el período entre

1975 y 2000 la teca no era de especial relevancia en países de América del Sur como Brasil,

Ecuador y Colombia, en los que, sin embargo, está cobrando una gran importancia en los

últimos años, como muestran las estadísticas de 2010 (Kollert y Cherubini 2012).

41

En América Central las plantaciones de teca cubren una superficie de 132.770 ha (3% del

total mundial) principalmente en Panamá, Costa Rica y Guatemala (55.000, 31.500 y 28.000 ha,

respectivamente) aunque también ha sido introducida en El Salvador, Nicaragua, Honduras y

Belize (9.760, 7.960, 450 y 100 ha, respectivamente) (Kollert y Cherubini 2012). A pesar de la

pequeña importancia relativa de las plantaciones de teca de América Central en el sector

mundial, los bosques plantados de teca sí que tienen una gran importancia en los países

centroamericanos en general, debido a su pequeña superficie. Así, Panamá es el tercer país del

mundo con mayor área plantada con teca relativa a su superficie, Costa Rica es el quinto y El

Salvador el noveno (Figura 2.8) (Kollert y Cherubini 2012).

Nieuwenhuyse et al. (2000) estiman que se podría optimizar la producción agraria de una

región de las llanuras caribeñas de Costa Rica si aproximadamente el 70% del paisaje estuviese

cubierto con plantaciones de teca. Estos autores aseveran que la producción de una madera de

calidad como la de la teca es más rentable que la de una madera como la de Gmelina arborea

(con un turno más corto pero con madera de peor calidad, i.e. menor precio) o que otros usos

agropecuarios, como los granos básicos o la ganadería extensiva. Esta alta rentabilidad

observada en las plantaciones de teca de América Central (Nieuwenhuyse et al. 2000; Pandey y

Brown 2000; De Camino et al. 2002; De Camino y Morales 2013) es la que ha llevado a la

región a convertirse en una de las zonas donde se han observado mayores incrementos en la

superficie plantada con esta especie. Así, las plantaciones de teca cubrían en el año 2010 16

veces lo que cubrían en el año 2005 en Guatemala, 14 veces en Panamá, 13 veces en Nicargaua,

5 veces en el El Salvador y 2 veces en Costa Rica (Figura 2.9) (Kollert y Cherubini 2012). Esta

alta rentabilidad de las plantaciones de teca también explica cómo esta especie ha sido plantada

por grandes empresas multinacionales que gestionan grandes plantaciones, pero también por

multitud de pequeños propietarios (De Camino y Morales 2013).

42

Figura 2.7. Países con mayor

superficie de plantaciones de

teca (Tectona grandis L.f.) del

mundo (modificado de Kollert y

Cherubini 2012)

Figura 2.8. Países con mayor

superficie de plantaciones de

teca (Tectona grandis L.f.) del

mundo en relación con la

superficie del país (modificado

de Kollert y Cherubini 2012)

Figura 2.9. Incremento en la superficie de plantaciones de teca (Tectona grandis L.f.) entre 1995 y 2010

(modificado de Kollert y Cherubini 2012)

La gran variedad de sitios en los que se ha establecido la teca hace que sea difícil establecer

generalidades al respecto del turno (entre 4 y 80 años) y el incremento anual medio esperado

para la especie (entre 2 y 30 m3 ha

-1 año

-1) (Tabla 2.2) (Kollert y Cherubini 2012). Turnos entre

20 y 30 años se consideran normales dado que este es el periodo de máxima producción. No

obstante, la gestión forestal con estos turnos relativamente cortos genera productos de pequeñas

dimensiones que se consideran de calidad media o baja, no válida para todos los usos que

requiere el mercado. Por el contrario, la madera de grandes dimensiones considerada de buena

calidad es muy apreciada en el mercado internacional, pero requiere rotaciones más largas,

1667

1269

390 214 146 128 73 65 55 45

0

500

1000

1500

2000

1.0

00

ha

45

21.8 16.4 14.1 12.6 12

4.9 4.9 4.5 3.1 2.8 2.2 1.9 1.8 1.7

0

10

20

30

40

50

Fact

or

de

crec

imie

nto

17.5

9.4 7.4 7 6.2 5.9 5.6 5.6 4.7 4.1

0

5

10

15

20

‰

43

como las de los bosques naturales en India que se gestionan con turnos de entre 50 y 150 años

(Kollert y Cherubini 2012).

De manera similar, el incremento anual medio esperado para la especie es muy variable,

oscilando entre 2 y 30 m3 ha

-1 año

-1 (Tabla 2.2) (Kollert y Cherubini 2012). Durante las últimas

décadas se ha generado una polémica con respecto a las tasas de crecimiento de las plantaciones

de teca ya que algunas empresas han llegado a describir tasas de crecimiento de hasta 50 y 80

m3 ha

-1 año

-1 a inversores que luego se encontraban con tasas de crecimiento reales mucho

menores, con los consecuentes problemas asociados (p.ej. Kollert y Cherubini 2012; De Camino

y Morales 2013). Kollert y Cherubini (2012) consideran que si se realiza una buena gestión en

un buen sitio se pueden alcanzar producciones de hasta 15 m3 ha

-1 año

-1 en turnos de 20-25 años,

lo que coincide aproximadamente con lo estimado por De Camino et al. (2002) aunque es un

poco más alto que lo observado por varios autores en América Central, que consiste en 100 –

150 m3 ha

-1 de volumen comercial en la corta final aproximadamente a los 20 años, a lo que hay

que sumar las claras. Como resumen, Kollert y Cherubini (2012) señalan que la mayoría de las

plantaciones de teca tienen una productividad baja, probablemente menor a 5 m3 ha

-1 año

-1, con

la excepción de algunas en sitios buenos y una gestión de intensidad media-alta. Los mismos

autores también hacen hincapié en que una productividad de 5 m3 ha

-1 año

-1 puede ser

considerada media-alta para los objetivos de una plantación de un pequeño propietario o de

plantaciones comunitarias pero que, sin embargo, será considerada como muy baja para una

gran plantación de inversores internacionales.

Bermejo et al. (2004) caracterizaron el crecimiento y la productividad de las plantaciones de

teca en Costa Rica y establecieron tres clases de sitio (19, 21 y 23) en base a la altura dominante

a los 10 años de edad. Para esta Tesis se ha ampliado el número de clases de sitio a cinco:

menor de 18, 19, 21, 23 y mayor de 24. Los modelos de crecimiento establecidos por Bermejo

et al. (2004) muestran tasas de crecimiento altas considerando las otras plantaciones de teca en

la región, ya que están en general por encima de la clase considerada como de calidad I por

Miller en Trinidad (Miller 1969, citado por Bermejo et al. 2004).

44

Tabla 2.2. Incremento Medio Anual (IMA) y turno o período de rotación en plantaciones de teca según la

región geográfica. Entre paréntesis se indica el número de países de los que se obtuvo información

(modificado de Kollert y Cherubini 2012)

Región IMA (m3 ha-1 año-1) Turno o rotación (años)

Min Max Min Max

Africa (7) 3 21 4 60

Asia (5) 2 14 20 80

Oceanía (2) 5 12 20 30

América del Sur (4) 10 27 20 30

América Central (5) 5 30 6 30

Caribe (3) 3 12 20 65

Mundial (26) 2 30 4 80

Torres et al. (2012) proponen otras curvas de índice de sitio para Colombia que muestran

crecimientos ligeramente inferiores a los encontrados por Bermejo et al. (2004) en Costa Rica,

mayores que los encontrados por Henao (1982, citado por Torres et al. (2012) y similares a los

de Keogh (1980, citado por Torres et al. (2012). Las curvas de crecimiento en altura dominante

de Torres et al. (2012) también son comparadas con las establecidas para la India y muestran un

rápido crecimiento en los primeros años en Colombia que luego disminuye y llega a ser menor

que el crecimiento estimado para la India en turnos más largos.

Además de los trabajos mencionados, varios autores han realizado trabajos y modelos

estadísticos para ser usados como herramientas en la gestión de plantaciones de teca teniendo en

cuenta aspectos selvícolas (p.ej. Pérez y Kanninen 2005 a, b; Adu-Bredu 2008) o genéticos

(p.ej. Jayaraman y Bhat 2011; Murillo et al. 2013). Por otro lado, se han realizado manuales de

gestión (Fonseca González 2004) y libros y trabajos de revisión que detallan cuestiones

generales de la gestión de estos sistemas (p.ej. Pandey y Brown 2000; De Camino et al. 2002;

Jayaraman y Bhat 2011; Kollert y Cherubini 2012; De Camino y Morales 2013). Además, se ha

generado TEAKNET (www.teaknet.org) como una red que pone en contacto a investigadores,

gestores, comerciantes, propietarios y cualquiera interesado en el sector de la teca mediante una

comunidad virtual organizada por su página web y, además, organiza conferencias periódicas

que sirven para actualizar conocimientos y contactos.

45

2.4. Suelo y nutrición de plantaciones de teca

Pese a la relevancia de la teca en el sector forestal internacional y a la importancia de la

fertilidad del suelo y la nutrición en la productividad de las plantaciones, la investigación en este

tema es relativamente escasa (Kumar 2011). No obstante, hay algunos autores que han

presentado estudios de la concentración y acumulación de nutrientes en varios tejidos (p.ej.

Kaul et al. 1979; Nwoboshi 1984; Ola-Adams 1993; Negi et al. 1995; Montero 1999; Oliveira

2003; Siddiqui et al. 2007, 2009; Behling 2009; Kumar et al. 2009; Kumar 2009, 2014),

referencias para la interpretación de análisis foliares (p.ej. Drechsel y Zech 1991, 1994; Zech y

Dreschel 1991) y varios trabajos que analizan las relaciones suelo-planta en plantaciones de teca

(para una revisión ver Kumar 2011; Alvarado 2012 b; Alvarado y Mata 2013). Varios autores

han propuesto modelos para estimar el índice de sitio potencial de un lugar en función de

variables edafoclimáticas (Vásquez y Ugalde 1994; Montero 1999; Mollinedo et al. 2005,

Thiele 2008). Estos trabajos presentan, en general, problemas metodológicos, pero coinciden en

señalar la importancia de los factores climáticos (sobre todo precipitación o número de meses

secos) y a veces fisiográficos sobre los edáficos, entre los que destacan los altos requerimientos

de Ca y la intolerancia a la acidez del suelo.

La teca es una especie normalmente considerada como calcícola y muy exigente en

propiedades edáficas, tanto químicas como físicas; suelos profundos (>90 cm), bien drenados y

con elevada fertilidad química (especialmente Ca y baja acidez) son normalmente asociados con

los requerimientos de esta especie (p.ej. Kumar 2011; Alvarado 2012 b; Alvarado y Mata 2013).

Pese a estas consideraciones, la teca ha sido plantada en una gran variedad de suelos incluyendo

algunos con problemas de fertilidad (p.ej. Ultsioles y Oxisoles) y otros que además presentan

dificultades añadidas debidas a las propiedades físicas del suelo (p.ej. Vertisoles) (Zech y

Drechsel 1991; Drechsel y Zech 1994; Ombina 2008; Kumar 2011; Alvarado 2012 b; Alvarado

y Mata 2013). Ombina (2008) menciona la existencia de plantaciones de teca en suelos de baja

fertilidad en Sudán del Sur, Zech y Dreschel (1991) encuentran suelos generalmente deficientes

en las de Liberia, mientras que Adekunle et al. (2011) describe las plantaciones de Nigeria

como con una alta fertilidad y Dreschel y Zech (1994) observan tanto alta como baja fertilidad

46

en las plantaciones de Togo, Benin, Côte d’Ivoire, Liberia y Nigeria. Kumar (2011) resume esta

variabilidad teniendo en cuenta fundamentalmente los suelos asiáticos y Alvarado (2012 b) y

Alvarado y Mata (2013) los latinoamericanos.

La evaluación y selección de sitio antes de establecer una plantación de teca resulta

fundamental para su éxito, principalmente teniendo en cuenta estudios de suelo (p.ej. Alvarado

2012 b; Alvarado y Mata 2013; Segura 2013). En ese sentido, el mal drenaje ha sido un

problema común en fincas en las que no se realizó una buena selección de sitio en las llanuras

del Norte de Costa Rica (con suelos poco profundos o muy compactados o incluso con plintita)

que luego ha causado numerosos problemas de muerte descendiente de los árboles (Arguedas et

al. 2009).

Muchos autores han señalado a la teca como poco tolerante a la acidez del suelo (p.ej. Zech y

Drechsel 1991; Drechsel y Zech 1994; Wehr et al. 2010; Zhou et al. 2012). Sin embargo,

Ombina (2008) señala que la especie ha sido plantada en un amplio rango de pH, desde 3,8 a

7,9. Alvarado y Fallas (2004) señalan como nivel crítico el umbral del 3% de Saturación de

Acidez por encima del cual se recomienda la aplicación de encalado para reducir la toxicidad y

aumentar el rendimiento productivo de la plantación (Figura 2.10): En ese sentido, los mismos

autores consideran que la Saturación de Ca debe ser superior al 68% para mantener altas

productividades (Figura 2.10) (Alvarado y Fallas 2004).

Figura 2.10. Efecto de la saturación de la acidez (izquierda) y de Calcio (derecha) del suelo sobre el

Incremento Medio Anual en altura en plantaciones de teca (Tectona grandis L.f.) con pH<6 en Costa Rica

(tomado de Alvarado y Fallas 2004)

47

Como se ha comentado anteriormente, el P también suele ser considerado como

determinante de la productividad de plantaciones forestales, aunque la teca suele ser

relativamente eficiente en el uso de este elemento (Alvarado 2012 a, b; Alvarado y Mata 2013).

Así, Zech y Drechsel (1991) identifican para este elemento un nivel crítico foliar de 0,13%

(±0.04) para plantaciones de teca en África Occidental, similar al que estos autores encuentran

su revisión bibliográfica: rango de deficiencia entre 0,1 y 0,13% y rango intermedio o adecuado

entre 0,12 y 0,21% (Drechsel y Zech 1991). La presencia de micorrizas, común en plantaciones

de teca, incrementa la producción de fosfatasas que favorece la mineralización y solubilización

del P orgánico y/o precipitado (Corryanti et al. 2007). Alvarado et al. (2004) seleccionaron

micorrizas en numerosas plantaciones de teca de Costa Rica y propusieron su inoculación en las

plántulas como medida para mejorar la disponibilidad y absorción de P por las plantas y mejorar

así la productividad de los sitios.

Varios autores han estudiado el efecto de la aplicación de enmiendas y fertilizantes sobre el

crecimiento y la productividad de plantaciones de teca (p.ej. Prasad et al. 1986 citado en Kumar

2011; Hernández et al. 1990; Mothes et al. 1991; Torres et al. 1993; Montero 1995; Singh 1997;

Alvarado y Fallas 2004; Zhou et al. 2012) encontrando resultados en general variados y a veces

contradictorios. Pese a la poca y confusa investigación al respecto, en general se considera que

al ser altos los requerimientos nutricionales de la teca, la fertilización es necesaria y de hecho se

suele realizar habitualmente siguiendo las siguientes recomendaciones:

Para plantaciones jóvenes en Kerala (Suroeste de India) se recomiendan dos

aplicaciones repetidas para un total de 163 kg ha-1

de urea, 375 kg ha-1

de roca fosfórica,

145 kg ha-1

de cloruro potásico, 105 kg ha-1

de cal agrícola finamente molida y 373 kg

ha-1

de sulfato de Mg en el primer año y la misma cantidad pero divida en cuatro veces

a lo largo del segundo y tercer año (Balagopalan et al. 2000, citado por Kumar 2011)

También para plantaciones en Kerala (Suroeste de India) otros autores recomiendan 30-

40 N g planta-1

, 15-20 g P2O5 planta-1

y 15-20 g K2O planta-1

al año desde que la

plantación tiene 2 años hasta que tiene 5 y después aplicar la misma cantidad cada 3 ó 4

años hasta los 10 ó 12 años (KAU 200, citado en Kumar 2011)

48

Para plantaciones en Costa Rica (y América Central en conjunto) se recomienda en

general la aplicación de una fórmula típica N-P-K (10-30-10 ó 12-24-12) al comienzo

de la época lluviosa acompañada de una dosis extra de N (en forma de urea) durante el

período de máximas lluvia hasta los 3-5 años en que la plantación cierra el dosel de

copas (Alvarado 2012 b). Sin embargo, en la práctica esta recomendación queda

reducida habitualmente a 50-150 g de una fórmula N-P-K o 5-15 g de un fertilizante de

liberación lenta (p.ej. osmocote) durante el establecimiento de la plantación.

Las plantaciones de teca normalmente se han considerado como causantes de degradación y

erosión de suelo; sin embargo, investigaciones recientes han demostrado como falso este mito y

han observado que las tasas especialmente altas que lo originaron ocurren generalmente en

plantaciones con una mala gestión (sobre todo las que usan el fuego como herramienta de

control de la vegetación acompañante, comunes en Asia) o aquellas que no han conseguido

restaurar las propiedades físicas del suelo que los usos anteriores habían degradado (p.ej.

ganadería mal gestionada) (para una revisión ver Fernández-Moya 2013).

CAPÍTULO 3

SOIL FERTILITY CHARACTERIZATION

OF TEAK (Tectona grandis L.f.) PLANTATIONS

IN CENTRAL AMERICA

51

3.1. Introduction

Teak (Tectona grandis L.f.) is an important species in the worldwide quality tropical

hardwood sector, with a total planted area of 4.3 ·106 ha (Pandey and Brown 2000; De Camino

et al. 2002; Kumar 2011; Kollert and Cherubini 2012). Teak has been extensively planted in

Central America (132,770 ha), mainly in Panama, Costa Rica and Guatemala (55,000, 31,500

and 28,000 ha, respectively) and it has also been introduced in El Salvador, Nicaragua,

Honduras and Belize (9,760 7,960, 450 and 100 ha, respectively) (Kollert and Cherubini 2012).

Despite the relatively minor importance of Central American plantations in the worldwide teak

sector, teak plantations have had quite far-reaching socio-economic and environmental effects in

Central America due to the small size of the countries. Panama is the third in terms of area

dedicated to teak plantation relative to the size of the country, while Costa Rica is the fifth and

El Salvador, the ninth (Kollert and Cherubini 2012). Arias (2004) highlighted the fact that forest

plantations (especially teak plantations) established by large or medium sized companies not

only provide environmental services but also play an important role in the sustainable

development of countries like Costa Rica, creating employment in rural areas where few other

job opportunities exist. In addition, many small landowners have also planted teak across the

region, and usually manage their plantations as a complementary crop alongside other land uses

within their farms.

Nieuwenhuyse et al. (2000) estimate that around 70% of the landscape should be covered by

teak plantations in order to maximize the regional income of their study area in the Caribbean

lowlands of Costa Rica. These authors report that the production of a valuable timber species,

such as teak, is more profitable than fast-growing low-quality wood species, such as Gmelina

arborea Roxb., or other land uses, such as basic grain and beef cattle ranching. Due to the high

profitability of teak plantations (Nieuwenhuyse et al. 2000; Pandey and Brown 2000; De

Camino et al. 2002), Central America is among the regions which have seen the greatest

increase in area dedicated to teak plantations. In 2010, the area occupied by teak plantations in

Guatemala was 16 times greater than in 1995; in Panama the area was 14 times larger, 13 times

larger in Nicaragua, 5 times larger in El Salvador and twice as large in Costa Rica (Kollert and

52

Cherubini 2012). These trends in Central America are even more marked if Latin America is

considered as a whole, since large increases have been reported in Ecuador and Brasil (Kollert

and Cherubini 2012), two countries where a sharp increase in the area occupied by teak

plantations is expected over the next few years.

Soil fertility is one of the major factors influencing forest productivity (e.g. Binkley 1986),

especially in teak plantations since teak is considered a calciphile species with high nutrient

requirements (see Kumar 2011; Alvarado 2012 b). Hence, high soil fertility is often assumed in

teak plantations. However, failure to select suitable sites for establishing large plantations along

with inadequate conditions in many small-scale farms have led to the current assortment of soils

in teak plantations across Central America.

The present work aims to (1) analyze and characterize the general soil patterns which may be

influencing teak plantations in Central America; (2) assess differences between countries and

sub-regions; (3) create a global framework to help contextualize the soil fertility analyses

conducted at sub-regional or farm level, and (4) determine the main problems associated with

soil fertility and use the findings to define further lines of research in the near future. No

attempt is made in this study to analyze the relationship between soil fertility and growth, as this

line of investigation is currently being addressed elsewhere. However, the rationale for the

present study is the need to gain a clearer understanding of teak growth performance at regional

scale.

3.2. Material and Methods

Study area

A set of teak (Tectona grandis L.f.) plantations was selected across Central America:

Guatemala, Costa Rica and Panama (Figure 3.1); from 17ºN in the Petén region of Guatemala to

8ºN in the southern region of Costa Rica and the Panama Canal Watershed. Although the

climate varies across the large study area, most of the study sites are in tropical or subtropical

moist forest (2,000 – 4,000 mm year-1

with 4 – 6 dry months) and tropical or subtropical wet

forest (4,000 – 8,000 mm year-1

with 0 – 3 dry months) life zones, according to the Holdridge

53

(1967) classification. A wide variety of USDA soil orders were also found throughout the study

area, including Inceptisols, Entisols, Vertisols, Andisols, Alfisols, Ultisols and perhaps Oxisols.

Most planted teak forests in Central America have been established on land previously used

for beef cattle ranching, although some sites had been used for agricultural crops such as corn or

bananas. The previous overgrazing of the land has generally led to soil compaction and erosion

in many planted teak forests in the region. This degradation has sometimes been wrongly

attributed to teak establishment and in many cases has caused high land preparation costs and a

decrease in the potential site productivity.

Management of planted teak forests varies depending on whether they are woodlots

belonging to small landowners or large scale company owned plantations. However, there are

general patterns that are common to most planted teak forests in Central America. The rotation

period is usually around 20 – 25 years, with an expected commercial volume of 100-150 m3 at

the final harvesting. Management of these plantations consists of continuous silvicultural

activities: land preparation, fertilization and liming during the establishment (at variable

dosages and formulas depending on the company), weed control, pruning and thinning

(approximately from 816-1111 trees ha-1

to 150-200 trees ha-1

at final felling); although some of

these activities are not always performed, depending on the management intensity adopted by

the owner or manager.

Figure 3.1. General

location of the

countries of the study

(Guatemala, Costa

Rica and Panama) in

Central America

54

Data collection, transformation and statistical analysis

A compilation of different local scale soil studies in which some of the authors had

participated was assembled. Soil fertility information from 684 sites was collected: 299 in Costa

Rica, 257 in Guatemala and 128 in Panama. Topsoil (0 – 20 cm) and subsoil (20 – 40 cm)

information was available for these sites and was averaged in order to estimate soil fertility at 0

– 40 cm. The following soil attributes were used for the present analysis: pH, available P,

available cations (Ca, Mg, K and acidity) and their derived attributes: Effective Cation

Exchange Capacity (“ECEC”), Ca Saturation (“Ca Sat”), Mg Saturation (“Mg Sat”), K

Saturation (“K Sat”), acidity Saturation (“A Sat”), Ca Mg-1

, Ca K-1

, Mg K-1

and (Ca+Mg) K-1

ratios. Most soil analyses were performed at the “Centro de Investigaciones Agronómicas”,

University of Costa Rica (“CIA-UCR”), where pH was determined in water 10:25, available Ca,

Mg and acidity were measured using KCl solution 1M 1:10 and available K and P were

analyzed with a modified Olsen solution pH 8,5 (NaHCO3 0,5 N, EDTA 0.01M, Superfloc 127)

1:10. This methodology is commonly known in Central America as “modified Olsen-KCl” and

is established as a routine activity in the internationally certified soil laboratory of the CIA-

UCR. Some of the soil samples from Panama were analyzed in the Tropical Agricultural

Research and Higher Education Center (“CATIE”) using the same methodology. For the

samples from Guatemala and some of the ones from Panama, pH was determined in water

10:25, and the other elements were analyzed using the Mehlich 3 methodology (HOAc 0.2M,

NH4NO3 0.25M, NH4F 0.015M, HNO3 0.013M, EDTA 0.001M, pH 2.5). In order to make the

results obtained by Mehlich 3 compatible with those obtained by modified Olsen-KCl, the

following transformations were used, based on the models proposed by Cabalceta (1995) and

Bertsch et al. (2005): (1) Ca(Olsen-KCl) = (1/1.05) · Ca(Mehlich3) [R2=0.95]; (2) K(Olsen-KCl) = (1/1.41) ·

K(Mehlich3) [R2=0.94]; (3) P(Olsen-KCl) = 2.829 + 0.647 · P(Mehlich3) [R

2=0.69].

A Principal Component Analysis (“PCA”) was performed to assess similarities between

sampled sites according to their soil fertility. Prior to performing the PCA, data were centered

and standardized using the mean and the standard deviation of each variable. The 684 soil sites

were used as rows and the 15 soil fertility variables as columns. A multivariate cluster analysis

55

was carried out in order to group the soil samples according to their similarities. Complete

linkage (or farthest neighbour) hierarchic clustering was performed using Euclidean distance to

measure the similarities in soil fertility between the sites and 10 groups were created. R

software was used for all the statistical analyses (R Development Core Team 2011). The

relationships between the different variables were also graphically explored by plotting each

variable against the others.

The critical values reported in the literature were used for soil analysis interpretation: (1) A

Sat = 3% and Ca Sat = 68% (Alvarado and Fallas 2004); (2) K Sat = 3.09% (for details see

chapter 7 of this thesis); and (3) the general values used in Costa Rica for the rest of the

variables (Bertsch 1998) (Table 3.1). In addition, some critical values were used based on the

experience of the authors in analyzing soil fertility in teak plantations throughout the region: Ca

= 10 cmol(+) L-1

, Mg = 3 cmol(+) L-1

and P = 5 mg L-1

(Table 3.1).

3.3. Results and discussion

The PCA results revealed a predictable antagonism between Ca-driven soil attributes with

respect to acidity and A Sat, Mg, Mg Sat, K, K Sat and P (Figure 3.2). Forest managers in the

region often form generalized ideas about certain countries or regions having better soils than

others. However, the multivariate analysis did not show any noticeable differences either

between the countries analyzed (Figure 3.2) or between regions within each country.

Conversely, the multivariate analysis revealed high variability within each region in the form of

a soil-fertility gradient. However, soils in Panama are more likely to show acidity problems

(low pH and Ca Sat, and high acidity and A Sat) whereas they generally have slightly higher P

content (Table 3.1, Figure 3.3). Soils from Costa Rica and Guatemala are rather similar, the

soils from Guatemala showing somewhat lower Mg content than those from Costa Rica, which

might result in Mg deficiency and differences in the cation balances (Table 3.1, Figure 3.3).

Similar K values were found in the three countries, the absolute values generally being adequate

although K sat values were low as the soils are very rich in Ca and Mg, inducing a generalized

K deficiency (Table 3.1, Figure 3.3).

56

Table 3.1. Summary of the soil fertility variables analyzed in teak (Tectona grandis L.f.) plantations in

Central America: Costa Rica, Guatemala and Panama. Mean, Confidence Interval at 95% (“CI”) and

range (minimum – maximum) are reported for each variable considering a country average. Critical

values for some of the variables are reported either from literature (Bertsch 1998; Alvarado and Fallas

2004; chapter 7 of this thesis) or based on the experience of the authors. Problematic values according to

those critical values are remarked in red bold type when it does not match either the literature or the

empirical critical values and orange bold type when it is adequate taking one of them into consideration

Regional

(n= 684)

Costa Rica

(n= 299)

Guatemala

(n= 257)

Panama

(n= 128)

Critical values

Literature Experience of the

authors

pH

Mean 6.0 5.9 6.6 4.8

CI (5.9, 6.0) (5.9, 6.0) (6.5, 6.7) (4.7, 4.9) 5.5

Range 3.6 - 8.4 4.4 - 7.8 4.2 - 8.4 3.6 - 7.6

Ca

(cmol(+) L-1)

Mean 20.4 25.3 19.7 10.5

CI (19.2, 21.7) (24.1, 26.6) (17.1, 22.4) (9.2, 11.7) 4.0 10.0

Range 0.7 - 143.9 1.2 - 59.9 0.7 - 143.9 1.2 - 43.6

Mg

(cmol(+) L-1)

Mean 5.7 7.8 2.1 8.1

CI (5.3, 6.1) (7.3, 8.3) (1.8, 2.4) (7.1, 9.0) 1.0 3.0

Range 0.2 - 33.5 0.5 - 23.4 0.2 - 16.5 1.0 - 33.5

K

(cmol(+) L-1)

Mean 0.3 0.3 0.3 0.4

CI (0.3, 0.3) (0.3, 0.4) (0.2, 0.3) (0.3, 0.4) 0.2

Range 0 - 3 0 - 3 0 - 1.9 0 - 1.6

Acidity

(cmol(+) L-1)

Mean 0.6 0.3 0.2 2.0

CI (0.5, 0.7) (0.3, 0.4) (0.1, 0.3) (1.4, 2.6) 0.5

Range 0 - 18.8 0.1 - 6.4 0 - 10.1 0.1 - 18.8

ECEC

(cmol(+) L-1)

Mean 27.0 33.8 22.3 20.9

CI (25.7, 28.4) (32.2, 35.3) (19.6, 25.0) (18.9, 22.9)

Range 1.6 - 149.5 4.7 - 80.2 1.6 - 149.5 3.8 - 69.5

Ca Sat

(%)

Mean 73.3 74.0 84.3 49.6

CI (72.0, 74.6) (73.0, 75.1) (82.6, 85.9) (47.2, 52.0) 68.0

Range 14.9 - 98.8 17.3 - 97.5 14.9 - 98.8 21.4 - 84.3

Mg Sat

(%)

Mean 21.5 22.9 11.9 37.8

CI (20.6, 22.5) (22.0, 23.7) (10.7, 13.2) (35.8, 39.8)

Range 0.7 - 70.2 1.6 - 45.2 0.7 - 70.2 12.8 - 67.8

K Sat

(%)

Mean 1.6 1.1 1.9 2.0

CI (1.4, 1.7) (1.0, 1.3) (1.6, 2.1) (1.8, 2.2) 3.09

Range 0 - 16.5 0.1 – 9.0 0 - 16.5 0.1 - 5.3

A Sat

(%)

Mean 3.6 2.0 1.9 10.6

CI (2.9, 4.3) (1.3, 2.7) (1.0, 2.8) (8.3, 13.0) 3.0

Range 0 - 72.9 0.1 - 60.5 0 - 72.9 0.2 - 64.5

Ca Mg-1

Mean 7.4 4.2 14.0 1.5

CI (6.5, 8.3) (3.7, 4.8) (12.1, 16.0) (1.4, 1.6)

Range 0.3 - 137.7 0.8 - 62.7 0.4 - 137.7 0.3 - 6.5

Ca K-1

Mean 147.3 213.2 119.1 50.1

CI (131.6, 163.0) (187.4, 238.9) (93.5, 144.8) (38.6, 61.7)

Range 3.2 - 1926.2 7 - 1518.5 3.2 - 1926.2 9.2 - 361.4

Mg K-1

Mean 44.2 71.7 14.2 40.0

CI (39.0, 49.3) (62.2, 81.1) (10.8, 17.7) (29.1, 50.8)

Range 0.7 - 493.2 1.8 - 492.8 0.7 - 340.5 5.5 - 493.2

(Ca+Mg) K-1

Mean 191.5 284.8 133.4 90.1

CI (171.7, 211.3) (250.2, 319.4) (105.5, 161.3) (68.8, 111.5)

Range 4 - 2119.1 9.6 - 1981.5 4 - 2119.1 16.2 - 752.8

P

(mg L-1)

Mean 7.9 7.6 8.1 8.5

CI (7.1, 8.7) (5.9, 9.2) (7.2, 8.9) (7.5, 9.5) 10.0 5.0

Range 0 - 78.6 0 – 78.0 2.8 - 78.6 0.1 - 20.1

ECEC: Effective Cation Exchange Capacity [ECEC=Acidity+Ca+Mg+K]. Ca Sat: Ca saturation [Ca

Sat=Ca/ECEC]. Mg Sat: Mg saturation [Mg Sat=Mg/ECEC]. K Sat: K saturation [K Sat=K/ECEC]. A

Sat: Acidity saturation [A Sat=Acidity/ECEC].

Although teak is considered a calciphile species, it has been planted at sites throughout

Central America with a wide range of soil pH values (3.6 – 8.4, Table 3.1). The findings of a

worldwide analysis by Ombina (2008) also reflect this situation (soil pH values 3.8 – 7.9). As

expected, the results of the present study show that pH can be used as a general indicator of the

57

Ca – acidity relationship in the analyzed soils, as acidity saturation falls to values below 3%

(critical level for the species, according to Alvarado and Fallas 2004) in soils with pH values

higher than 5.5, sharply increasing in soils with lower pH (Figure 3.3). This tendency is in

accordance with the classical theory regarding soil acidity in tropical regions (e.g. Kamprath

1984). Accordingly, soils with pH lower than 5.5 are more likely to have low Ca Sat and show

high Mg Sat, while soils with pH higher than 5.5 generally have Ca Sat values higher than 68%

(critical level for the species, Alvarado and Fallas 2004). Although cation saturation indexes

varied according to ECEC values and the general soil cation balance, when Ca content is higher

than 20 – 25 cmol(+) L-1

, the cases of A Sat problems seem to be less common (Figure 3.3).

However, high Ca values can also be a problem, as available P content decrease when Ca

exceeds levels of around 50 – 60 cmol(+) L-1

(Figure 3.3), probably because P would be

precipitated as Ca phosphates. Similarly, P content is higher where there is no noticeable soil

acidity problem (< 0.5 cmol(+) L-1

) and P availability becomes a problem when acidity is high

(>2.5 cmol(+) L-1

, approximately) (Figure 3.3), probably because P would be precipitated as Al

and/or Fe phosphates.

Figure 3.2. Results of

the PCA of the 684 soil

samples according to

the 15 soil fertility

variables analyzed in

teak (Tectona grandis

L.f.) plantations in

Central America: Costa

Rica, Guatemala and

Panama

58

Figure 3.3. Relationship between some of the soil fertility variables analyzed in teak (Tectona grandis

L.f.) plantations in Central America: Costa Rica, Guatemala and Panama

59

The 684 soil samples were divided into 10 groups according to their similarities as regards

soil fertility using cluster analysis (Table 3.2). Most groups showed a K deficiency (K Sat <

3.09%) except G-6, G-9 and G-10, the latter being the only group with no soil fertility problems

although it only comprised 10 samples (Table 3.2). G-6 only had slight problems of acidity

whereas G-9 had problems of acidity as well as available Ca and Mg (Table 3.2). Despite K

deficiency, G-2 had generally adequate cation values although their P content might be in the

critical range, as with the G-8 samples, although this group also showed a slightly lower Mg

content. The other groups (G-1, G-3, G-4, G-5 and G-7) presented different combinations of

nutrient deficiencies and toxicities, which in simple terms can be described as low soil fertility

(Table 3.2).

Teak is considered to be a species with high soil nutrient requirements; deep, well-drained

soils with high chemical fertility (especially Ca) and low acidity are usually considered

necessary for the successful growth of this species (Montero 1999; Alvarado and Fallas 2004;

Mollinedo et al. 2005; Kumar 2011; Alvarado 2012 b). However, teak plantations are found on

a wide variety of soils, including many with serious problems of fertility (e.g. Ultisols and

Oxisols) or with added difficulties due to their physical properties (e.g. Vertisols) (Zech and

Drechsel 1991; Drechsel and Zech 1994; Ombina 2008; Kumar 2011; Alvarado 2012 b). The

soils analyzed in the present study also showed high variability, including Inceptisols, Entisols,

Vertisols, Andisols, Alfisols, Ultisols and perhaps Oxisols. The wide variety of soil fertility

levels in teak plantations of Central America has also been observed in Africa. Ombina (2008)

reported teak plantations on low fertility soils in Southern Sudan, Zech and Drechsel (1991)

found generally deficient soils in Liberia, while Adekunle et al. (2011) described relatively high

fertility soils in Nigeria. Similarly, Drechsel and Zech (1994) found a variety of rich and poor

soils in Togo, Benin, Côte d’Ivoire, Liberia and Nigeria. Asian teak plantations have also been

established on a wide variety of soils (Kumar 2011).

60

Table 3.2. Summary of the soil fertility variables analyzed in teak (Tectona grandis L.f.) plantations in

Central America: Costa Rica, Guatemala and Panama. Mean, Confidence Interval at 95% (“CI”) and

range (minimum – maximum) are reported for each variable considering each of the 10 groups

differentiated by multivariate cluster analysis based on the soil fertility attributes. Critical values for some

of the variables are used either from literature chapter 7 of this thesis) or based on the experience of the

authors. Problematic values according to those critical values are remarked in red bold type when it does

not match either the literature or the empirical critical values and orange bold type when it is adequate

taking one of them into consideration

Group 1 2 3 4 5 6 7 8 9 10

(n=135) (n=452) (n=26) (n=9) (n=4) (n=12) (n=24) (n=10) (n=2) (n=10)

pH

Mean 4.9 6.2 6.1 5.1 4.7 6.2 6.1 8.0 5.3 6.2

CI (4.7, 5.0) (6.2, 6.3) (6, 6.3) (5.0, 5.2) (4.3, 5.1) (5.9, 6.5) (5.9, 6.4) (7.8, 8.1) (5.1, 5.5) (6.1, 6.4)

Range 3.6 - 7.4 4.4 - 8.4 5.4 - 6.9 4.8 - 5.4 4.2 - 5.1 5.2 - 6.9 5.5 - 7.3 7.4 - 8.2 5.2 - 5.4 5.9 - 6.6

Ca

(cmol(+) L-1)

Mean 9.5 21.0 33.2 9.0 1.4 25.3 25.8 102.7 1.4 30.6

CI (8.4, 10.6) (19.8, 22.2) (29.8, 36.6) (6.2, 11.8) (0.9, 1.9) (20.3, 30.3) (22.4, 29.2) (90.3, 115) (0.7, 2.2) (25.8, 35.4)

Range 1.2 - 29.3 1.4 - 73.3 17.6 - 55.9 2.2 - 16.2 0.7 - 1.9 8.0 - 42.5 9.7 - 47.0 80.0 - 143.9 1.0 - 1.8 21.3 - 45.8

Mg

(cmol(+) L-1)

Mean 6.5 4.6 17.9 9.3 1.0 7.9 7.2 2.4 0.3 7.0

CI (5.7, 7.3) (4.3, 5.0) (16.1, 19.8) (6.2, 12.4) (0.4, 1.5) (6.0, 9.8) (5.3, 9.1) (1.7, 3.1) (0.2, 0.4) (5.3, 8.7)

Range 0.3 - 19.8 0.2 - 18.3 12.4 - 33.5 2.5 - 18.6 0.2 - 1.5 1.9 - 14.2 0.2 - 14.1 0.7 - 3.7 0.3 - 0.4 4.4 - 13.1

K

(cmol(+) L-1)

Mean 0.5 0.2 0.1 0.1 0.1 1.2 0 0.6 0.4 2.4

CI (0.4, 0.5) (0.2, 0.2) (0.1, 0.2) (0.1, 0.2) (0, 0.1) (0.9, 1.5) (0, 0) (0.3, 0.9) (0.3, 0.6) (2.2, 2.7)

Range 0 - 1.9 0 - 1.4 0.1 - 0.4 0.1 - 0.2 0 - 0.1 0.1 - 1.9 0 - 0.1 0.2 - 1.9 0.3 - 0.5 2.0 - 3.0

Acidity

(cmol(+) L-1)

Mean 1.2 0.2 0.6 13.4 4.4 0.6 0.2 0 0.3 0.3

CI (1, 1.4) (0.2, 0.2) (0.1, 1.0) (11.3, 15.6) (3.1, 5.7) (0.1, 1.0) (0.1, 0.2) (0, 0) (0.1, 0.6) (0.2, 0.4)

Range 0 - 7.1 0 - 3.5 0.1 – 5.0 10.0 - 18.8 2.6 - 5.6 0.1 - 2.9 0 - 0.6 0 - 0.1 0.2 - 0.5 0.2 - 0.5

ECEC

(cmol(+) L-1)

Mean 17.7 26.0 51.8 31.9 6.9 34.9 33.1 105.7 2.5 40.3

CI (16, 19.4) (24.7, 27.4) (47.6, 56.1) (25.9, 37.8) (4.6, 9.1) (29.4, 40.5) (28.2, 38.1) (93.0, 118.4) (1.8, 3.3) (35.0, 45.6)

Range 2.2 - 45.7 1.6 - 75.3 35.4 - 80.2 15.0 - 41.1 3.6 - 8.7 17.1 - 52.2 10.5 - 56.0 83.6 - 149.5 2.1 - 2.9 28.1 - 53.9

Ca Sat

(%)

Mean 54.0 80.2 63.9 26.8 20.5 71.3 80.7 97.1 56.0 75.6

CI (51.3, 56.6) (79.2, 81.1) (60.9, 66.9) (22.2, 31.4) (18.3, 22.8) (64.9, 77.7) (76.5, 84.9) (96.3, 97.9) (41.9, 70.1) (71.8, 79.3)

Range 23.6 - 91.9 40.5 - 98.5 32.7 - 71.5 14.9 - 40.1 17.3 - 22.4 47.0 - 85.6 61.4 - 98.0 95.5 - 98.8 48.8 - 63.2 66.4 - 86.8

Mg Sat

(%)

Mean 34.2 17.6 34.8 27.9 13.0 23.0 18.8 2.3 13.0 17.5

CI (31.8, 36.7) (16.7, 18.5) (32, 37.6) (22, 33.7) (6.7, 19.2) (17.9, 28.1) (14.7, 22.9) (1.6, 3.0) (12.2, 13.8) (14.0, 21.0)

Range 3.3 - 70.2 1.1 - 58.8 26 - 62.1 16.8 - 45.2 5.0 - 20.7 9.3 - 35.3 1.8 - 36.8 0.7 - 3.8 12.6 - 13.4 8.6 - 26.3

K Sat

(%)

Mean 3.0 1.1 0.3 0.4 1.0 3.3 0.1 0.6 15.9 6.2

CI (2.7, 3.3) (1.1, 1.2) (0.2, 0.3) (0.2, 0.6) (0.5, 1.5) (2.6, 4.0) (0.1, 0.1) (0.4, 0.8) (14.7, 17.2) (5.5, 6.9)

Range 0.6 - 9.7 0.1 - 4.7 0.1 - 0.6 0.1 – 1.0 0.6 - 1.6 0.8 - 5.0 0 - 0.2 0.2 - 1.2 15.3 - 16.5 4.3 - 7.7

A Sat

(%)

Mean 8.8 1.1 1.0 44.9 65.5 2.4 0.4 0 15.1 0.8

CI (7.2, 10.4) (0.9, 1.3) (0.3, 1.8) (36.1, 53.6) (60.4, 70.7) (0, 5.1) (0.2, 0.6) (0, 0) (0, 31.3) (0.6, 0.9)

Range 0.1 - 44.2 0 - 16.4 0.1 - 8.7 24.4 - 67.4 60.5 - 72.9 0.4 - 17 0 - 1.8 0 - 0.1 6.9 - 23.4 0.3 - 1.3

Ca Mg-1

Mean 2.5 8.3 1.9 1.0 2.1 3.9 9.8 57.1 4.3 4.9

CI (2.0, 3.1) (7.4, 9.2) (1.7, 2.1) (0.8, 1.2) (0.7, 3.5) (2.6, 5.1) (4.6, 14.9) (33.7, 80.4) (3.5, 5.1) (3.5, 6.4)

Range 0.3 - 27.9 0.7 - 93.5 0.5 - 2.6 0.6 - 1.4 0.8 - 4.1 1.3 - 9.2 1.7 - 54.6 24.8 - 137.7 3.9 - 4.7 2.5 - 10.1

Ca K-1

Mean 23.8 137.7 307.9 94.4 26.1 26.1 987.2 222.1 3.5 12.7

CI (21.5, 26.0) (126.7, 148.7) (257.6, 358.1) (50.7, 138.0) (12.4, 39.9) (18.3, 34.0) (850, 1124) (145.0, 299.3) (2.9, 4.1) (10.7, 14.6)

Range 6.6 - 79.0 12.7 - 715.1 106.2 - 561.2 15.5 - 202.4 10.8 - 38.1 12.4 - 62.6

627.2 -

1926.2 77.2 - 468.6 3.2 - 3.8 9.9 - 20.3

Mg K-1

Mean 17.4 36.6 171.4 102.8 15.6 10.3 235.8 5.0 0.8 2.9

CI (15.2, 19.5) (32.4, 40.8) (135.9, 206.9) (41.9, 163.6) (6.9, 24.2) (3.5, 17.2) (177.8, 293.8) (2.8, 7.2) (0.8, 0.8) (2.2, 3.6)

Range 0.7 - 67.8 1.2 - 230.1 56.2 - 493.2 17.5 - 301.9 4.3 - 23.4 2.0 - 47.1 18.9 - 492.8 1.5 - 12.0 0.8 - 0.8 1.8 - 5.1

(Ca+Mg) K-1

Mean 41.1 174.3 479.2 197.1 41.7 36.4 1223 227.1 4.3 15.6

CI (37.3, 45) (159.8, 188.8) (403.0, 555.5) (94.0, 300.2) (20.4, 63) (22.2, 50.7) (1065, 1381) (148.3, 306.0) (3.7, 4.9) (13.5, 17.7)

Range 7.9 - 123.2 19.4 - 798.4 162.3 - 860.4 33.0 - 504.3 22.1 - 61.3 18.8 - 109.7

653.8 -

2119.1 79.2 - 480.6 4 - 4.6 11.9 - 22.3

P

(mg L-1)

Mean 10.7 5.6 1.3 0.7 1.8 56.1 4.6 9.5 8.8 51.7

CI (9.8, 11.7) (5.0, 6.1) (0.6, 2.0) (0.1, 1.3) (0, 3.6) (48.9, 63.2) (2.6, 6.6) (7.0, 12.0) (8.2, 9.5) (37.8, 65.6)

Range 0.1 - 30.0 0 - 43.5 0 - 7.7 0.1 – 3.0 0.6 - 4.6 41.5 - 78.6 0 - 14.1 2.8 - 15.9 8.5 - 9.2 23.5 - 78.0

ECEC: Effective Cation Exchange Capacity [ECEC=Acidity+Ca+Mg+K]. Ca Sat: Ca saturation [Ca

Sat=Ca/ECEC]. Mg Sat: Mg saturation [Mg Sat=Mg/ECEC]. K Sat: K saturation [K Sat=K/ECEC]. A

Sat: Acidity saturation [A Sat=Acidity/ECEC].

Despite the commonly high nutrient requirements of the species and the general belief that

soils under teak plantations tend to be fertile, our data reveals that most sites in Central America

exhibit soil deficiencies of some kind, especially related to K, acidity and P (Tables 3.1 and 3.2,

61

Figure 3.3). Site selection based on soil fertility was not a widespread practice when the first

large teak plantations of the region were established in the 80s. Although several protocols were

later developed (Keogh 1987; Müller et al. 1998; Núñez 2001), many teak plantations of the 80s

and 90s were established on poor sites due to corruption, technical negligence and/or

administrative problems. Poor site selection is considered one of the causes of the failure of

several financial investments based on plantations in Central America. These failures have left

Central American plantations with a poor reputation in certain countries (e.g. Netherlands),

where many people lost their savings after investing in plantations which were sold as financial

products with expected high rates of return and which did not live up to expectations. Numerous

sites in the northern lowlands of Costa Rica which were acquired for teak plantations had soil

fertility and/or physical problems (e.g. shallow soils with bad drainage and plinthite), leading to

dieback syndrome in the following years (Arguedas et al. 2009), even though the soils were

considered adequate at the time and high teak growth was anticipated. Similarly, many of the

initial teak plantations in Guatemala were established on sites with poor soils, although since the

implementation of the site selection protocol (Segura et al. 2013) the soil quality at recently

acquired sites has improved. Despite the fact that soil fertility is carefully evaluated in most teak

plantations established today, it seems that managers still have to deal with soil fertility

deficiencies (particularly as regards K, acidity and P) at many sites acquired several years ago.

Fernández-Moya et al. (2014) already highlighted the importance of K and P in teak

plantations in Central America, as the accumulation in tree tissues and exporting of both

elements through timber harvesting is very high in relation to the amounts available in the soil.

In addition, Fernández-Moya et al. (see chapter 7 of this thesis) state that P and K deficiencies

are directly related to a decrease in teak performance in Panama. However, the underlying cause

may not be K deficiency but rather a problem of cationic unbalance due to the high values of Ca

and/or Mg (Tables 1 and 2). Two further problems exist with regard to K management in the

studied systems: (i) susceptibility to leaching due to the high levels of precipitation throughout

the study area, and (ii) fixation, especially if the soil mineralogy is dominantly vermiculite,

abundant in many soils in the region considered as highly fertile. In addition, Mg deficiencies

62

have been observed by forest managers in the field, although this is not reflected by the results

of the present study.

Although some high available P values are found, the results show a general P deficiency in

the soils under teak plantations in the region. Most of the P rich sites probably result from a

residual effect of the fertilization applied during the previous land use. This is especially

common in Guatemala, where teak plantations have commonly been established on sites

previously used for agriculture (e.g. maize…), whereas in Costa Rica and Panama, they have

generally been established on former grasslands. Phosphorus is recognized as an important

limiting nutrient in forest plantations (see Fox et al. 2011) and is considered to be generally

deficient in tropical forests, which is why tropical tree species are usually P-efficient (Vitousek

1984; Hedin et al. 2009). Although P fertilization is usually considered to be a requirement,

especially in large-scale plantations, the presence of mycorrhizas has been identified as

necessary for the production of phosphatases (Corryanti et al. 2007), which improve the

mineralization rates of organic-P, resulting in higher levels of available P. Alvarado et al.

(2004) collected mycorrhizas in teak plantations throughout Costa Rica and proposed the

inoculation of seedlings as a way to improve P uptake and enhance productivity, particularly in

acid soils.

Many authors have reported soil acidity-toxicity problems in teak plantations (e.g. Zech and

Drechsel 1991; Drechsel and Zech 1994; Wehr et al. 2010; Zhou et al. 2012) and liming is

recommended when A Sat is higher than 3% (Alvarado and Fallas 2004). Although large

investors now take into account soil requirements and the poor adaptation of teak to acid soils

when acquiring land for establishing new plantations, many small-landowners cannot afford to

pay the prices demanded for agricultural land and have to make do with the land they already

own, which in many cases presents low soil fertility and acidity problems (e.g. red acidic

Ultisols). The establishment of teak plantations on this marginal land is a mistake which has

been frequently made. Moreover it has been found that in some cases the establishment of

native species such as Terminalia amazonia and Swietenia macrophylla could be more

profitable than teak plantations (Griess and Knoke 2011).

63

Despite their importance to forest growth, the present study cannot take into consideration

either physical soil properties or micronutrient availability due to the shortage of data in this

regard. Within a small region with a homogeneous climate, a number of variables such as bulk

density, water retention capacity or the position on the slope can affect the amount of available

water in a given plot over the course of the year. Similarly, large amounts of certain

micronutrients can be found in teak biomass (Fernández-Moya et al. 2013, 2014) and despite a

lack of concern with respect to this circumstance, it may imply (1) a hidden soil fertility

deficiency; and (2) a problem in future rotations.

64

CAPÍTULO 4

NUTRIENT CONCENTRATION AGE DYNAMICS

OF TEAK (Tectona grandis L.f.) PLANTATIONS IN

CENTRAL AMERICA

Este capítulo ha sido publicado como:

Fernández-Moya J, Murillo R, Portuguez E, Fallas JL, Ríos V, Kottman F, Verjans JM, Mata R,

Alvarado A. 2013 Nutrient concentration age dynamics of teak (Tectona grandis L.f.)

plantations in Central America. Forest Systems 22 (1): 123-133.

(Anexo I)

67

4.1. Introduction

Teak (Tectona grandis L.f.) has been planted extensively in Central America, acquiring

socio-economical relevance due to its productivity (Pandey and Brown 2000; De Camino et al.

2002). Reference foliar nutrient concentrations have been summarized for teak (Drechsel and

Zech 1991; Boardman et al. 1997), and a preliminary Diagnosis and Recommendation

Integrated System (DRIS) has been developed for West African planted teak forests (Drechsel

and Zech 1994). However, appropriate knowledge regarding teak nutrition is still required for a

better management of the plantations to attain high productivity and sustainability.

The concentrations of nutrients in tissues depend mainly on species, environmental factors

(climate and soil availability) and plantation management. Whether comparing different species

or within a single species, genetic requirements, root distribution and age (developmental stage)

are usually the most important factors affecting nutrient absorption. At early growth stages, tree

nutrition is considered crucial to sustain high growth rates and rapid expansion of the crown and

roots. After the crown is fully developed, if seedling nutrition has been adequate, tree

requirements during the remainder of the rotation are assumed to be satisfied by environmental

inputs, nutrient recycling and nutrient translocation (Miller 1981, 1984, 1995). Nutrients in

foliage are considered to represent 20-40% of the total amount of nutrients in a stand, and lower

nutrient concentrations found in the tree bole (e.g. Miller 1995). Kumar et al. (2009) found teak

reproductive parts to contain the highest nutrient concentrations, while twigs and foliage also

showed high values; the Ca and B concentrations in bark are higher than in other tissues.

Foliar concentration is considered a useful parameter to evaluate the nutritional status of a

stand and as a reference to evaluate plantation fertilizer recommendations because (a) its

variation is highly dependent on site and soil parameters; (b) it reflects the current nutrient

supply; (c) it allows diagnosis of nutritional deficiencies when they are not severe enough to

cause visually observable symptoms and, thus, allows action to be taken before the effects on

productivity are significant; and (d) deficiency symptoms are confused easily with other effects

when visual guidelines are used (Mead 1984; Drechsel and Zech 1991; Barker and Pilbeam

2006; Lehto et al. 2010).

68

The nutrient concentrations presented in this paper for dominant and co-dominant teak trees,

which were developed for a variety of soils and environmental conditions in Central America

(Costa Rica and Panama), should be taken as reference values for evaluating the nutritional

status of similar teak plantations in the region.

4.2. Materials and methods

Study sites

Three teak (Tectona grandis L.f.) plantations were studied in Central America: two in Costa

Rica (Guanacaste and San Carlos) and one in Panama (Panama Canal watershed) (Figure 4.1).

The three areas are classified as tropical wet forest according to Holdridge’s life zones, with

similar mean annual rainfall (2500–3100 mm), although in Guanacaste the dry season is longer

than at the other two sites. The soils of the study areas are also similar, although the northern

region of Costa Rica is less fertile than the other sites and it has higher soil acidity (Table 4.1).

The studied stands were chosen to be representative of properly managed teak plantations in

Central America. In general, management of these plantations consists on continuous

silvicultural activities: weed control, pruning, thinning regimen (approximately from 800-1000

trees ha-1

at establishment to 150-200 trees ha-1

at final felling) and fertilization during the

establishment. The use of clones is common in recent years, so they were not sampled in the

study. A commercial volume of 100-150 m3 is expected for this kind of plantations in

approximately 20 years rotation.

Figure. 4.1. Locations

of the study teak

(Tectona grandis L.f.)

plantations: Guanacaste

(Costa Rica); northern

region (Costa Rica) and

Panama Canal

Watershed (Panama).

69

Table 4.1. Summary of soil properties at the different study areas; means and coefficients of variation (in

parentheses) are provided. Soil information was only available for 23 of the 28 sampled stands.

Northern region,

Costa Rica

(n=11)

Guanacaste,

Costa Rica

(n=9)

Canal Zone,

Panama

(n=3)

TOTAL

(n=23)

pH 5.11 (6) 5.90 (6) 6.70 (12) 5.63 (12)

Acidity (cmol(+) L-1

) 0.70* (5) 0.31 (30) 0.15 (33) 0.48 (81)

Ca (cmol(+) L-1

) 4.45 (44) 21.36 (28) 20.97 (38) 13.22 (74)

Mg (cmol(+) L-1

) 1.46 (47) 6.89 (54) 5.25 (64) 4.08 (89)

K (cmol(+) L-1

) 0.13* (109) 0.33 (87) 0.36 (82) 0.24 (101)

ECEC (cmol(+) L-1

) 6.74 (31) 28.90 (32) 26.72 (40) 18.02 (71)

Acidity Saturation (%) 11.96* (84) 1.22 (58) 0.65 (55) 6.28* (139)

P (mg L-1

) 3* (114) 3* (146) 2* (0) 3* (124)

Zn (mg L-1

) 2* (84) 3 (58) 3 (107) 2* (77)

Cu (mg L-1

) 8 (19) 11 (85) 4 (83) 9 (71)

Fe (mg L-1

) 165 (23) 37 (81) 65 (154) 102 (75)

Mn (mg L-1

) 43 (171) 38 (82) 19 (101) 38 (142)

Organic matter (%) 4.6 (27) 3.8 (28) 4.6 (16) 4.3 (27)

Sand (%) 24.9 (23) 23.4 (56) 29.0 (31) 24.8 (38)

Silt (%) 18.4 (13) 36.9 (42) 36.8 (43) 28.0 (51)

Clay (%) 56.7 (11) 39.7 (22) 34.3 (50) 47.1 (27)

ECEC: effective cation exchange capacity. * = Values outside the adequate reference soil levels (Bertsch

1998).

Field sampling and design

False time series (chronosequences) method was used to analyze nutrient concentration

dynamics of teak trees from 1 to 19 years of age. Despite of the critiques of this method

(Johnson and Miyanishi 2008), it was considered to be valid as all the studied stands are

considered to be similar in environmental conditions (soil and climate) and management

practices.

A total of 28 stands were analyzed, seven in Panama, 12 in the northern region of Costa Rica

and nine in Guanacaste (Costa Rica). In order to analyze a maximum yield research experiment

(Bertsch 1998; Alvarado 2012 b), dominant and co-dominant trees were selected: (a) without

visible symptoms of diseases or nutritional deficiencies and (b) that were representative of the

best-performing trees of the plantations, assuming optimal nutrition and a full expression of

genetic potential. By analyzing nutrient concentration in the most productive soils without soils

deficiencies (and in dominant or co-dominant trees in a site), the maximum species

requirements are assessed; hence, if the considered minimum inputs for these high-fertility sites

70

are applied in sites of lower fertility where tree nutrient uptake would be lower, the productivity

of the plantation is still achieved (Bertsch 1998; Alvarado 2012 b).

In stands younger than ten years of age, two trees were sampled per stand, whereas only one

tree was sampled in older stands. Trees were felled and tree components (bole, bole’s bark,

foliage and primary and secondary branches) were analyzed for nutrient concentration. Tissue

samples (1 kg per tissue per tree) were collected and analyzed at the “Centro de Investigaciones

Agronómicas” of the University of Costa Rica (hereafter CIA-UCR) to determine nutrient

concentrations (N, P, Ca, Mg, K, S, Fe, Mn, Cu, Zn and B, hereafter referred to as nutrients)

after samples were dried and water content was assessed. Dry combustion was used to measure

the N concentration, and wet digestion and atomic spectrometry were used to extract and

determine other nutrients (Bertsch 1998). Primary and secondary branches were weighted

averaged and are reported as “branches”. Foliage samples were collected as representative

homogeneous mixture of all the foliage of the sampled trees. All the field work was performed

during July-September, at the trees optimal nutritional status during the period of maximum

growth activity, to avoid effects of seasonality. In order to estimate soil nutrient availability,

topsoil samples were collected (0-20 cm). Soil information was only available for 23 of the 28

sampled stands (Table 4.1). Soil samples were analyzed at CIA-UCR to determine: pH (in

water), Ca, Mg, K, P, Fe, Cu, Zn, Mn, exchangeable acidity and Al, following the KCl-modified

Olsen method, as described by Díaz-Romeu and Hunter (1978). Organic matter was determined

by the wet combustion method of Walkey and Black, as described by Briceño and Pacheco

(1984). Soil texture was determined using the modified Bouyoucos method, as described by

Forsythe (1975).

Statistical analysis

Generalized linear mixed models (GLMMs) were used to study the relationships between the

concentration of nutrients (N, P, Ca, Mg, K, S, Fe, Mn, Cu, Zn and B) in each tissue (bole, bark,

bole and bark, branches, foliage and total) and tree age. The use of GLMMs was required, as

most of the study variables did not approach the normal distribution hypothesized in traditional

71

models. A total of 44 response variables analyzed and the distribution approached by the data

used to construct the GLMMs. The exponentially distributed variables were modeled using a

Gamma distribution approach with α=1.

To evaluate the best fitted model for each study variable, a total of 83 different models were

constructed, selecting the one with lowest deviance. Three groups of models were constructed:

(1) a null model considering only an intercept [yi = b0]; (2) a model considering an intercept in

addition to age as an explanatory variable [(yi) λ = b1 · age + b0]; and (3) a model without an

intercept [(yi) λ = b1 · age ]. For groups (2) and (3), 41 different power link functions [g(μ)=μ

λ]

were tested for each one, with λ varying between λ=2 to λ=-2 and a λgap=0.1. When no model

including age as a parameter was statistically significant, or when the data did not follow any of

the studied distribution functions, the resulting model included only an intercept representing

the mean of the variable, and no age effect was taken into account.

The sampled stands in each study area were spatially correlated. The spatial correlation was

taken into account by including a random effect for the study area, modeling the working

correlation matrix with a first-order autoregressive structure. The goodness-of-fit of the models

was estimated by measuring the percentage difference between the deviance of the model and

the deviance of a model with no covariates (hereafter referred to as efficiency, EF), which is a

pseudo-R2 measure reported for GLMMs. All statistical analyses were performed using SAS 9.0

(SAS Institute Inc., 2002). All statistical tests throughout the text are considered significant with

α=0.05.

4.5. Results

N concentrations decreased with age in all tissues (Figure 4.2, Table 4.2); the N

concentration was higher in the foliage (1.7-2.3%) than in the other tissues. Ca concentrations

did not show any relationship with age in any tissue (Figure 4.3, Table 4.2) but were highest in

the bark (1.9%), followed by the foliage (1.3%) and finally the bole (0.1%). K concentrations

decreased with age in the bole, bark and branches but were constant in the foliage (Figure 4.4,

Table 4.2). K-bark (0.6-1.6%) was higher than in the other tissues during the early growth years,

72

but after 14 years, K-foliage was highest (0.88%). Mg-bole and Mg-branches decreased with

age, whereas Mg-foliage increased and Mg-bark was constant (Figure 4. 5, Table 4.2). The Mg-

foliage (0.23-0.34%) was highest, although Mg-bark was also high (0.23%) compared with Mg-

bole and Mg-branches. P concentrations did not show any relationship with age in any tissue

(Table 4.2). P-foliage (0.16%) was higher than in the other tissues (<0.08%). S-bark decreased

with age but remained constant in the other tissues (Table 4.2). In most tissues, the

macronutrient concentrations generally followed the trend N>Ca=K>>Mg>P>S.

Figure 4.2. Tissue N

concentration (%)

related to tree age

(years). Points represent

trees sampled at three

different locations:

Guanacaste, Costa Rica (

); northern region,

Costa Rica ( ) and

Panama ( ). Lines

represent fitted models

(Table 4.2).

Figure 4.3. Tissue Ca

concentration (%) related

to tree age (years). Points

represent trees sampled

at three different

locations: Guanacaste,

Costa Rica ( ); northern

region, Costa Rica ( )

and Panama ( ). Lines

represent fitted models

(Table 4.2).

73

Table 4.2. Regressions between tissues’ nutrient concentration and tree age (years) in 1 to 19 years old

teak (Tectona grandis L.f.) plantations in Costa Rica and Panama. Below specified models are in the form

[y = b0 + (b1 · age)1/λ

], where the response variables (y) are the nutrients concentration at tree tissues.

When no model including age as a parameter was statistically significant, the model only included an

intercept (b0) representing the mean of the variable.

Tissues Macronutrient (%) b0 b0

[Std. error] b1

b1

[Std. error] λ EF (%)

Foliage

N (%) 0.1845 0.0170 0.0073 0.0004 -2 34

Ca (%) 1.3363 0.1062

K (%) 0.8759 0.0736

Mg (%) 0.0501 0.0151 0.0033 0.0003 -2 90

P (%) 0.1589 0.0198

S (%) 0.1181 0.0047

Fe (mg kg-1) 129.6087 22.9492

Mn (mg kg-1) 42.5507 1.7941

Cu (mg kg-1) 11.0761 0.4361

Zn (mg kg-1) 31.9978 3.7041

B (mg kg-1) 19.6158 0.6613

Bark

N (%) 1.4506 0.1306 0.1258 0.0246 -2 36

Ca (%) 1.9098 0.2409

K (%) 1.2218 0.0234 -0.0205 0.0042 0.4 30

Mg (%) 0.2349 0.0072

P (%) 0.0772 0.0112

S (%) 32.0732 1.6898 2.2538 0.3823 -1.4 23

Fe (mg kg-1) 97791.3100 19837.4200 -5106.0 1043.8660 2 29

Mn (mg kg-1) 0.0004 0.0002 0.0002 0.0001 -2 43

Cu (mg kg-1) 3.4615 0.3050

Zn (mg kg-1) 29.8471 4.5945

B (mg kg-1) 30.6635 1.1832

Bole

N (%) 6.1237 0.9564 0.6100 0.1253 -2 26

Ca (%) 0.1093 0.0057

K (%) 0.6179 0.0197 -0.0229 0.0029 0.5 72

Mg (%) 18.9991 3.2787 7.2370 1.5985 -1.7 30

P (%) 0.0645 0.0162

S (%) 0.0446 0.0098

Fe (mg kg-1) 72.4639 23.0074

Mn (mg kg-1) 1.2500 0.3113

Cu (mg kg-1) 2.0907 0.2691

Zn (mg kg-1) 10.3356 4.4177

B (mg kg-1) 2.7143 0.2763

Branches

N (%) 2.4636 0.4475 0.2449 0.0420 -2 36

Ca (%) 0.9122 0.0544

K (%) 0.4282 0.0458

Mg (%) 14.5100 3.9542 5.9444 0.3472 -2 39

P (%) 0.0751 0.0221

S (%) 0.0673 0.0064

Fe (mg kg-1) 162.7561 8.7137

Mn (mg kg-1) 13.9337 1.2417

Cu (mg kg-1) -0.0060 0.0026 0.0085 0.0010 -2 48

Zn (mg kg-1) 3.3847 0.1286 -0.0648 0.0088 0 28

B (mg kg-1) 11.1654 0.1765

EF (%): Model efficiency, pseudo R2 estimate for Generalized Linear Mixed Models.

74

Fe-bark decreased with age but remained constant in the other tissues (Table 4.2); in young

trees, Fe-bark (28-304 mg kg-1

) was higher than in other tissues, although after 14 years, Fe-

branches (163 mg kg-1

) was highest. Mn-bark decreased with age but remained constant in the

other tissues (Table 4.2). Cu-branches decreased with age but remained constant in the other

tissues (Table 4.2). Zn-branches decreased with age but remained constant in the other tissues

(Table 4.2). Zn-foliage (32 mg kg-1

) was higher than the Zn concentration in the other tissues,

although the Zn-bark (30 mg kg-1

) was also high as was Zn-branches in younger trees (9-28 mg

kg-1

). B concentrations did not show any relationship with age in any tissue (Table 4.2); B-bark

(31 mg kg-1

) was highest, although B-foliage (20 mg kg-1

) was also high compared to B-

branches (11 mg kg-1

) and B-bole (3 mg kg-1

). In general, the micronutrient concentrations in

the tissues followed the trend Fe>>Mn>Zn=B>Cu.

Figure 4.4. Tissue K

concentration (%) related

to tree age (years). Points

represent trees sampled

at three different

locations: Guanacaste,

Costa Rica ( ); northern

region, Costa Rica ( )

and Panama ( ). Lines

represent fitted models

(Table 4.2).

Figure 4.5. Tissue Mg

concentration (%) related

to tree age (years). Points

represent trees sampled

at three different

locations: Guanacaste,

Costa Rica ( ); northern

region, Costa Rica ( )

and Panama ( ). Lines

represent fitted models

(Table 4.2).

75

Ca was found to be the most concentrated nutrient in the tree bark; its concentration in the

bark remained constant with age, which was also found for Mg, P, S, Cu, Zn and B. In contrast,

the N, K, Fe and Mn concentrations in the bark declined with tree age. N was the most

concentrated nutrient in the tree bole and had decreasing concentrations with age. K-bole was

also high in young trees but decreased sharply with age, with very low values at the end of the

rotation. Mg-bole declined slowly with age due to the low values found in young trees. The

relatively low concentrations of Ca, P, S, Fe, Mn, Cu, Zn and B in the bole showed no

relationship with tree age. K was the most concentrated nutrient in young tree branches but

decreased sharply with tree age. Ca-branches was highest in trees older than approximately 3

years. The concentrations of other elements in the tree branches, including N, Mg, Cu and Zn,

also declined with age, whereas those of Ca, P, S, Fe, Mn and B did not show any relationship

with tree age. N was the most concentrated nutrient in the foliage and decreased with age, while

the other elements did not show any relationship with age, with the exception of Mg, which

increased with tree age.

4.6. Discussion

In West African planted teak forests, Drechsel and Zech (1994) observed decreases in foliar

N and P concentrations with tree age; Montero (1999) also reported foliar N, P and K

concentrations decrease in Costa Rica. Similarly, Siddiqui et al. (2009) found foliar N, P, K and

Zn concentration decrease with age, whereas Ca and Mn increase. Siddiqui et al. (2007) also

reported a significant reduction in nutrient concentrations in teak roots with trees age. These

previously reported patterns do not correspond to the dynamics found in the present study, in

which the only foliar nutrient concentration found to decline with tree age was N (Figure 4.2,

Table 4.2), whereas there was a tendency for the Mg concentration to increase with tree age

(Figure 4.5, Table 4.2).

N concentrations decreased with age in all tissues (Figure 4.2), especially in the foliage (2.3

to 1.7%), as has been previously reported for teak (Montero, 1999) and is considered to be a

general trend in plant nutrition (Barker and Pilbeam 2006; Yuan et al. 2007). High N

76

concentrations in young trees could be due to the large requirements of plants at this fast-

growing stage, as N is usually related to plant growth (Fölster and Khanna 1997; Barker and

Pilbeam 2006). However, as greater soil N availability leads to higher plant N concentrations,

the high N concentration at the beginning of the rotation could also be explained by the large

amount of N available from the soil at this stage, which could be supplied either by large

amounts of organic residues combined with high mineralization rates and/or by fertilization.

Thus, a plant would absorb large amounts of N, store it as a reservoir and use it later during the

rotation by translocation from one tissue to another (Miller 1984; Yuan et al. 2007, Yuan and

Chan 2010). Hence, the application of N fertilizer at this stage could be futile or could cause

even greater losses due to leaching, resulting in contamination and economic losses (Fölster and

Khanna 1997).

The decreases in N concentration observed in several tissues with tree age (Figure 4.2) could

be explained by (a) decreasing plant requirements as plants age and decline in productivity and

require less N to support these lower growth rates; (b) a growth dilution effect, as plant biomass

increases with age and usually tends to allocate more structural and storage materials containing

little N (Yuan et al. 2007); or (c) a decline in the soil N supply, which would result in lower N

uptake and greater nutrient translocation using the nutrients stored during younger years.

However, declines in N concentration with age are also considered one of the causes of age-

related declines in forest productivity because (i) N is usually considered to be the limiting

nutrient in forest ecosystems, particularly in young tropical soils (Hedin et al. 2009); (ii) the N

mineralization rate in older forest soils is lower than in younger stands, causing a diminishing

soil N supply in older stands; and (iii) plant N supply and forest nutrition are generally related to

the photosynthesis rate, so a decrease in plant N supply would cause lower net primary

production (Gower et al. 1996; Ryan et al. 1997; Binkley et al. 2002). Understanding whether N

concentration declines are a cause or a consequence of planted forest productivity declines is an

important issue that should be addressed in further research. One way to resolve this issue could

be to establish a fertilization experiment in mature planted teak forests (combined with a

thinning program), monitoring the N concentration before and after fertilization and evaluating

77

whether an increase in the N concentration would eventually result in an increase in growth

rates.

High Ca concentrations in teak bark have also been reported by other authors (Nwoboshi

1984; Totey 1992; Negi et al, 1995). The Ca concentration in teak bark tended to increase with

tree age in the trees sampled in Guanacaste (Costa Rica) and Panama, but we could not fit a

sound statistical model reflecting this tendency, as some trees in the northern region of Costa

Rica presented low Ca bark concentrations (Figure 4.3). These low concentrations probably

reflect Ca deficiencies or certain disorders, as the soils in this region exhibited lower values of

available soil Ca and high acidity (Table 4.1). However, the foliar Ca concentration of the trees

in this region was adequate and comparable to that of trees from other regions (Figure 4.3). If

Ca deficiency occurs it would mainly affect new leaves (Barker and Pilbeam 2006), so trees

with lower bark Ca concentrations could have suffered a Ca deficiency in the past but then

recovered, thus exhibiting adequate nutritional status at sampling time. This phenomenon could

be explained by lime application at intermediate tree ages.

Zech and Drechsel (1991) consider values of 0.96-1.21 for the Ca:K ratio in foliage as

adequate for healthy teak trees; the average ratio value found in our study was 1.53, reflecting a

nutrient imbalance involving a Ca excess and/or a K deficiency. It could also reflect a difference

between African planted teak forests and the investigated Central American teak stands in terms

of environmental, management or even genetic differences. The foliar K concentration

(0.88±0.07%) fell at the lower end of the range (0.80-2.32%) considered as adequate (Drechsel

and Zech 1991; Boardman et al. 1997), higher than the values reported by Negi et al. (1990) in

India (0.83%) and Benin (0.29%). K is a mobile nutrient that plays key roles in photosynthesis

and CO2 assimilation (Barker and Pilbeam 2006) and exerts a regulatory effect on stomatal

movement and transpiration rates; thus, foliar K requirements would be expected to increase

with tree age because modifying transpiration rates to control increased hydraulic resistance and

sustaining a higher photosynthesis rate are two of the key physiological mechanisms underlying

the plant response to the aforementioned age-related decline in plant productivity (Gower et al.

1996; Ryan et al. 1997; Binkley et al. 2002). However, the foliar K concentration showed no

78

relationship with tree age, in spite of the foliar K decreases associated with age in the sampled

trees reported by Montero (1999). The decreasing K concentrations in the bole, bark and

branches observed with increasing tree age (Figure 4.4) are probably related to the constant

foliar K concentration, as the increasing foliar K requirements are probably supplied by the K in

those tissues by nutrient translocation.

An increase in the foliar Mg concentration with tree age, as found in this study, has also been

reported by Montero (1999) in teak plantations in Costa Rica. This increase could be related to

the decline in the Mg concentrations in the bole and branches because Mg is probably

translocating from these tissues to leaves. Foliar Mg may increase with age to meet the

physiological demands of older trees; these physiological needs include the following: a) to

sustain high photosynthesis efficiency; b) to partially inhibit excess photophosphorylation; and

c) to regulate leaf stomatal conductance (Gower et al. 1996; Gholz and Lima 1997; Ryan et al.

1997; Binkley et al. 2002; Barker and Pilbeam 2006). However, some of the foliar Mg

concentrations found in young trees (Figure 4.5) are considered low relative to the adequate

reference range values (0.20-0.37%) proposed for teak in the literature (Drechsel and Zech

1991; Boardman et al. 1997); in fact, during the dry season in the northern region of Costa Rica,

symptoms of foliar Mg deficiency are common. Hence, the translocation of Mg from the bole

and branches and the increase in the foliar Mg concentration can be considered as mechanisms

to achieve an adequate Mg level to ensure plant productivity.

The P concentration was higher in the foliage than in the other tissues analyzed, as

previously reported by Nwoboshi (1984), showing no tendency associated with tree age. The

average foliar P concentration lies at the lower limit of the reference range (0.14-0.25%) for

adequate values reported in the literature for teak (Drechsel and Zech 1991; Boardman et al.

1997), although many of the foliage samples showed values lower than this reference. S was

found to be concentrated mainly in the teak foliage, showing values in the lower end of the

reference range (0.11-0.23%) considered as adequate for teak (Drechsel and Zech 1991;

Boardman et al. 1997).

79

The tissue Fe and Mn concentrations showed high variability, probably due to sample

contamination with soil during fieldwork; however, prior to statistical analysis, the values

determined as too high were considered as outliers and removed from the dataset. This

contamination was most noticeable in the bole, the branches and especially the bark, where the

Fe and Mn concentration declines with tree age were probably caused by a decrease in the

proportion of contaminated vs. properly collected samples as the biomass increased. The foliar

Fe (58-390 mg kg-1

) and Cu (10-25 mg kg-1

) concentrations were within the ranges considered

adequate for teak, while the foliar Mn (50-112 mg kg-1

) values were below it (Drechsel and

Zech 1991; Boardman et al. 1997).

The foliar Zn concentration lies within the range (20-50 mg kg-1

) considered as adequate by

other authors (Drechsel and Zech 1991; Boardman et al. 1997), although Zn is usually deficient

in plants growing in highly weathered soils (Barker and Pilbeam 2006), such as the ones in our

study areas. The overall average among the locations showed low available soil Zn; this low

average was influenced by the low values found in the northern region of Costa Rica, as the soil

Zn values in Guanacaste and Panama are considered adequate (Table 4.1). Montero (1999)

reported an increase in the foliar Zn concentration with age, in contrast with the lack of a

relationship found in this study, where the variability of the data was high. The foliage B

concentration lies within the range (15-45 mg kg-1

) considered adequate by other authors

(Drechsel and Zech 1991; Boardman et al. 1997), which is higher than the requirements

reported for other species (Lehto et al. 2010). Of all tested tissues, the highest B concentration

was in the bark, as has been reported for other tree species (Lehto et al. 2010).

Generally, relatively high values of microelements are required to maintain an appropriate

nutritional status in teak and to ensure forest productivity and sustainability, although little

attention has been paid to this issue in other studies of teak nutrition (Nwoboshi 1984; Totey

1992; Negi et al. 1995; Kumar et al. 2009). Tropical soils are usually characterized as highly

weathered soils that are rich in Fe or Mn but generally deficient in Zn, B, Cu, and Mo (Barker

and Pilbeam, 2006). B is commonly deficient in soils throughout the world and is difficult to

80

evaluate in routine soil fertility analyses (Lehto et al. 2010). Hence, special care should be taken

to evaluate the B and Zn status in planted forests throughout the tropics.

Tissue nutrient concentrations, especially those for foliage, are considered to be a

management tool for evaluating the nutritional status of planted trees (Mead 1984; Drechsel and

Zech 1991; Barker and Pilbeam 2006; Lehto et al. 2010). Foliar concentrations have been

reported to be remarkably useful for this purpose because they are sensitive indicators of

nutritional deficiencies due to their direct relation with productivity, as foliage is where

photosynthesis takes place (Mead 1984). Table 4.2 summarizes the models and values that we

put forth for consideration as adequate concentration reference levels to be used in nutrient

management of planted teak forests in Central America. By selecting dominant and co-dominant

trees within well-managed and highly productive plantations, we sampled trees with an

appropriate nutritional status and higher nutrient requirements than average, so if the plantations

are managed to ensure the aforementioned levels, and if fertilizers are added accordingly, the

trees would have good nutritional status. Hence, the nutrient concentration values found in this

study can be taken as a reference to evaluate the nutritional status of similar teak plantations in

the region, i.e., as teak nutrition guidelines for Central America.

Decreases in N concentration with tree age are considered to be either a cause or a

consequence of the decline in productivity associated with increasing tree age; the K and Mg

concentrations could also be related to this phenomenon, which is a key issue in applied

ecology. Future research about these relationships should be performed with the aim of

achieving higher growth rates throughout the rotation period, which would allow shorter cycles

to be used.

CAPÍTULO 5

NUTRIENT ACCUMULATION AND EXPORT IN

TEAK (Tectona grandis l.f.) PLANTATIONS OF

CENTRAL AMERICA

83

5.1. Introduction

Due to growing needs for timber and wood products, forest plantations have increased

around the world and have gained economical relevance. At the same time there has been an

increase in concerns regarding the sustainability of planted forests, especially those managed

under a regime of short rotations (eg. Nambiar 1995). In particular, the relationship between

forest nutrition and sustainable timber production has become an important issue for the

management of less studied species in countries such as Costa Rica and China (eg. Ma et al.

2007, Arias et al. 2011, Quiong et al. 2011).

Teak (Tectona grandis L.f.) plantations have been widely established in Central America,

initially in Costa Rica and Panama (De Camino et al. 2002) and more recently in Guatemala, El

Salvador and Nicaragua. Teak has become an important species in the worldwide quality

tropical hardwood sector (e.g. Pandey and Brown 2000), with a total planted area of 4.3 ·106 ha

(not including the natural areas), of which 132,780 ha are in Central America (3%) and 86,500

in Panama and Costa Rica (Kollert and Cherubini 2012). In contrast to the rotations of 40-80

years that are used in Asia and Africa, in Central America the species is intensively managed in

rotations of 20-25 years, usually in carefully selected productive sites, with an expected

commercial industrial volume of 10 m3 ha

-1 year

-1 (Pandey and Brown 2000, De Camino et al.

2002). In this kind of short rotation intensively managed forest plantations, nutrient

management is a key issue for attaining sustainability and maintaining yields for future rotations

(Poels 1994, Evans and Turnbull 2004). Appropriate knowledge regarding teak nutrition is

required to improve plantation management and to attain high productivity and sustainability.

Nutrient accumulation increases as stands age, mainly induced by biomass accumulation;

however, nutrient uptake during early years is considered crucial to sustain the high growth

rates and the rapid expansion of both crown and roots required to maintain an appropriate

nutritional status throughout the entire rotation length (eg. Miller 1981, Laclau et al. 2003). In

general, foliage is the tree tissue with the highest nutrient concentration and it is considered to

contain 20-40% of total stand nutrients; while tree stems are assumed to have relatively low

concentrations of nutrients (eg. Miller 1984, 1995). However, the high amount of biomass

84

accumulated in the tree stem makes it an important sink of nutrients and, as a consequence, loss

of nutrient through wood removal at harvesting is a major cause of nutrient impoverishment of

forest sites (eg. Fölster and Khanna 1997; Worrel and Hampson 1997). While N, P and Mg are

mainly accumulated in the tree stem, bark and roots are considered to be Ca sinks (Nwoboshi

1984). Nutrient uptake depends mainly on the demand of the species and its ability to access

nutrients as well as the potential of the site, especially the soil, to supply nutrients. In calcareous

soils in India, the most absorbed nutrients by teak were Ca>K>N>Mg>P=S (Negi et al. 1995),

while they were K>N>Ca>>Mg≥P in less fertile soils in Africa (Nwoboshi 1984) and

N>Ca>K>Mg>P>Na>S>Cl in a different study site in India (Kumar et al. 2009).

It has been long recognized that, in order to understand the relationship between soil and

forest nutrition, it is first necessary to evaluate the quantities of nutrients taken up by the

growing forest and removed from the site during timber extraction (Rennie 1955). However,

Fölster and Khanna (1997) pointed out a traditional and general lack of concern of this problem

in planted forests. Soil-plant relation research in agriculture has traditionally analyzed when,

where, and at what rates nutrients are accumulated by plants in order to accomplish an efficient

and environmentally acceptable nutrient management (eg. Sadler and Karlen 1995, Bertsch

1998). This kind of study (nutrient absorption curves or nutrient accumulation dynamics with

age) traditionally used in agriculture, are considered to be promising tools to analyze forest

nutrition in intensively managed planted forests in the tropics (Ranger et al. 1995, Alvarado

2012 a). In order to provide information for the near-maximum accumulation rates of any given

crop, nutrient absorption studies are carried out in sites where near-maximum yield of the crop

is achieved (eg. Sadler and Karlen 1995, Bertsch 1998). By analyzing nutrient accumulation in

the most productive soils without nutrient deficiencies (and in dominant or co-dominant trees in

a site), the maximum species requirements are assessed. Therefore, if the minimum inputs

calculated for these high-fertility sites are applied in sites of lower fertility where tree nutrient

uptake would be lower, the sustainability of the plantation is still achieved (Bertsch 1998,

Alvarado 2012 a).

85

Nutrient accumulation dynamics of a species can be used to estimate (i) the nutrient removal

by thinning or harvesting, (ii) the maximum nutrient absorption of the species during one

rotation, (iii) the amount of nutrients left at a site after harvesting, which will recycle and be

reused during the next rotation, and (iv) the minimum nutrient inputs (fertilizers) the system

requires to be sustainably managed (Ranger et al. 1995, Bertsch 1998, Alvarado 2012 a). Hence,

in order to assess the nutrient sustainability of teak plantations, we conducted a study to

measure the amount of nutrients accumulated by the trees and exported during wood harvest, by

analyzing nutrient accumulation dynamics at different ages and the allocation patterns in highly

productive teak plantations in Central America (Costa Rica and Panama) throughout a rotation

period.

5.2. Material and methods

Study sites

Three teak (Tectona grandis L.f.) plantations were studied in Central America: two in Costa

Rica (Guanacaste and northern region) and one in Panama (Panama Canal watershed) (Figure

5.1). The three areas are classified as tropical wet forest according to Holdridge’s life zones

(Holdridge 1967), with similar mean annual rainfall (2500–3100 mm), although in Guanacaste

the dry season is longer than at the other two sites. The soils of the study areas are also similar,

although the northern region of Costa Rica is less fertile than the other sites and it has higher

soil acidity (Table 5.1).

Figure 5.1. Locations of

the study teak (Tectona

grandis L.f.)

plantations: Guanacaste

(Costa Rica); northern

region (Costa Rica) and

Panama Canal

Watershed (Panama).

86

Table 5.1. Summary of soil properties at the different study areas; means and coefficients of variation (in

parentheses) are provided. Soil information was only available for 23 of the 28 sampled stands.

Northern region,

Costa Rica

(n=11)

Guanacaste,

Costa Rica

(n=9)

Canal Zone,

Panama

(n=3)

TOTAL

(n=23)

pH 5.11 (6) 5.90 (6) 6.70 (12) 5.63 (12)

Acidity (cmol(+) L-1

) 0.70* (5) 0.31 (30) 0.15 (33) 0.48 (81)

Ca (cmol(+) L-1

) 4.45 (44) 21.36 (28) 20.97 (38) 13.22 (74)

Mg (cmol(+) L-1

) 1.46 (47) 6.89 (54) 5.25 (64) 4.08 (89)

K (cmol(+) L-1

) 0.13* (109) 0.33 (87) 0.36 (82) 0.24 (101)

ECEC (cmol(+) L-1

) 6.74 (31) 28.90 (32) 26.72 (40) 18.02 (71)

Acidity Saturation (%) 11.96* (84) 1.22 (58) 0.65 (55) 6.28* (139)

P (mg L-1

) 3* (114) 3* (146) 2* (0) 3* (124)

Zn (mg L-1

) 2* (84) 3 (58) 3 (107) 2* (77)

Cu (mg L-1

) 8 (19) 11 (85) 4 (83) 9 (71)

Fe (mg L-1

) 165 (23) 37 (81) 65 (154) 102 (75)

Mn (mg L-1

) 43 (171) 38 (82) 19 (101) 38 (142)

Organic matter (%) 4.6 (27) 3.8 (28) 4.6 (16) 4.3 (27)

Sand (%) 24.9 (23) 23.4 (56) 29.0 (31) 24.8 (38)

Silt (%) 18.4 (13) 36.9 (42) 36.8 (43) 28.0 (51)

Clay (%) 56.7 (11) 39.7 (22) 34.3 (50) 47.1 (27)

ECEC: effective cation exchange capacity. * = Values outside the adequate reference soil levels (Bertsch

1998).

The stands studied were chosen to be representative of properly managed teak plantations in

Central America. In general, management of these plantations consists on continuous

silvicultural activities: weed control, pruning, thinning regimen (approximately from 800-1000

trees ha-1

at establishment to 150-200 trees ha-1

at final felling) and fertilization during the

establishment. The use of clones is common in recent years. A commercial volume of 100-150

m3 ha

-1 is expected for this kind of plantation after approximately 20 years rotation.

Field sampling and design

False time series (chronosequences) method was used to analyze nutrient accumulation

dynamics of teak trees from 1 to 19 years of age. Johnson and Miyanishi (2008) define this

method as an inference of a time sequence of development made from a series of plots or stands

differing in age. Despite of the critiques of this method (Johnson and Miyanishi 2008), it was

considered to be valid as all the studied stands are considered to be similar in environmental

conditions (soil and climate) and management practices. Hence, stands of different ages

87

(between 1 and 19 years old) were studied and it is assumed that they represent the average

time-pattern of the plantations analyzed.

A total of 28 stands were analyzed, seven in Panama, 12 in the northern region of Costa Rica

and nine in Guanacaste (Costa Rica). In order to analyze a maximum yield research experiment

(Sadler and Karlen 1995, Bertsch 1998, Alvarado 2012 a), dominant and codominant trees were

selected: (a) without visible symptoms of diseases or nutritional deficiencies and (b) that were

representative of the best-performing trees of the plantations, assuming optimal nutrition and a

full expression of genetic potential. In stands younger than ten years of age, two trees were

sampled per stand, whereas only one tree was sampled in older stands. Trees were felled and

biomass accumulated at different tree components (bole, bole’s bark, foliage and primary and

secondary branches) was weighed in the field. A random subsample (1 kg per tissue per tree)

was then carefully collected representative homogeneous mixture of the whole sample of each

tissue. More detail can be found in Fernández-Moya et al. (2013). Dry biomass (hereafter

biomass) was then calculated based on field measurements and sample water content estimated

in laboratory. All the fieldwork was performed during July-September, at the tree’s optimal

nutritional status during the period of maximum growth activity, to avoid effects of seasonality.

Tissues samples analyzed at the “Centro de Investigaciones Agronómicas” of the University

of Costa Rica (hereafter CIA-UCR) to determine nutrient concentrations (N, P, Ca, Mg, K, S,

Fe, Mn, Cu, Zn and B, hereafter referred to as nutrients) after samples were dried and water

content was assessed. Dry combustion was used to measure the N concentration, and wet

digestion and atomic spectrometry were used to extract and determine other nutrients (Bertsch

1998). Table 2 summarizes nutrient concentration age dynamics (more detail can be found in

Fernández-Moya et al. 2013).

Nutrient accumulation and allocation in tissues biomass were the objective variables of the

present work. Those were calculated by multiplying nutrient concentration (Table 2, Fernández-

Moya et al. 2013) by biomass. Primary and secondary branches were weighted averaged and are

reported as “branches” {Nutrient i accumulation (branches) = [Biomass (primary branches) ·

Nutrient i concentration (primary branches) + Biomass (secondary branches) · Nutrient i

88

concentration (secondary branches)] · (Biomass (primary branches) + Biomass (secondary

branches))-1

}. Bole and bark nutrient accumulation was also weighted averaged and reported as

“bole and bark” {Nutrient i accumulation (bole and bark) = [Biomass (bole) · Nutrient i

concentration (bole) + Biomass (bark) · Nutrient i concentration (bark)] · (Biomass (bole) +

Biomass (bark))-1

}. Similarly, “total” nutrient accumulation represents a weighted average from

all the sampled tissues.

As no detailed information about the thinning regimen and the trees density dynamics with

age was available, in order to upscale individual tree measurements to estimate stand’s values,

tree stocking at different stands age were considered as 1000, 300 and 150 trees ha-1

at 1-5, 10

and 19 years respectively. These values are considered as average values normally used in

plantations in Central America. Even though plant density influence tree nutrient uptake

because of competition for soil nutrients, we consider plan density for a given age as relatively

homogeneous between the three study sites as the three companies follow similar management

patterns.

In order to estimate soil nutrient availability, topsoil samples were collected (0-20 cm),

where more than half of the teak roots are situated (Srivastava et al. 1986, Behling 2009). Five

soil sub-samples were taken from each site (without litter) and mixed into one composed soil

sample for each site. Soil information was only available for 23 of the 28 sampled stands (Table

1). Soil samples were collected during at the same time of tree components specified before.

Soil samples were analyzed at CIA-UCR to determine: pH, Ca, Mg, K, acidity and Al, P, Fe,

Cu, Zn, Mn. pH was determined in water 10:25; acidity, Al, Ca and Mg in KCl solution 1M

1:10; P, K, Zn, Fe, Mn and Cu in modified Olsen solution pH 8,5 (NaHCO3 0,5 N, EDTA

0.01M, Superfloc 127) 1:10. Organic matter was determined by combustion method as

described by Horneck and Miller (1998). Soil texture was determined using the modified

Bouyoucos method, as described by Forsythe (1975).

Teak roots accounts for 5-30% of total tree nutrient accumulation (Ola-Adams 1993,

Siddiqui et al. 2007, Behling 2009) so total planted teak forests accumulation can be estimated

as 105-130% of the above mentioned for the aboveground biomass. Belowground biomass was

89

not taken into account in this study because it can be considered as left at the site after final

harvesting, mineralized and used by the next rotation (recycled).

Statistical analysis

Generalized linear mixed models (GLMMs) were used to study the relationships between

nutrient (N, P, Ca, Mg, K, S, Fe, Mn, Cu, Zn and B) accumulation in each tissue (bole, bark,

bole and bark, branches, foliage and total) and tree age. The use of GLMMs was required, as

most of the study variables did not approach the normal distribution hypothesized in traditional

models. The probability distribution of each of the 72 response variables analyzed was studied

prior to construct the GLMMs. The exponentially distributed variables were modeled using a

Gamma distribution approach with α=1.

To evaluate the most suitable model for each study variable, a total of 83 different models

were constructed, selecting the one with lowest deviance. Three groups of models were

constructed: (1) a null model considering only an intercept [yi = b0]; (2) a model considering an

intercept in addition to age as an explanatory variable [(yi) λ = age + b0]; and (3) a model without

an intercept [(yi) λ

= age]. For groups (2) and (3), 41 different power link functions [g(μ)=μλ]

were tested for each one, with λ varying between λ=2 to λ=-2 and a λgap=0.1. When no model

including age as a parameter was statistically significant, or when the data did not follow any of

the studied distribution functions, the resulting model included only an intercept representing

the mean of the variable, and no age effect was taken into account.

The sampled stands in each study area were spatially correlated. The spatial correlation was

taken into account by including a random effect for the study area, modeling the working

correlation matrix with a first-order autoregressive structure. The goodness-of-fit of the models

was estimated by measuring the percentage difference between the deviance of the model and

the deviance of a model with no covariates (hereafter referred to as efficiency, EF), which is a

pseudo-R2 measure reported for GLMMs. All statistical analyses were performed using SAS 9.0

(SAS Institute Inc. 2002). All statistical tests throughout the text are considered significant with

α=0.05.

90

5.3. Results

Aboveground biomass allocation

The fitted models showed an estimated total aboveground biomass of 87 kg, 277 kg and 807

kg per tree at 5, 10 and 19 years of age, respectively (Table 5.3, Figure 5.2). Bole was the tissue

where most biomass accumulated, accounting for 60% of the total tree biomass at age 5, 10 and

19 (Table 5.3, Figure 5.2). Branches also accounted for large amounts of biomass compared to

total tree biomass but it increased with tree age: 19%, 24% and 30% at 5, 10 and 19 years

respectively (Table 5.3, Figure 5.2). However, foliage and bark percentage of total tree biomass

decreased with age although the net biomass increased (Table 5.3, Figure 5.2): (i) Bark biomass

was 8, 19 and 42 kg tree-1

accounting for 9%, 7% and 5% of the total tree biomass at 5, 10 and

19 years, respectively; (ii) foliage biomass was 10, 17 and 29 kg tree-1

accounting for 11%, 6%

and 4% of the total tree biomass at 5, 10 and 19 years, respectively.

Figure 5.2. Tree tissues biomass accumulation (kg tree-1

/ g tree-1

) related to tree age (years) in teak

plantations (Tectona grandis L.f.). Points represent sampled trees at different locations: Guanacaste ( );

Northern Region of Costa Rica ( ); Panama ( ). Lines represent fitted models (Table 5.3).

91

Table 5.3. Estimation of the relation between biomass accumulation (kg) and tree age (years) in 1 to 19

year old teak (Tectona grandis L.f.) plantations in Costa Rica and Panama. Below specified models are in

the form [y = (b1 · age)1/λ

], where the response variables (y) are the biomass accumulation in tree tissues

Tissues Model

Biomass accumulation

5 years 10 years 19 years

b1 b1 [Std. error] λ EF (%) kg* % of Total kg* % of Total kg* % of Total

Foliage 3.0515 0.2845 1.2 69 10 (8, 11) 11.1 17 (15, 20) 6.2 29 (25, 34) 3.7

Bark 1.0461 0.0258 0.8 49 8 (7, 8) 9.1 19 (18, 20) 6.8 42 (39, 45) 5.2

Bole 2.1629 0.1062 0.6 91 53 (45, 62) 60.7 168 (142, 196) 60.7 489 (413, 570) 60.7

Bole and bark 2.3399 0.1032 0.6 91 60 (52, 69) 69.1 191 (165, 220) 69.1 558 (480, 641) 69.1

Branches 0.8106 0.0423 0.5 82 16 (13, 20) 18.8 66 (53, 80) 23.7 237 (191, 288) 29.4

Total 2.9196 0.1275 0.6 91 87 (75, 100) 277 (238, 317) 807 (695, 925)

EF (%): Model efficiency, pseudo R2 estimate for Generalized Linear Mixed Models.

* Biomass estimated for the different ages with the Confidence Interval between parenthesis (α=0.05).

Nutrient accumulation and allocation

The fitted models (Table 5.4, Figures 5.3 and 5.4) allow to estimate the nutrient

accumulation at different tree tissues of teak trees based on their age. N mainly accumulated in

foliage during the first years although bole was the most important sink after 6-7 years as bole N

accumulation increased sharply with age while foliage N accumulation increased slowly with

age. Ca accumulated in branches following a tendency close to that of bole and bark. Bark Ca

accumulation was higher than in foliage or bole from 10 years old on, while foliage was higher

before that. K accumulated in branches in trees older than 6-7 years whereas in younger trees it

mainly accumulated at tree bole; bole K accumulation was very similar to that of bark and they

were both slightly higher than in foliage. Mg accumulated in tree bole throughout the entire

rotation, accumulation in branches was also high; bark and foliage Mg accumulation were

similar and lower than in the other tissues. P and S also accumulated at tree bole and branches

while bark and foliage accumulation was low. Fe accumulated at tree bole and branches while

foliage accumulation was lowest following a similar pattern to that of bark. Mn also

accumulated in tree branches although foliage accumulation was high and even higher than in

branches at trees younger than 8-9 years. Cu mainly accumulated in tree bole while bark

accumulation was lowest and foliage accumulation was higher at trees younger than 5 years old.

Zn accumulated in tree bole and branches while bark and foliage accumulation was low. B

92

mainly accumulated in tree branches although bark accumulation was high compared to foliage

accumulation, which was lowest.

Figure 5.3. Tree foliage nutrient accumulation

(kg tree-1

or g tree-1

) related to tree age (years) in

teak plantations (Tectona grandis L.f.). Points

represent sampled trees at three different

locations: Guanacaste, Costa Rica ( ); Northern

Region, Costa Rica ( ); Panama ( ). Lines

represent fitted models (Table 5.4).

93

Figure 5.4. Tree bole and bark nutrient

accumulation (kg tree-1

or g tree-1

) related to tree

age (years) in teak plantations (Tectona grandis

L.f.). Points represent sampled trees at three

different locations: Guanacaste, Costa Rica ( );

Northern Region, Costa Rica ( ); Panama ( ).

Lines represent fitted models (Table 5.4).

94

Table 5.4. Regressions between tissues’ nutrient accumulation and tree age (years) in 1 to 19 year old

teak (Tectona grandis L.f.) plantations in Costa Rica and Panama. Below specified models are in the form

[y = (b0 + b1 · age)1/λ

], where the response variables (y) are the nutrient accumulation in tree tissues.

When no model including age as a parameter was statistically significant, the model only included an

intercept (b0) representing the mean of the variable.

Tissues Macronutrient

(kg) b0 b1 b1 [Std. error] λ EF (%)

Micronutrient (g)

b0 b1 b1 [Std.

error] λ

EF (%)

Foliage

N 0.0218 0.0022 1.4 55 Fe 0.2088 0.0149 0.9 66

Ca 0.0205 0.0003 1.1 66 Mn 0.0522 0.0100 1.6 39

K 0.0018 0.0002 2.0 41 Cu 0.0062 0.0011 1.4 47

Mg 0.0049 0.0004 1.0 74 Zn 0.0452 0.0029 1.3 60

P 0.0009 0.0001 1.3 63 B 0.0240 0.0040 1.3 60

S 0.0006 0.0001 1.3 58

Bark

N 0.0186 0.0002 0.8 43 Fe 0.4200 0.084 1.2 23

Ca 0.0545 0.0041 0.5 47 Mn 0.0352 0.0019 1.1 33

K 0.0109 0.0011 1.2 35 Cu 0.0081 0.0010 0.7 39

Mg 0.0054 0.0002 0.9 43 Zn 0.0564 0.0073 0.9 39

P 0.0116 B 0.0632 0.0026 0.6 50

S 0.0017 0.0001 0.9 43

Bole

N 0.0574 0.0028 0.7 89 Fe 0.4318 0.0751 0.6 51

Ca 0.0354 0.0014 0.6 90 Mn 0.0960

K 0.0079 0.0013 1.6 45 Cu 0.0434 0.0045 0.5 71

Mg 0.0236 0.0012 0.7 82 Zn 0.9382

P 0.1133 B 0.0526 0.0060 0.7 75

S 0.0209 0.0038 0.6 67

Bole and

bark

N 0.0688 0.0027 0.7 89 Fe 0.500 0.0575 0.6 60

Ca 0.0767 0.0018 0.6 82 Mn 0.0447 0.0009 1.1 45

K 0.0265 0.0052 1.3 52 Cu 0.043 0.005 0.6 53

Mg 0.0288 0.0016 0.7 84 Zn 1.3757

P 0.1249 B 0.0949 0.0062 0.6 77

S 0.0234 0.0037 0.6 72

Branches

N 0.0412 0.0023 0.7 71 Fe 0.3773 0.0108 0.6 63

Ca 0.0761 0.0052 0.5 83 Mn 0.0836 0.0083 0.6 62

K 0.0401 0.0032 0.7 73 Cu 0.0274 0.0023 0.8 66

Mg 0.0210 0.0012 0.6 80 Zn 0.0910 0.0104 0.7 58

P 0.0391 B 0.0753 0.0044 0.6 80

S 0.0142 0.0016 0.6 73

Total

N 0.1165 0.0044 0.8 87 Fe 0.9331 0.0742 0.7 72

Ca 0.1281 0.0024 0.6 91 Mn 0.2045 0.0144 1.2 54

K 0.0676 0.0091 1.3 93 Cu 0.0711 0.0085 0.8 68

Mg 0.0425 0.0013 0.7 89 Zn 0.2692 0.0608 0.9 53

P 0.02180 B 0.1657 0.0060 0.6 85

S 0.0254 0.0033 0.7 82

EF (%): Model efficiency, pseudo R2 estimate for Generalized Linear Mixed Models.

95

In the first years, total nutrient accumulation showed a tendency as P > K > N > Ca > Mg =

S > Fe > Mn > Zn > B > Cu; the fifth year it was N > Ca > K > P > Mg > S > Fe > Zn > Mn > B

> Cu; the tenth year it was Ca > N > K > Mg > P > S > Fe > Zn > B > Mn > Cu; and the

nineteenth year it accumulated Ca > N > K > Mg > S > P > Fe > B > Zn > Mn > Cu. The

estimation of the accumulation of some nutrients (mainly P) may not be well represented as it is

based on an average of sampled trees from all ages, as no statistical model could be fitted with

age as explanatory variable (Table 5.4). The estimated total tree nutrient accumulation at

approximately one rotation period (19 years old) is 2.7 kg N, 4.4 kg Ca, 1.2 kg K, 700 g Mg,

200 g P, 400 g S, 61 g Fe, 3 g Mn, 1 g Cu, 6 g Zn and 7 g of B tree-1

(Table 5.5).

Nutrient export

Timber extraction as thinning or final harvesting implies export of nutrients allocated at bole

and bark. Nutrient export varied from thinning to final harvesting as tree bole and bark nutrient

allocation varied with age, as well as tree stocking (Table 5.6). Final harvesting constitutes a

major nutrient output from the system, considering that 19 year old tree bole and bark nutrient

accumulation was: 1.9 kg N, 1.5 kg Ca, 600 g K, 300 g Mg, 400 g P, 200 g S, 43 g Fe, 0.9 g

Mn, 0,7 g Cu, 1.4 g Zn and 2.7 g B tree-1

(Table 5.6).

5.4. Discussion

Aboveground biomass allocation

Reported tree and stand biomass are similar (Pérez and Kanninen 2003) or higher (Kaul et

al. 1979, Kumar 2009, Kumar et al. 2009) than those reported in other studies. This tendency

matches was expected because only the best performing trees in each sampled stand were

selected. However, stand biomass is low compared to that reported by a spacing trial in south-

western Nigeria (Ola-Adams 1993). Most of tree biomass is accumulated in the tree bole,

accounting for 60% by itself and 69% of total tree biomass when considering bole and bark

biomass, similar results to those were observed by other authors (Kaul et al. 1979, Pérez and

Kanninen 2003).

96

Table 5.5. Nutrient accumulation in 1 to 19 year old teak (Tectona grandis L.f.) plantations in Costa Rica

and Panama. Total nutrient accumulation is estimated from the statistical models summarized at Table 4

and represents the sum of the nutrients accumulated in bole, bark, branches and foliage. Individual tree

estimations (Table 5.4) are used to calculate stand values assuming the following densities: 1000, 300 and

150 trees ha-1

at 1-5, 10 and 19 years respectively.

Nutrient 5 yr 10 yr 19 yr

nutrient tree-1 * nutrient ha-1 nutrient tree-1 * nutrient ha-1 nutrient tree-1 * nutrient ha-1

N (kg) 0.51 (0.46, 0.56) 508.89 1.21 (1.10, 1.32) 363.10 2.70 (2.45, 2.95) 404.99

Ca (kg) 0.48 (0.45, 0.51) 475.92 1.51 (1.42, 1.60) 453.28 4.40 (4.14, 4.68) 660.59

K (kg) 0.43 (0.34, 0.52) 434.14 0.74 (0.58, 0.89) 221.98 1.21 (0.96, 1.45) 181.85

Mg (kg) 0.11 (0.10, 0.12) 109.42 0.29 (0.27, 0.32) 88.36 0.74 (0.67, 0.80) 110.52

P** (kg) 0.22 217.97 0.22 65.39 0.22 32.70

S (kg) 0.05 (0.03, 0.07) 52.45 0.14 (0.09, 0.20) 42.35 0.35 (0.23, 0.49) 52.98

Fe (g) 9.03 (7.09, 11.10) 9027.53 24.30 (19.08, 29.89) 7290.09 60.79 (47.72, 74.76) 9118.48

Mn (g) 1.02 (0.90, 1.13) 1018.72 1.82 (1.60, 2.02) 544.54 3.10 (2.74, 3.45) 464.83

Cu (g) 0.27 (0.20, 0.36) 274.50 0.65 (0.47, 0.85) 195.87 1.46 (1.04, 1.89) 218.46

Zn (g) 1.39 (0.73, 2.09) 1391.18 3.01 (1.57, 4.52) 901.53 6.13 (3.2, 9.21) 919.77

B (g) 0.73 (0.65, 0.82) 730.84 2.32 (2.05, 2.60) 696.08 6.76 (5.98, 7.58) 1014.42

* Biomass estimated for the different ages with the Confidence Interval between parenthesis (α=0.05).

** No statistically sound model could be fitted between total P accumulation and age (Table 5.4).

Therefore, estimated P accumulation is based in the average from the sampled trees; all collected data was

taken into account, no difference could be made based on tree age.

Tree biomass increases sharply with tree age (Table 5.3, Figure 5.2), whereas stand biomass

shows fast growth during an initial establishment stage (from 6 to 87 Mg ha-1

at years 1 and 5,

respectively) and a smaller increment afterwards (from 87 to 121 Mg ha-1

at years 5 and 19,

respectively). Little stand biomass variation with tree age or spacing has been shown for teak

and other species (Ola-Adams 1993, Pérez and Kanninen 2003, Peri et al. 2008). Tree foliage

biomass increase with tree age reaching 29 kg tree-1

when trees are 19 years old, whereas its

contribution to total tree biomass decreases from 11% to 6% and 4% at 5, 10 and 19 years of

age respectively, a decreasing tendency showed also in other studies (Pérez and Kanninen 2003,

Kumar 2009). Stand foliage biomass also decreased from 9.7 Mg ha-1

in the fifth year to 4.4 Mg

ha-1

at age 19, which could be related to declining stand growth capacity – age-related decline

in productivity – (Gower et al. 1996, Ryan et al. 1997, Binkley et al. 2002). This also coincides

with the declining tendency shown between leaf biomass and tree spacing (Ola-Adams 1993),

as tree spacing increases with age in the sampled plantations due to the thinning regimen.

97

Table 5.6. Nutrient export by timber extraction in 1 to 19 year old teak (Tectona grandis L.f.) plantations

in Costa Rica and Panama, and comparison between nutrient export and total accumulation (Table 5.5) at

19 yr old plantations. Nutrient export is estimated from the statistical models summarized in Table 5.4

and represents the sum of the nutrients accumulated in bole and bark. Individual tree estimations (Table

5.4) are used to calculate stand values assuming the following densities: 1000, 300 and 150 trees ha-1

at 1-

5, 10 and 19 years respectively.

Nutrient export (bole and bark) Total accumulation

at 19 yr old

plantations (nutrient ha-1)**

Nutrient export

compared to total

accumulation at 19 yr old plantations (%)

5 yr 10 yr 19 yr

nutrient tree-1 * nutrient tree-1 * nutrient tree-1 * nutrient ha-1

N (kg) 0.22

(0.19, 0.24)

0.59 (0.52, 0.65)

1.47

(1.31, 1.63) 219.93 404.99 54

Ca (kg) 0.20

(0.19, 0.22)

0.64

(0.59, 0.69)

1.87

(1.73, 2.02) 280.98 660.59 43

K (kg) 0.21

(0.15, 0.27)

0.36 (0.25, 0.46)

0.59

(0.41, 0.76) 88.48 181.85 49

Mg (kg) 0.06

(0.05, 0.07)

0.17

(0.14, 0.20)

0.42

(0.36, 0.49) 63.39 110.52 57

P*** (kg) 0.15 0.15 0.15 23.02 32.70 70

S (kg) 0.03

(0.02, 0.04)

0.09

(0.05, 0.14)

0.26

(0.14, 0.41) 38.85 52.98 73

Fe (g) 4.61

(3.01, 6.46)

14.62 (9.55, 20.52)

42.61

(27.84, 59.8) 6391.91 9118.48 70

Mn (g) 0.26

(0.25, 0.27)

0.48

(0.46, 0.50)

0.86

(0.83, 0.89) 129.3 464.83 28

Cu (g) 0.08

(0.05, 0.11)

0.24 (0.16, 0.34)

0.71

(0.46, 1.01) 107.1 218.46 49

Zn*** (g) 1.41 1.41 1.41 211.9 919.77 23

B (g) 0.29

(0.23, 0.35)

0.92

(0.73, 1.12)

2.67

(2.13, 3.27) 400.67 1014.42 39

* Nutrient accumulation estimated for the different ages with the Confidence Interval between parenthesis

(α=0.05). ** Total accumulation represents the sum of the nutrients accumulated in bole, bark, branches

and foliage of a mature stand (19 yr) near to harvesting (Table 5.5). *** No statistically sound model

could be fitted between bole and bark P and Zn accumulation and age (Table 5.4). Hence, estimated bole

and bark P and Zn accumulation are based in the average from the sampled trees; all collected data was

taken into account, no difference could be made based on tree age.

Nutrient accumulation and allocation

N accumulates mainly at foliage during the first years because N foliage concentration is

high and foliage is an important component of tree biomass at the beginning of the rotation;

however, as foliage becomes a less important tree component and N foliage concentration

decreases (Table 5.2, Fernández-Moya et al. 2013), the bole becomes the most important N sink

in plantations older than 6-7 years old, as it is an important biomass sink. A similar pattern is

followed by Ca as it is mostly accumulated in the foliage in young trees (same as reported by

Ola-Adams 1993) where the bark biomass is low and then is mainly accumulated in tree bark in

98

older plantations (and bole-and-bark consequently), which is the general pattern observed for

teak and other species (Nwoboshi 1984, Peri et al. 2008, Arias et al. 2011, Qiong et al. 2011).

Other elements (K, Mg, P, S, Fe, Cu, Zn and B) showed a tendency to accumulate mainly at tree

bole (or bole and bark) and branches at all ages, probably because tissue concentration of those

elements are lower (Table 5.2, Fernández-Moya et al. 2013) and so is the influence of nutrient

concentration in nutrient accumulation in biomass. In general, this is consistent with the results

of other studies such as Nwoboshi (1984) who reported that N, P and Mg mainly accumulate in

the tree stem, while bark is considered as Ca sink; or other authors who estimated that bole

wood is the main sink for all nutrients considered (Ola-Adams 1993, Kumar et al. 2009).

The relative importance of the different elements in tree nutrient accumulation also varies

with tree age. Young teak trees accumulate P > K > N > Ca > Mg = S > Fe > Mn > Zn > B >

Cu, whereas at the end of the rotation they have absorbed Ca > N > K > Mg > S > P > Fe > B >

Zn > Mn > Cu. However, the high P accumulation at young trees are probably overestimated by

the proposed model (Table 5.4) as we could not fit an appropriate model and the average for all

ages had to be used; for the same reason, P accumulation at later in the rotation is probably

underestimated (Table 5.4, Figure 5.3 and 5.4). In general, this is consistent with the results of

other studies where the nutrients most absorbed by teak were K > N > Ca >> Mg ≥ P

(Nwoboshi 1984), N = K > Ca > Mg > P (Ola-Adams 1993), Ca > K > N > Mg > P = S (Negi et

al. 1995), Ca > K > N > Mg > P > S (Behling 2009) or N > Ca > K > Mg > P > Na > S > Cl

(Kumar et al. 2009). Hence, the general pattern between diverse studies is that teak mostly

accumulates Ca, N and K. However, the most absorbed nutrient of those Ca, N and K varies

between studies, probably depending on soil availability at each particular study-site.

The proposed models (Table 5.4) allow managers to calculate the amount and allocation of

nutrients accumulated by a well performing tree at different ages. As an example, a stand of 150

trees ha-1

at age 19 would accumulate 405 kg N ha-1

, 661 kg Ca ha-1

, 182 kg K ha-1

, 111 kg Mg

ha-1

, 33 kg P ha-1

, 53 kg S ha-1

, 9 kg Fe ha-1

, 465 g Mn ha-1

, 218 g Cu ha-1

, 920 g Zn ha-1

, 1 kg B

ha-1

.

99

The nutrients accumulated (kg ha-1

) in 5 year old teak stands (Table 5.5) represent between

70 to more than 100% of nutrients accumulated in 19 year old plantations (Table 5.5), validating

Miller’s theory (Miller 1981, Miller 1984, Miller 1995) that once canopy closure occurs, if

nutrition has been appropriate, nutrient uptake decreases and nutrition is mainly based on

translocation between tissues of the same tree and nutrient recycling, although nutrient export

by thinning should be taken into consideration too. Alvarado (2012 b) estimates than more than

70% of N required by a forest plantation could be provided by residues mineralization and wet

and dry N deposition. Atmospheric N input could be estimated, based on reviewed data, as 230

kg ha-1

during 20 years of the rotation period of a planted forest (Fölster and Khanna 1997,

Alvarado 2012 b) which is approximately 57% of N accumulation at final stages of the rotation

(Table 5.5). However, stand N accumulation at earlier stages is higher (509 kg ha-1

at year 5,

Table 5.5) and atmospheric deposition during 5 years could be estimated as only 57.5 kg ha-1

(11% of estimated plantation accumulation) pointing out a possible N deficit at early plantation

stages that should be supplied by fertilizers if it could not be supplied by soil N. Atmospheric

inputs vary from site to site in a significant way but even though a general value from literature

cannot be used to close a nutrient balance, it point out the probable nutrient deficit in teak

plantations compared with nutrient accumulation and exports and, hence, the need to do more

detailed studies to design a nutritional plan including forest fertilization. Fertilization can be

considered as another nutrient input to the system but, at this moment, it is usually very low.

The low values of soil available K and extremely low in the case of P (Table 5.1) contrast

with the relatively high K and P accumulation in tree biomass (Table 5.5). This could be

explained just by the methodology used to estimate K and especially P topsoil availability or it

could be caused by one of the following hypothesis: a) the best performing trees may benefit

from a particular site condition with a soil nutrient availability higher than the average, which

allows them to have better growth, and/or maybe a deeper root system which allows them to

explore a larger soil volume; b) teak roots may produce phosphatases which improve the

mineralization rates of organic-P resulting in higher levels of available P than showed by soil

analysis (Corryanti et al. 2007); c) those elements could be limiting teak productivity; d)

100

nutrient input as atmospheric deposition could be playing a key role in plantation nutrition of

those element.

Well performed teak trees generally accumulate high quantities of N, Ca and K. As Ca soil

availability is high if site selection was appropriate, special attention should be paid to other

elements such as N, usually considered as limiting productivity of terrestrial ecosystems; and K,

which could become a problem due to its interaction with soil Ca and the usual high values of

the latter. In addition, P and B (Lehto et al. 2010) have also been identified as limiting in forest

soils and should be carefully considered. To a lesser extent other elements such as Mg showed

moderately high requirements and its interaction with Ca could be a problem in some

environments.

Nutrient export

The proposed models (Table 5.4) allow managers to estimate nutrient export by trees

removal at different plantation ages. Hence, if a thinning is performed in teak plantations of a

certain age and managers can estimate the number of removed trees, then they can estimate the

nutrient removal for this operation by simply multiplying number of trees and nutrient removal

per tree. Detailed thinning information was not available for the present study and hence there

are only estimations about nutrient export by final harvesting at the end of the rotation (Table

5.6).

Nutrient export by timber extraction at the end of the rotation represents approximately half

of the estimated tree nutrient accumulation of mature stands, varying between 23 and 73%,

depending on the element (Table 5.6). The other half of the nutrient absorbed by the tree could

remain at the site to be recycled and used during the following rotation, if an appropriate

residues management is done (Fölster and Khanna 1997). Timber extraction by final felling

constitutes a major nutrient output from the system, as harvesting (bole and bark) 150 trees ha-1

at age 19 would export 220 kg N ha-1

, 281 kg Ca ha-1

, 88 kg K ha-1

, 63 kg Mg ha-1

, 23 kg P ha-1

,

39 kg S ha-1

, 6 kg Fe ha-1

, 129 g Mn ha-1

, 107 g Cu ha-1

, 212 g Zn ha-1

, 401 g B ha-1

. Final felling

nutrient extraction could be of special relevance in cases such as P and K, because of their low

101

soil availability (Table 5.1), which could become a problem after several rotations. Estimated N

output at final felling (Table 5.6) is also high; however there is no information about soil N

content to allow us to speculate about the sustainability of this element. Reported data for

nutrient export is calculated as nutrient accumulated in bole and bark tissues; therefore, it is

only roughly estimating nutrient export, as a percentage of non-commercial stem is left at the

site after final felling.

Nutrient export repeated during several rotations could be a cause of soil nutrients depletion

(eg. Miller 1984, Fölster and Khanna 1997, Evans and Turnbull 2004) and it can be the cause of

the decrease in forest productivity after several rotations noticed long ago by many foresters

throughout history (Rennie 1955, Evans 2009). The need to replace (by fertilization) the nutrient

output has been traditionally ignored by forest managers (Fölster and Khanna 1997), although

FSC (2004) and several other authors (eg. Rennie 1955, Worrel and Hampson 1997)

recommend the application of fertilizer to sustain short-cycle plantation productivity. Some

authors report Plantation Stability Indices (Fölster and Khanna 1997, Arias et al. 2011) as a

good measure to evaluate soil nutrient mining by forestry plantations. We consider them as a

good indicator about more research need to be paid to, and especially that they should be used

to evaluate the sustainability of plantations in certification schemes or when they are included in

payment for environmental services programs.

Many authors (eg. Rennie 1955; Fölster and Khanna 1997; Ma et al. 2007) have proposed to

debark tree stems at the plantation site aiming to minimize nutrient exports although this could

be a non-profitable and time-consuming practice which may not be adopted by forests

companies. As approximately half of the nutrients translocate from leaves to other tree tissues

before senescence (Aerts 1996), bole and bark may have a higher nutrient concentration during

the dry season (defoliated teak period) when timber extraction is usually done. Hence, a detailed

study of tree nutrient translocation could result in the programming of final felling during a time

of the year when nutrient exports are minimal. If this would be confirmed and bole and bark

nutrient concentration would be higher in the dry season compared with the rainy season, that

102

would suppose an underestimation of the nutrient export reported in the present work. Hence,

future work should be necessary to take this possible influence into account.

The proposed models (Table 5.4) allow managers to calculate the amount of nutrients

accumulated in bole and bark tree biomass depending on tree age, and can be used to estimate

the nutrient extraction in different thinning management scenarios. Estimated nutrient export

(accumulated at bole and bark) by teak thinning or final felling (Table 5.6) should be a reference

of the minimum nutrient inputs the managers should add in via fertilization along a rotation

period, depending on the site’s soil and nutrient environmental inputs and outputs, thinning

regimen and other silvicultural variables. Nutrient cycling in teak plantations should be

analyzed in more detail in order to evaluate the sustainability of the system. This study serve as

a first general approach to the subject in Central America and reflects the needs to properly

analyze not only the direct output via timber harvest (including the output via thinning) but the

complete input and output factor which would allow us to close a nutrient balance for the

system.

CAPÍTULO 6

MODIFYING HARVESTING TIME AS A TOOL

TO REDUCE NUTRIENT EXPORT

BY TIMBER EXTRACTION

105

6.1. Introduction

Forest plantations have globally increased during recent decades to cover nowadays 264 ·

106 ha, 7% of global forest area, in response to the growing global demand for timber, pulp,

energy and other goods (Evans 2009; FRA 2010). Meanwhile, forest managers have been

increasingly concerned about the relationship between forest nutrition, soil management and

sustainable timber production, with the challenge of maintaining high productivity rates through

several rotations, especially in short-rotation plantations (e.g., Nambiar 1995; Fox 2000).

Repeated nutrient export by timber harvesting during several rotations could cause soil nutrients

depletion (eg. Miller 1984; Fölster and Khanna 1997; Worrel and Hampson 1997; Merino et al.

2005) and it can be the origin of the decrease in forest productivity after several rotations

noticed throughout history by many foresters (eg. Rennie 1955; Evans 2009). This need of

replacing the nutrient output (e.g. by fertilization) has been traditionally ignored by forest

managers (Fölster and Khanna 1997), although FSC (2004) and several other authors (eg.

Rennie 1955; Worrel and Hampson 1997; Merino et al. 2005) recommend the application of

fertilizer to sustain short-cycle plantation productivity. Although nutrient inputs (mainly wet

and dry depositions and minerals weathering) may partially compensate for these nutrient

extractions, this is usually not enough when the extraction is high and/or the rotations are short

(eg. Worrel and Hampson 1997).

Teak (Tectona grandis L.f.) is an important species in the worldwide quality tropical

hardwood sector (Pandey and Brown 2000; De Camino et al. 2002; De Camino and Morales

2013), with a total planted area of 4.3 ·106 ha, of which 132,780 ha are in Central America

(Kollert and Cherubini 2012). Problems of excessive nutrient export have also been reported for

teak plantations (Fernández-Moya et al. 2011, 2014) and could be related with the decline in

productivity noticed by some managers in the second and/or third rotations (Kollert and

Cherubini 2012). Timber extraction by final felling at the end of the 20 years rotation

constitutes a major nutrient output from the system, as harvesting (wood and bark) 150 teak

trees ha-1

at age 19 would export 220 kg N ha-1

, 281 kg Ca ha-1

, 88 kg K ha-1

, 63 kg Mg ha-1

, 23

kg P ha-1

, 39 kg S ha-1

, 6 kg Fe ha-1

, 129 g Mn ha-1

, 107 g Cu ha-1

, 212 g Zn ha-1

, 401 g B ha-1

.

106

Although this is a large output itself, the total nutrient extraction along one rotation is even

bigger as the timber extracted by thinnings, with a higher proportion of bark, also needs to be

taken into account. Final felling nutrient extraction in teak plantations could be of special

relevance in cases such as P and K, because of their low soil availability (Fernández-Moya et al.

2014). Due to the big relative relevance of teak plantations in Central American countries

(Kollert and Cherubini 2012), this species was chosen to analyze possible solutions to the

noticed problem of nutrient export and soil depletion.

In order to solve the lack of sustainability and the possible decrease in long term

production caused by nutrient depletion, many authors (eg. Rennie 1955; Fölster and Khanna

1997; Ma et al. 2007) have traditionally proposed to debark tree stems at the plantation site

aiming to minimize nutrient export. However, this is considered as an expensive, non-profitable

and time-consuming practice by many forest managers and nowadays it is not an option to

forests companies in Central America. On the other hand, as teak is a deciduous species,

nutrient allocation in the different tree tissues would be affected by the (re) translocation

processes related with leaves senesce (for a revision see Aerts 1996). Hence, timber may

hypothetically have higher nutrient concentration during the dry season (defoliated teak period),

when timber extraction is usually done, than in the growing season when nutrient would be

allocated in foliage (Fernández-Moya et al. 2014). Consequently, if that hypothesis would be

confirmed, modifying harvesting schedules would be an efficient and cheap tool to easily reduce

nutrient export by timber harvesting, minimizing the size of the abovementioned problem. The

present study analyzes foliar and trunk (wood and bark) nutrient concentration intra-annual

dynamics in teak planted forest in Guanacaste (Costa Rica), in order to check the formulated

hypothesis.

6.2. Material and methods

Study area

The present case study was carried out in a teak (Tectona grandis L.f.) planted forest,

property of the Cabalceta family, in San Juan de Santa Cruz, Guanacaste, Costa Rica. The

107

region is bioclimatically classified as tropical moist forest according to Holdridge’s life zones

(Holdridge 1967); with a climate characterized by an average annual precipitation of 2,000 –

3,000 mm and 4 to 6 dry months (December-April). The area has relatively fertile, reddish

clayey soils described as Typic “Paralithic” Haplustalfs and Ultic Haplustalfs (Soil Survey Staff

2010) by Thiele (2008).

The study was specifically settled in the “El Mango” stand (10.211º N, 85.573º W) with an

elevation of 115-120 m asl. This is a mature stand planted over previously grazed land in 1989,

i.e. 23 years old in 2012 at the beginning of the study. Low intensity thinning has led to an

excessive density of 200 – 250 trees ha-1

in this stand. A total of 12 topsoil samples were

collected (0-40 cm) and analyzed at CIA-UCR. Organic matter was determined by combustion

method as described by Horneck and Miller (1998); pH was determined in water 10:25; acidity,

Ca and Mg in KCl solution 1M 1:10; P, K, Zn, Fe, Mn and Cu in modified Olsen solution pH

8,5 (NaHCO3 0,5 N, EDTA 0.01M, Superfloc 127) 1:10. Soil texture was determined using the

modified Bouyoucos method, as described by Forsythe (1975). Topsoil (0-40 cm) fertility

results are summarized in Table 6.1, showing P, K and Zn deficiencies.

Sampling design and laboratory analyses

Foliar and trunk nutrient concentration were monitored along nine sampling times from

June 2012 to August 2013 (Table 6.2). Nine trees were randomly selected in order to obtain

three composite samples of leaves and trunk at each sampling time; hence, they were grouped in

three groups of three trees each. Foliar samples were taken between 6 and 7 m high using a

sampling stick. Trunk samples were taken at 1.3 m high using an increment borer, hence

representative proportion of wood and bark were sampled. Diameter at Breast Height (1.3 m

high, hereafter DBH) of sampled trees was measured and it is reported in Table 6.2. Figure 6.1

shows pictures of the evolution of the stand at different sampling times along the study period.

Three samples of senesced leaves were collected from the soil surface (i.e. litterfall) in January,

February and March (for a total of 9 samples) to estimate nutrient concentration after falling of

the trees and compare with the other abovementioned foliar analysis.

108

Table 6.1 Topsoil (0-40 cm) attributes in 12 sampled plots in a teak (Tectona grandis L.f.) plantation in

Santa Cruz, Guanacaste, Costa Rica

Meand and 95% confidence interval Critical values *

pH 5.8 (5.5, 6.1) 5.5

C org [%] 2.18 (1.93, 2.43)

Sand [%] 41 (33, 49)

Silt [%] 19 (14, 24)

Clay [%] 39 (36, 42)

Acidity [cmol (+) L-1] 0.18 (0.17, 0.19) 0.5

Ca [cmol (+) L-1] 18.3 (17.08, 19.52) 4

Mg [cmol (+) L-1] 6.56 (5.77, 7.35) 1

K [cmol (+) L-1] 0.37 (0.28, 0.46) 0.2

ECEC [cmol (+) L-1] 25.41 (23.5, 27.32) 5

A Sat [%] 0.7 (0.6, 0.8) 3

K Sat [%] 1.46 3.09

P [mg L-1] 2 (1, 3) 10

Zn [mg L-1] 1 (1, 1) 3

Cu [mg L-1] 10 (8, 12) 1

Fe [mg L-1] 49 (36, 62) 10

Mn [mg L-1] 17 (13, 21) 5

ECEC: Effective Cation Exchange Capacity [ECEC=Acidity+Ca+Mg+K]. AS: Acidity Saturation

[AS=Acidity/ECEC]. * Critical values as reference levels to evaluate soil fertility in Costa Rica, as

reported by Bertsch (1998) for general crops, by Alvarado and Fallas (2004) for AS and by Fernández-

Moya et al. (chapter 7 of the present Thesis) in teak plantations.

Table 6.2. Diameters at Breast Height (DBH, cm) of trees sampled to monitor within year foliar and

trunk nutrient concentration in teak (Tectona grandis L.f.) plantations in Guanacaste, Costa Rica

Sampling date DBH (Sample B)

(trees 1, 2 and 3)

DBH (Sample B)

(trees 4, 5 and 6)

DBH (Sample C)

(trees 7, 8 and 9)

Total average

(trees 1 to 9)

26-Jun-12 35.1 ± 5.9 29.4 ± 2.0 28.1 ± 5.8 30.9 ± 3.2

26-Aug-12 32.2 ± 6.5 31.8 ± 3.5 33.2 ± 1.1 32.4 ± 2.2

7-Oct-12 32.5 ± 4.1 32.7 ± 4.0 32.0 ± 3.6 32.4 ± 1.9

3-Dec-12 31.4 ± 3.0 33.1 ± 1.9 31.4 ± 2.0 32.0 ± 1.3

27-Jan-13 33.0 ± 4.3 36.2 ± 3.1 33.1 ± 5.8 34.1 ± 2.5

24-Feb-13 30.1 ± 5.5 31.9 ± 1.6 32.1 ± 4.7 31.4 ± 2.2

21-Mar-13 28.6 ± 5.1 28.4 ± 0.7 30.9 ± 5.1 29.3 ± 2.2

28-Jun-13 32.3 ± 1.5 33.2 ± 3.9 32.7 ± 2.5 32.7 ± 1.4

12-Aug-13 32.5 ± 4.9 32.5 ± 3.8 33.1 ± 2.6 32.7 ± 1.9

Tissues samples were analyzed at the Centro de Investigaciones Agronómicas of the

University of Costa Rica (CIA-UCR) to determine nutrient concentrations (N, P, K, Ca, Mg, S,

Fe, Cu, Zn, Mn and B). Samples were dried and then dry combustion was used to measure the N

concentration, and wet digestion and atomic spectrometry were used to extract and determine

the other nutrients (Bertsch 1998).

109

Figure 6.1. Pictures of the sampled teak (Tectona grandis L.f.) planted forest in Guanacaste, Costa Rica,

at June 2012 (a and b) and January 2013 (c and d)

6.3. Results and discussion

Foliar concentration dynamics

Foliar nutrients concentration varies during the year with different intensity depending on

the element, even though the pattern presents very high variability (Figures 6.2 and 6.3). This

variation could be mostly attributed to the widely analyzed retranslocation processes (eg. van

den Driessche 1984; Aerts 1996; Fife et al. 2008). In addition, litterfall nutrient concentration

(Figure 6.4) generally show lower values compared to nutrient concentration in green non

senescing leaves, matching with the abovementioned nutrient resorption or (re)translocation

theory. This process has been considered as a major nutrient conservation mechanism which

allows the plants to keep a reservoir readily available for further plant growth; while, on the

other hand, non resorbed nutrients circulate through the litterfall and some part of them would

eventually become available for future plant uptake (e.g. Vitousek 1984; Aerts 1996). These

110

processes become especially important in tropical rainy climates where the non resorbed

nutrients circulating through litterfall would translate in high nutrient lixiviation losses and

hence site impoverishment (e.g. Bruijnzeel 1991), even though the fast organic matter

decomposition dynamics reduces the time that the nutrients take to become available to plants

again (e.g. Vitousek 1984).

Figure 6.2. Macronutrient foliar concentrations variations with time (from June 2012 to August 2013) in

a teak (Tectona grandis L.f.) planted forest in Guanacaste, Costa Rica. Mean and 95% Confidence

Interval is shown for each sampling time.

111

Aerts (1996), in a profuse literature review, has already reported that around 54% of foliar

N and 50% of foliar P (the only two elements he analyzed) retranslocate before senescence as a

general pattern in deciduous species, even though he also found out a high variability in the

data. Specifically, Singh (2004) have already reported nutrient resorption associated to teak

foliage senescence. In the present case study, foliar N concentration remained between 1.5 and

1.8% between June and December 2012 and suddenly decreased to 0.8 and 1% in January and

February, when leaves are senescing (Figure 6.2). On the contrary, as Ca, Fe and Cu are non-

mobile nutrients they are accumulated on the leaves and non retranslocation process is

noticeable (Figures 6.2 and 6.3). This pattern has been observed by other authors, at least

regarding Ca lack of retranslocation (eg. Saur et al. 2000; Fife et al. 2008).

Figure 6.3. Micronutrient foliar

concentrations variations with time

(from June 2012 to August 2013) in

a teak (Tectona grandis L.f.) planted

forest in Guanacaste, Costa Rica.

Mean and 95% Confidence Interval

is shown for each sampling time.

112

Figure 6.4. Litter nutrient concentration variations with time (from January 2013 to March 2013) in a

teak (Tectona grandis L.f.) planted forest in Guanacaste, Costa Rica. Mean and 95% Confidence Interval

is shown for each sampling time.

Retranslocation can be also observed in the P, K and Zn foliar concentration results

(Figures 6.2 and 6.3). However, compared with N, the process followed a much less clear

pattern for these nutrients due to the big difference noticed between their foliar concentration in

the 2012 growing season compared with 2013 (Figures 6.2 and 6.3). While P resorption matches

in time with the observed N resorption between December and January, K resorption is

observed slightly later (between January and February) and Zn a little bit earlier (between

October and December) (Figures 6.2 and 6.3). K is a mobile nutrient that plays key roles in

photosynthesis and CO2 assimilation and exerts a regulatory effect on stomatal movement and

113

transpiration rates. Hence, normal K concentration values are probably sustained until last

senescing stages (January - February) to maintain transpiration. Foliar K losses could also be

expected in the rainy period as this element can be washed out. However, this tendency is not

observed in the collected data although it may be behind the higher variability of K foliar

concentration in October compared to August and December (Figure 6.2).

The establishment of foliar concentration references is considered as a useful tool to

evaluate the nutritional status of a stand and foliar analysis has been used for diagnostic

purposes and for designing nutritional and fertilization plans (e.g. Richards and Bevege 1972;

Mead 1984; Drechsel and Zech 1991). However, as foliar concentration varies within the year

(Figures 6.2 and 6.3), the foliar concentration reference should establish a clear temporal

framework to be applied and this should be carefully taken into account by the references users,

i.e. forest managers. Richards and Bevege (1972) have long ago identified two additional

problems related to the practical use of foliar concentration references, which match with the

present results (Figures 6.2 and 6.3): (1) big differences between trees, and (2) big differences

between years, which can be even bigger than between tree differences. Raupach (1967) and

Richards and Bevege (1972) recommend sampling at least six trees from each stand and analyze

them either singly or in composites samples of three trees in order to obtain a representative

estimation which take this tree-to-tree variation into account. In order to be careful with this

methodological issue, the abovementioned recommendations were taken into account in the

present study, analyzing three composites (of three trees each) samples each time. Further

research should be done, extending the present experimental design through several years in

order to take this other year-to-year variation into account.

Trunk concentration dynamics

An increase in N trunk concentration is observed corresponding with the decrease in N

concentration in senescing leaves previously commented (Figures 6.2 and 6.5). N trunk

concentration remains between 0.18 and 0.19% between August and December 2012 and

suddenly increases to 0.24, 0.21 and 0.29% in January, February and March, respectively

114

(Figure 6.5). A similar but somewhat less intense pattern can be observed in P trunk

concentration, which remains between 0.05 and 0.06% between August and December 2012 and

suddenly increases to 0.1 and 0.07% in January and February, respectively (Figure 6.5). A

slightly different pattern can be observed in K trunk concentration, which is estimated to be 0.07

and 0.09% in August and October 2012 and suddenly increases to be between 0.12 and 0.16

during the rest of the year (Figure 6.5). On the other hand, no noticeable pattern has been

observed for the trunk concentration of the other nutrients analyzed (Figures 6.5 and 6.6).

Figure 6.5. Macronutrient trunk (wood and bark) concentrations variations with time (from June 2012 to

August 2013) in a teak (Tectona grandis L.f.) planted forest in Guanacaste, Costa Rica. Mean and 95%

Confidence Interval is shown for each sampling time.

115

The increase of N and P trunk concentration between December and January corresponds

directly with the foliar concentration decrease shown at this same time (Figures 6.2 and 6.5).

This matches with the previously commented resorption process, as nutrients are retranslocated

to the wood, bark, branches and roots when the leaves start the senescing process (eg. van den

Driessche 1984; Aerts 1996; Fife et al. 2008). This pattern is different compared with the

followed by K foliar concentration, which show the resorption process slightly later (between

January and February, as previously stated) and the bigger increase in trunk concentration

slightly earlier, between October and December (Figures 6.2 and 6.5). Hence, the main variation

Figure 6.6. Micronutrient trunk

(wood and bark) concentrations

variations with time (from June 2012

to August 2013) in a teak (Tectona

grandis L.f.) planted forest in

Guanacaste, Costa Rica. Mean and

95% Confidence Interval is shown

for each sampling time.

116

of K trunk concentration may not be caused by the leaves senescing resorption processes but

because of other physiological mechanism. Low K trunk concentration period could be seen as a

large decrease between June and August instead of the abovementioned increase between

October and December (Figure 6.5). Actually, a decrease in K foliar concentration is also

registered between June and August both in 2012 and 2013 (Figure 6.2). Flowering (between

June to September) could be the main cause of these decreases tendencies both in foliar and

trunk concentrations, being the K mobilized from those tissues to the flowers. Flowers have

already been reported as a nutrient sink tissue with the highest nutrient concentration in teak

plantations (Kumar et al. 2009).

Trunk nutrient concentration is very low compared to foliar concentration (Figures 6.2, 6.3,

6.5 and 6.6). On the other hand, the high amount of biomass accumulated in the tree stem makes

it an important sink of nutrients and, as a consequence, loss of nutrient through wood removal at

harvesting is considered as a major cause of nutrient impoverishment of forest sites (eg. Rennie

1955; Fölster and Khanna 1997; Worrel and Hampson 1997). The results show that the trunk

nutrient concentration varies along the year (Figures 6.2, 6.3, 6.4 and 6.5), verifying the

hypothesis stated in the Introduction section (Fernández-Moya et al. 2014). Thus, harvesting

time is influencing the amount of nutrients extracted through wood removal. However, teak is

usually harvested in the dry period to minimize the impact over the site and because of

operational and logistic reasons (De Camino and Morales 2013), which can affect managers

decision making, as it further discussed. Once the previous hypothesis has been confirmed,

modifying harvesting time would have different consequences depending on the alternative

timing considered:

(1) Timber harvested in the dry period (January to March) would have an average N-P-K

concentration of 0.25, 0.07 and 0.14%, respectively, while it would be 0.19, 0.05 and

0.08%, respectively, if it would be harvested between August and October, with a

consequent reduction of 24, 29 and 43% of the N-P-K exported by timber harvesting,

respectively.

117

(2) Timber hypothetically harvested in December would have an average N-P-K

concentration of 0.18, 0.05 and 0.12%, respectively, with a consequent reduction of 28,

29 and 14% of the N-P-K exported by timber harvested in the dry period (January to

March), respectively.

Based on the present results, harvesting between August and October would cause a major

reduction in N-P-K export, due to the lower concentration of those nutrients on the wood and

bark at that period of the year. However, this period may present some logistic problems for

harvesting operations as it matches with the rainy period of normal years. A possible solution

may be to perform some of the harvesting operations at this time (e.g. cutting), left the trees out

in the field and extract them during the dry period; although this option should be examined

carefully as it may have some outbreak risk or produce a significant increase the operational

costs. As another alternative, starting the harvest in December (slightly earlier), at the begging

of the dry period but before leaves senescence, would be a good option too as it has not any

operational inconveniency and wood and bark concentration are also lower than later in the dry

period.

Nutrient export by timber harvesting has been reported to be of especial relevance for

precisely the cases of N, P and K in the various studied teak plantations in Central America

(Fernández-Moya et al. 2014). N and Ca are the nutrients with highest values of nutrient export

(220 kg N ha-1

, 281 kg Ca ha-1

) but Ca does not show any sustainability problem due to its high

soil reservoir (Fernández-Moya et al. 2014). N sustainability could not be estimated in previous

studies due to the lack of soil N information but it is considered to be the limiting nutrient in

many forest ecosystems (e.g. Hedin et al. 2009). The estimated amount of P and K exported by

timber harvesting (88 kg K ha-1

and 23 kg P ha-1

) is much lower than N but they represent an

important sustainability problem for the system as the soils where the teak plantations are

generally established in Central America have very low P and K contents (Fernández-Moya et

al. 2014). Hence, a reduction of N-P-K export by timber extraction would consequently be of

great importance for the long time site productivity.

118

The present paper shows some promising results which could cause an improvement on

nutrient management and planted forests sustainability. However, a more detailed study along

several years, with higher number of samples and more locations would be recommended to

confirm this marked tendencies. For further research about this topic is also worthwhile to point

out the need to analyze the differences in nutrients trunk concentration with respect to the height

of sampling (e.g. Helmisaari and Siltala 1989; Saint-Andre et al. 2002). Trunk sampling for the

present paper was performed at a height of 1.3 m (Diameter at Breast Height); sampling trunk at

higher points may show an even more intense influence of the leaves senescence retranslocation

processes as the sampling point would be nearer to the crown. In addition, a wider approach

should also consider the nutrient export by seeds which could be a major output from a system

when those are collected to be sold (as it is the case for the present study case), although little

seasonal variation can be expected on this process. Similarly the root dynamics should also be

considered as small roots senesce and regrowth are probably following a similar pattern than the

observed in the foliage. Finally, teak phenology variability depending on the climate of a

specific site should be also taken into account. For example, teak shows a marked deciduous

behavior in Northern Costa Rica as it has been discussed in the present paper while it performs

as an evergreen in the more humid Sourthern Costa Rica where no summer is noticed.

Several authors have studied the temporal and spatial (radial and height variations) nutrient

dynamics within stem wood, pointing out a generalized tendency of nutrient translocation from

senescing sapwood and hardwood to active sapwood (Helmisaari and Siltala 1989; Andrews et

al. 1999; Laclau et al. 2001;Meerts 2002; Saint-Andre et al. 2002). However, no other study to

our knowledge has evaluated the possibility of using this natural phenomenon in order to

minimize nutrient removal by wood harvesting.

CAPÍTULO 7

RELATIONSHIPS BETWEEN NUTRIENT SOIL

AVAILABILITY, FOLIAR CONCENTRATION AND

TREE GROWTH IN TEAK (Tectona grandis L.f.)

PLANTATIONS

121

7.1. Introduction

Teak (Tectona grandis L.f.) is an important species in the quality tropical hardwood sector

worldwide (Pandey and Brown 2000; De Camino et al. 2002; Kumar 2011), with a total planted

area of 4.3 ·106 ha, of which 132,780 ha are in Central America (3%) and 55,000 in Panama,

making it the ninth country in terms of area of planted teak forest (Kollert and Cherubini 2012).

The research activity directed towards increasing the productivity of teak plantations has

focused on silvicultural and genetic approaches (e.g. Bermejo et al. 2004; Pérez and Kanninen

2005a,b; Adu-Bredu 2008; Jayaraman and Bhat 2011). Given the socio-economic importance of

the species (Kumar 2011), relatively little attention has been paid to nutrition and soil

management in teak plantations, although a number of authors have made reference to foliar

nutrient concentrations in teak (Drechsel and Zech 1991, 1994; Boardman et al. 1997;

Fernández-Moya et al. 2013) and several studies concerned with plant-soil relationships have

been undertaken (for a revision see Kumar 2011; Alvarado 2012 b). Nevertheless, further

research into teak nutrition is still required in order to improve plantation management, increase

productivity and enhance sustainability.

The use of environmental information to predict potential site productivity is a common

practice in forestry, and many different techniques have been used for this purpose (West 2006).

Several authors have proposed models to estimate site index for teak, taking into account

climate and soil parameters, although various deficiencies can be identified in their

methodology (Vásquez and Ugalde 1994; Montero 1999; Mollinedo et al. 2005). Foliar

concentration is also considered a useful parameter to evaluate the nutritional status of a stand

because (a) its variation is highly dependent on site and soil parameters; (b) it reflects the

current nutrient supply; (c) it allows diagnosis of nutritional deficiencies when they are not

severe enough to cause visually observable symptoms and, thus, it allows action to be taken

before the effects on productivity are significant; and (d) deficiency symptoms are easily

confused with other effects when visual guidelines are used (Mead 1984; Drechsel and Zech

1991; West 2006). Foliar analysis has been used for diagnostic purposes, establishing critical

levels or ranges within the simplified concept that a plant which has a nutrient concentration

122

below the critical level will be less productive than plants with a higher concentration (see

Richards and Bevege 1972). Relationships between foliar nutrient concentration and forest

growth have been studied in many species such as pines, eucalypts and others (e.g. Richards and

Bevege 1972; Lamb 1977; Álvarez-Álvarez et al. 2011). Critical levels are commonly used as a

conceptual framework in plant nutrition (originally for foliar analysis but also extrapolated to

soil properties), distinguishing four ranges of values (deficiency, critical, luxury and toxicity

ranges) depending on the relationship between the studied variable and the growth or yield of

the studied plant.

The present study stems from the need of forest managers for a tool to help them accurately

interpret soil and foliar laboratory analyses, with the aim of optimising the nutrient use by trees

and achieving the maximum sustainable productivity. Hence, foliar nutrient concentration and

soil availability were analysed in a number of stands in which growth was categorized as good,

medium or poor, in order to study the relationship between nutrition and tree growth in teak

plantations in Panama.

7.2. Material and methods

Study area

The study area was located at the Ecoforest teak (Tectona grandis L.f.) plantations in the

Panama Canal Watershed. The area is classified as tropical wet forest according to Holdridge’s

life zones (Holdridge 1967), with a humid climate characterized by mean annual rainfall of

2,500–3,100 mm with 4 dry months. The soils of the study area are generally clayey and acidic

red soils of low fertility (Ultisols) associated with low fertility Inceptisols and Entisols. The

studied plantations are representative of company-managed teak plantations in the region. The

management of these plantations consists of continuous silvicultural activities: weed control,

pruning, thinning regime (from approximately 1,111 trees ha-1

at establishment to 150-200 trees

ha-1

at final felling) and fertilization during the establishment. The use of clones has become

common practice in recent years, but they were not employed in the studied plantations. A

123

commercial volume of 100-150 m3 is expected for the final harvesting in this kind of plantation,

with a rotation of approximately 20-25 years.

Field sampling and design

The sampling design comprised a total of 89 permanent plots selected by the company

managers, taking into account their apparent rate of growth: 29 considered good, 30 medium

and 30 poor. These plots belonged to a network of permanent plots established by the company

for monitoring plantation growth; hence, the following growth information was available:

diameter at breast height (“DBH”), height (“H”) and age. As plot age varies from 3 to 8 years,

the mean annual increment in DBH and H (“DBH MAI” and “H MAI”, respectively) were

calculated in order to compare plots of different ages. However, as stand age has a strong

influence on MAI, the comparison between plots of different ages is only relative. Site index

(“SI”) was calculated as a measure of relative growth using a modified version of the original

models proposed by Bermejo et al. (2004). The modification consists of extending the original

classification (three classes SI19, SI21 and SI23, related to tree height at a base age of 10 years)

to five classes: SI<18 (considered very poor growth), SI19 (considered poor growth), SI21

(considered medium growth), SI23 (considered high growth) and SI>24 (considered very high

growth).

Foliage sampling was conducted in the upper third of the canopy, taking samples from the

four orientations on the end portion of the branches of 10 dominant trees in each sampled

permanent plot. Sampled leaves were close to maximum size and displayed no symptoms of

ageing, fungi or illness. One composite foliage sample was taken at each of the selected plots

and analyzed at “Agrotec Laboratorios Analíticos” (Costa Rica) for nutrient concentration (N, P,

K, Ca, Mg, S, Na, Fe, Mn, B, Cu, Zn, Mo, Al, Si) using dry combustion and atomic

spectrometry (Bertsch 1998). Five soil micro-pits (0-40 cm) were dug at each sampling site (one

at the center and one at each corner of the plot) and two soil samples were collected from each:

topsoil (0-20 cm) and subsoil (20-40 cm). The five topsoil and subsoil samples were then

composited to obtain one topsoil and one subsoil sample per plot. These soil samples were

analyzed at “Agrotec Laboratorios Analíticos” (Costa Rica) for NO3, NH4, Ntotal, available

124

nutrients (P, K, Ca, Mg, SO4, Na, Fe, Mn, B, Cu, Zn, Mo, Al), acidity, Effective Cation

Exchange Capacity (ECEC), pH, K Saturation (K Sat), Ca Saturation (Ca Sat), Mg Saturation

(Mg Sat), Na Saturation (Na Sat), acidity Saturation (A Sat), Organic Matter (OM), Ca Mg-1

, Ca

K-1

, Mg K-1

and (Ca+Mg) K-1

using pH in CaCl2, the Kjeldahl method for N and Mehlich 3

methodology for the other available nutrients. Soil properties estimated at two depths were

averaged to estimate soil properties for the total depth (0 – 40 cm). Foliar and soil sampling

were carried out in August 2007, during the first stage of the rainy period, at a moment of

maximum photosynthetic activity.

Statistical analysis

Generalized Linear Models were used to analyze, one-by-one, the relationships between

DBH MAI and H MAI (response variables) and each of the foliar nutrient concentrations (15

explanatory variables) as well as each of the soil properties (29 explanatory variables); hence,

44 models per response variable (a total of 88 models). In order to explore the relationship

between each combination of response and exploratory variables, 78 models were fitted for each

combination, taking into account several family distributions (normal and gamma), link

functions (inverse and potential from λ=0 to λ=2 with a λgap=0.1) and model forms (with and

without intercept). The best model for each combination of response and exploratory variables

was chosen using the Akaike Criteria which takes into consideration the deviance of the model

and the degrees of freedom. The goodness-of-fit of the models was estimated by measuring the

difference in percentage between the deviance of the model and the deviance of a model with no

covariates (hereafter efficiency, EF).

Site index was considered a multinomial variable with five levels. Hence, a multinomial

regression approach was used (Fox and Weisberg 2011) to analyze, one-by-one, the

relationships between SI and each of the 44 explanatory variables (foliar nutrient concentrations

and soil properties). In addition, two tree regression models were adjusted to analyze the

relationships between 1) SI and all 15 foliar nutrient concentrations, and 2) SI and all 29 soil

properties.

125

All the statistical analyses were done using R (R Development Core Team 2011). All

statistical tests are considered significant throughout the text at α=0.1, except where the contrary

is stated.

7.3. Results

Relationships between foliar nutrient concentration and teak growth

Table 7.1 summarizes details the foliage nutrient concentration of the study sites. DBH MAI

and H MAI showed a positive relationship with foliar concentration of N, K and P, and a

negative relationship with Mg, Mn and B (Table 7.2). However, the relationships with N, K, Mg

and B were very weak (EF < 20%) and are considered less important than those with P and Mn

(Table 7.2, Figure 7.1). Relationships were also identified between site index and foliar P (p-

value = 0.005), K (p-value = 0.017), Mg (p-value = 0.013), S (p-value < 0.001), Mn (p-value =

0.005), Cu (p-value = 0.034) and Si (p-value = 0.047). The fitted regression-tree model

identified P as the nutrient with the highest influence on site index (Figure 7.2). A threshold of

0.125% foliar P concentration was identified to separate most sites classified SI<18 from those

classified SI>24, with a cross-validation value of CV=0.665 (Figure 7.2).

Relationships between soil nutrient availability and teak growth

Table 7.3 summarizes the soil information of the study sites. DBH MAI displayed a positive

relationship with soil OM, pH, N, NO3, Ca, Mg, K, ECEC, K Sat, Ca Sat Fe, Mn, and B, and a

negative relationship with soil acidity, Al, Na Sat, A Sat, Mg K-1

, SO4 and Mo (Table 7.4).

However, many of these relationships were too weak; the only ones considered strong enough to

be taken into consideration were those with EF>20%: pH, Ca, K, acidity, Na Sat A Sat, Fe and

B (Table 7.4, Figure 7.3). Similarly, H MAI showed a positive relationship with soil OM, pH,

N, NO3, Ca, Mg, K, ECEC, K Sat, Ca Sat, Fe and B, and a negative relationship with soil

acidity, NH4, Al, Na Sat, A Sat, Ca K-1

, Mg K-1

, (Ca+Mg) K-1

, SO4 and Mo (Table 7.4). Of

these, the only ones considered significant (with an EF>20) were the relationships with pH, K,

acidity, Na Sat, A Sat and Fe (Table 7.4, Figure 7.3).

126

Table 7.1. Foliar nutrient concentration at the sampled sites in teak (Tectona grandis L.f.) plantations in

Panama (n=89)

Nutrient Mean Confidence interval Coefficient of variation (%) Min Max

N (%) 1.9 (1.85, 1.95) 13 1.36 2.61

P (%) 0.12 (0.12, 0.12) 8 0.09 0.16

K (%) 0.66 (0.63, 0.69) 26 0.26 1.02

Ca (%) 1.06 (1.01, 1.11) 25 0.50 1.84

Mg (%) 0.26 (0.24, 0.28) 38 0.12 0.57

Na (%) 0.2 (0.2, 0.2) 10 0.14 0.25

S (%) 0.01 (0.01, 0.01) 0 0.00 0.01

Al (mg kg-1) 42.75 (41.49, 44.01) 14 33.04 72.23

Fe (mg kg-1) 50.44 (45.4, 55.48) 48 20.79 135.30

Mn (mg kg-1) 28.84 (27.57, 30.11) 21 15.55 53.88

B (mg kg-1) 8.01 (7.63, 8.39) 23 4.18 12.78

Cu (mg kg-1) 28.3 (26.5, 30.1) 31 15.69 66.96

Zn (mg kg-1) 0.35 (0.33, 0.37) 23 0.20 0.64

Mo (mg kg-1) 36.16 (34.09, 38.23) 28 19.72 93.82

Si (mg kg-1) 601.66 (567.22, 636.1) 28 330.70 1,017.00

Table 7.2. Regressions between Diameter at Breast Height and Height Mean Annual Increment (DBH

MAI and H MAI, respectively) and foliar nutrient concentration at the sampled sites in teak (Tectona

grandis L.f.) plantations in Panama (n=89). Below specified models are in the form [y = (b0 + b1 · x)1/λ

and, when λ=0, y = exp(b0 + b1 · x)], where the response variables (y) are DBH MAI and H MAI and the

foliar nutrient concentration are the explanatory variables (x). When no model including age as a

parameter was statistically significant, the model only included an intercept (b0). Models with efficiency

(EF) higher than 20% are stressed using bold font letter.

Nutrient

DBH MAI H MAI

b0 b0

[Std. error] b1

b1

[Std. error] λ EF (%) b0

b0

[Std. error] b1

b1

[Std. error] λ EF (%)

N (%) 3.214 0.159 2 0.04 3.916 0.206 2 0.06

P (%) 7.549 0.186 0 0.20 8.374 0.199 0 0.20

K (%) 3.893 1.201 3.314 1.841 1 0.00 3.282 1.462 6.294 2.299 2 0.07

Ca (%) 2.464 0.063 1 0.00 2.721 0.074 1 0.00

Mg (%) 7.827 0.822 -6.548 2.641 2 0.06 10.166 1.034 -10.264 3.207 2 0.09

S (%) 2.464 0.063 1 0.00 2.721 0.074 1 0.00

Na (%) 2.464 0.063 1 0.00 2.721 0.074 1 0.00

Fe (mg kg-1) 2.464 0.063 1 0.00 2.721 0.074 1 0.00

Mn (mg kg-1) 0.313 0.022 0.002 0.000 -1 0.21 0.275 0.021 0.002 0.000 -1 0.21

B (mg kg-1) 0.300 0.050 0.004 0.002 -1 0.05 0.266 0.048 0.004 0.002 -1 0.05

Cu (mg kg-1) 2.464 0.063 1 0.00 2.721 0.074 1 0.00

Zn (mg kg-1) 2.464 0.063 1 0.00 2.721 0.074 1 0.00

Mo (mg kg-1) 2.464 0.063 1 0.00 2.721 0.074 1 0.00

Al (mg kg-1) 2.464 0.063 1 0.00 2.721 0.074 1 0.00

Si (mg kg-1) 2.464 0.063 1 0.00 2.721 0.074 1 0.00

EF (%): Model efficiency, pseudo R2 estimate for Generalized Linear Mixed Models.

127

Table 7.3. Soil properties (0-40 cm) at the sampled sites in teak (Tectona grandis L.f.) plantations in

Panama (n=89)

Soil property Mean Confidence interval Coefficient of variation (%) Min Max

OM (%) 2.32 (2.13, 2.51) 39 0.59 4.17

pH 4.44 (4.34, 4.54) 11 3.60 5.77

N (mg kg-1) 0.36 (0.31, 0.41) 68 0.07 2.09

NO3 (mg kg-1) 0.23 (0.19, 0.27) 90 0.00 1.55

NH4 (mg kg-1) 0.13 (0.11, 0.15) 86 0.00 0.58

Ca (cmol(+) L-1) 10.52 (9.15, 11.89) 63 1.26 30.80

Mg (cmol(+) L-1) 7.71 (6.75, 8.67) 60 1.01 18.74

K (cmol(+) L-1) 0.66 (0.57, 0.75) 65 0.12 2.24

Na (cmol(+) L-1) 0.19 (0.18, 0.2) 32 0.10 0.46

Al (cmol(+) L-1) 16.47 (15.48, 17.46) 29 10.53 33.02

Acidity (cmol(+) L-1) 1.02 (0.89, 1.15) 59 0.12 2.59

ECEC (cmol(+) L-1) 20.09 (17.92, 22.26) 52 4.06 47.60

K Sat (%) 3.33 (3.08, 3.58) 36 1.33 7.08

Ca Sat (%) 50.78 (48.36, 53.2) 23 24.04 75.67

Mg Sat (%) 37.43 (35.35, 39.51) 27 18.64 65.85

Na Sat (%) 1.15 (1.03, 1.27) 49 0.38 3.83

A Sat (%) 7.31 (5.85, 8.77) 96 0.46 40.89

Ca Mg-1 1.53 (1.38, 1.68) 48 0.37 4.03

Ca K-1 17.51 (16.04, 18.98) 40 6.85 38.65

Mg K-1 13.30 (11.76, 14.84) 56 4.05 48.10

(Ca+Mg) K-1 30.81 (28.29, 33.33) 39 11.79 70.49

P (mg kg-1) 14.23 (13.51, 14.95) 24 8.72 26.68

SO4 (mg kg-1) 34.26 (29.73, 38.79) 64 15.70 162.79

Fe (mg kg-1) 138.73 (129.54, 147.92) 32 64.99 259.20

Mn (mg kg-1) 120.26 (97.47, 143.05) 91 12.47 583.25

B (mg kg-1) 1.40 (1.32, 1.48) 28 0.70 2.25

Cu (mg kg-1) 3.73 (3.38, 4.08) 45 0.29 9.08

Zn (mg kg-1) 3.97 (2.68, 5.26) 156 0.74 49.95

Mo (mg kg-1) 0.65 (0.62, 0.68) 23 0.43 1.13

Figure 7.1. Diameter at Breast Height and Height Mean Annual Increment (DBH MAI and H MAI,

respectively) relationships with foliar P and Mn concentration at the sampled sites in teak (Tectona

grandis L.f.) plantations in Panama (n=89). Lines represent fitted models (Table 7.2).

128

Figure 7.2. Regression

tree between Site Index

and foliar nutrient

concentration at the

sampled sites in teak

(Tectona grandis L.f.)

plantations in Panama

(n=89).

Figure 7.3. Diameter at Breast Height and Height Mean Annual Increment (DBH MAI and H MAI,

respectively) relationships with various soil properties (0-40 cm) at the sampled sites in teak (Tectona

grandis L.f.) plantations in Panama (n=89). Lines represent fitted models (Table 7.4).

129

Table 7.4. Regressions between Diameter at Breast Height and Height Mean Annual Increment (DBH

MAI and H MAI, respectively) and various soil properties (0-40 cm) at the sampled sites in teak (Tectona

grandis L.f.) plantations in Panama (n=89). Below specified models are in the form [y = (b0 + b1 · x)1/λ

and, when λ=0, y = exp(b0 + b1 · x) when λ=0], where the response variables (y) are DBH MAI and H

MAI and the foliar nutrient concentration are the explanatory variables (x). When no model including age

as a parameter was statistically significant, the model only included an intercept (b0) representing the

mean of the response variable. Models with efficiency (EF) higher than 20% are stressed using bold

font letter.

Nutrient

DBH MAI H MAI

b0 b0

[Std. error] b1

b1

[Std. error] λ EF (%) b0

b0

[Std. error] b1

b1

[Std. error] λ EF (%)

OM (%) 4.115 0.724 0.857 0.319 2 0.07 4.756 0.922 1.163 0.411 2 0.08

pH 0.507 0.010 0.9 0.24 0.554 0.012 0.9 0.21

N (mg kg-1) 0.439 0.015 -0.089 0.029 -1 0.08 0.397 0.015 -0.078 0.029 -1 0.06

NO3 (mg kg-1) 0.796 0.034 0.434 0.110 0 0.17 2.014 0.068 0.909 0.254 0.8 0.16

NH4 (mg kg-1) 2.464 0.063 1 0.00 0.347 0.015 0.166 0.095 -1 0.04

Ca (cmol(+) L-1) 3.885 0.491 0.216 0.051 2 0.20 4.783 0.657 0.259 0.067 2 0.16

Mg (cmol(+) L-1) 0.463 0.019 -0.007 0.002 -1 0.13 0.420 0.019 -0.006 0.002 -1 0.11

K (cmol(+) L-1) 0.484 0.015 -0.110 0.016 -1 0.31 0.778 0.040 0.325 0.051 0 0.33

Na (cmol(+) L-1) 2.464 0.063 1 0.00 2.721 0.074 1 0.00

Al (cmol(+) L-1) 0.281 0.036 0.008 0.002 -1 0.13 11.109 1.196 -0.222 0.062 2 0.10

Acidity (cmol(+) L-1) 0.326 0.016 0.083 0.016 -1 0.26 0.293 0.016 0.077 0.015 -1 0.24

ECEC (cmol(+) L-1) 1.520 0.051 0.010 0.002 0.6 0.18 4.572 0.741 0.145 0.039 2 0.15

K Sat (%) 0.484 0.029 -0.023 0.008 -1 0.09 0.466 0.026 -0.029 0.007 -1 0.15

Ca Sat (%) 2.610 1.223 0.069 0.025 2 0.08 3.568 1.638 0.076 0.033 2 0.06

Mg Sat (%) 2.464 0.063 1 0.00 2.721 0.074 1 0.00

Na Sat (%) 0.303 0.021 0.094 0.019 -1 0.24 0.271 0.021 0.089 0.018 -1 0.23

A Sat (%) 0.354 0.013 0.008 0.002 -1 0.25 0.320 0.013 0.007 0.002 -1 0.22

Ca Mg-1 2.464 0.063 1 0.00 2.721 0.074 1 0.00

Ca K-1 2.464 0.063 1 0.00 0.324 0.025 0.003 0.001 -1 0.04

Mg K-1 0.372 0.021 0.003 0.001 -1 0.04 0.325 0.021 0.003 0.001 -1 0.06

(Ca+Mg) K-1 2.464 0.063 1 0.00 0.306 0.026 0.002 0.001 -1 0.07

P (mg kg-1) 2.464 0.063 1 0.00 2.721 0.074 1 0.00

SO4 (mg kg-1) 0.341 0.020 0.002 0.001 -1 0.15 0.302 0.019 0.002 0.001 -1 0.00

Fe (mg kg-1) 0.045 0.002 2 0.39 0.049 0.002 1.9 0.42

Mn (mg kg-1) 5.235 0.468 0.007 0.004 2 0.06 2.721 0.074 1 0.00

B (mg kg-1) 0.561 0.034 -0.108 0.022 -1 0.21 0.507 0.033 -0.097 0.022 -1 0.18

Cu (mg kg-1) 2.464 0.063 1 0.00 2.721 0.074 1 0.00

Zn (mg kg-1) 2.464 0.063 1 0.00 2.721 0.074 1 0.00

Mo (mg kg-1) 8.957 1.230 -4.408 1.731 2 0.06 10.665 1.633 -4.980 2.313 2 0.04

EF (%): Model efficiency, pseudo R2 estimate for Generalized Linear Mixed Models.

Site index showed a similar relationship with soil properties, displaying significant

relationships with N (p-value = 0.066), NO3 (p-value = 0.011), K (p-value < 0.001), Ca (p-value

= 0.002), Mg (p-value = 0.004), SO4 (p-value < 0.001), Fe (p-value < 0.001), B (p-value <

0.001), Mo (p-value = 0.004), Al (p-value < 0.001), acidity (p-value = 0.001), ECEC (p-value =

130

0.003), pH (p-value = 0.002), K Sat (p-value = 0.008), Ca Sat (p-value = 0.001), Mg Sat (p-

value = 0.002), Na Sat (p-value < 0.001), A Sat (p-value < 0.001), Ca Mg-1

(p-value = 0.02), Mg

K-1

(p-value < 0.001) and (Ca+Mg) K-1

(p-value = 0.003). However, the regression tree

technique identified Na Sat as having the greatest influence on Site Index, followed by K Sat

and soil acidity with a cross-validation value of CV=0.584 (Figure 7.4). Sites with soil Na Sat

above 1.1 are probably classified as SI<18, while sites with soil Na Sat lower than 1.1 and K Sat

higher than 3.09 would be classified as SI>24 (Figure 7.4). The differences between sites

classified as SI19 and SI21 were less clear, both being characterized by Na Sat < 1.1 and K Sat

< 3.09, although soil acidity at SI21 sites is > 0.75 while in the case of SI19 sites it is < 0.75

(Figure 7.4).

Figure 7.4. Regression tree between Site Index and soil attributes at the sampled sites in teak (Tectona

grandis L.f.) plantations in Panama (n=89).

131

7.4. Discussion

Relationships between foliar nutrient concentration and teak growth

Teak growth showed no relationship with the foliar concentration of Ca, S, Na, Fe, Cu, Zn,

Mo, Al and Si (Table 7.2); hence it is assumed that the concentration of these elements is within

the sufficiency to luxury range. A weak positive relationship was identified between teak

growth and the foliar concentration of N and K, so it is assumed that these are within the critical

range (Table 7.2), whereas P would appear to be in the deficiency range since the relationship in

this case is stronger (Table 7.2, Figure 7.1). On the other hand, Mg, Mn and B would seem to be

in the toxicity range, as they show a slightly negative relationship with tree growth, although

this is not in accordance with the adequate values cited in the literature (Drechsel and Zech

1991; Zech and Drechsel 1991; Boardman et al. 1997).

The critical range threshold for P concentration can be determined as 0.125% (Figure 7.2), as

plantations with a foliar P content higher than this are likely to be classified as IS>24 (very good

sites). This threshold is very similar to that reported by Zech and Drechsel (1991) for teak

plantations in West Africa (>0.13%, 0.18±0.04), as well as to the adequate range identified in

plantations of Costa Rica and Panama (>0.12%, 0.16±0.04) by Fernández-Moya et al. (2013),

and to those summarized in a review of literature concerning teak by Drechsel and Zech (1991):

deficiency range between 0.1-0.13% and intermediate or adequate range between 0.12 and

0.21%. P limitation is common in many soils worldwide, as P fixation occurs when Ca

phosphates are formed at high soil pH or when Fe and Al phosphates are formed in highly

weathered acidic soils, common in tropical areas. Hence, P is recognized as an important

limiting nutrient in forest plantations (for a revision see Fox et al. 2011). Fernández-Moya et al.

(2014) reported high accumulation of P in the aboveground biomass of planted teak forests in

contrast to the low soil P availability and suggest the following possible explanations: a) the

methodology used to estimate soil P availability; b) the sampled trees may have benefitted from

specific conditions in certain parts of the site, where soil nutrient availability was higher than

the average sampled, thus leading to better growth, and/or a deeper root system which allows

them to explore a larger volume of soil; c) teak roots may produce phosphatases which improve

132

the mineralization rates of organic-P, resulting in higher levels of available P than those

detected in the soil analysis, especially where mycorrhizal activity is present (Corryanti et al.

2007); d) P input as atmospheric deposition could be playing a key role in plantation nutrition

with this element; e) P could be limiting teak productivity. The present study would appear to

support the latter hypothesis, as higher foliar P content is related with higher site index, MAI H

and MAI DBH (Figures 7.1 and 7.2, Table 7.2). Alvarado et al. (2004) collected mycorrhizae

throughout teak plantations in Costa Rica and proposed the inoculation of seedlings as a way to

improve P uptake and enhance productivity, particularly in acid soils.

Together with P, other variables, such as N, Ca and soil acidity, have been identified as

limiting the productivity of teak plantations (e.g. Zech and Drechsel 1991; Drechsel and Zech

1994). Although N availability is usually considered to be the main limiting factor for growth in

many terrestrial ecosystems, it seems to be less critical in tropical ecosystems in comparison

with other nutrients such as P (Vitousek 1984; Hedin et al. 2009). Alvarado (2012 b) also

minimizes the role of N in limiting the productivity of teak plantations in Central America, as he

hypothesized that around 70% of N requirements could be provided through residue

mineralization along with wet and dry N deposition.

Relationships between soil nutrient availability and teak growth

As regards the relationship between teak growth and the analyzed soil properties, no

relationship was found with P, Na, Cu, Zn, Mg Sat and Ca Mg-1

(Table 7.4), so we can assume

that these properties fall within the sufficiency and luxury ranges. The soil properties OM, N,

NO3, Mg, ECEC, K Sat, Ca Sat, Mn, NH4, Al, Ca K-1

, Mg K-1

, (Ca+Mg) K-1

, SO4 and Mo

showed slight positive or negative relationships with teak growth, so it can be assumed that

these are either in the critical range or in the upper zone of the toxicity range, close to the luxury

range (Table 7.4). Meanwhile pH, Ca, K, Fe and B fall within the deficiency range, since the

relationships in this case are positive and stronger (Table 7.4, Figure 7.3). Acidity, Na Sat and A

Sat fall within the toxicity range. Soil Na Sat, K Sat and acidity can also be used to predict SI

classes (Figure 7.4). A value for Na Sat = 1.1% is identified as a threshold: soils with higher Na

133

Sat are likely to be classified as SI<18; on the other hand, soils with Na Sat < 1.1 and K Sat >

3.09% would be classified as SI>24 (Figure 7.4). Soils with low Na Sat (<1.1%) and low K Sat

(<3.09%) remained between the extremes and received a poorer classification; acidity being the

most important variable in this case for classifying them as SI 19 and SI 21 (Figure 7.4).

Soil Na Sat seems to be the most important limiting factor affecting teak growth among all

those analyzed (Figure 7.4). Na toxicity is usually found in saline areas but to our knowledge,

no other cases of Na toxicity have been reported in teak plantations. Soil K Sat, however, does

have a strong influence on tree growth (Figure 7.4), reflecting the expected tendency.

Fernández-Moya et al. (2014) also highlight the high K accumulation in the aboveground

biomass of planted teak forests in contrast to the generally low soil K availability in the region.

In addition, as soils where teak is usually planted generally contain high levels of Ca and Mg

(Table 7.3), the equilibrium between the base cations is vital to teak growth, as evidenced by the

importance of K Saturation rather than absolute value.

Low tolerance to soil acidity in teak plantations has been reported by many authors (e.g.

Zech and Drechsel 1991; Drechsel and Zech 1994; Alvarado and Fallas 2004; Wehr et al. 2010;

Alvarado 2012 b). This behavior is well documented and a threshold of A Sat=3% has been

proposed (Alvarado and Fallas 2004) as a critical level for acidity saturation in teak plantations,

teak growth being severely reduced in soils with higher levels of acidity. The negative

relationship of acidity and A Sat with DBH MAI and H MAI identified in this study (Table 7.4,

Figure 7.3) is in accordance with the assumed trend but this is not reflected in a similar negative

relationship between foliar Al and teak growth (Table 7.2). This absence of relationship

between foliar Al and growth was also reported by Drechsel and Zech (1994), who found foliar

Al to be a bad indicator of Al toxicity in teak plantations. However, in contrast to the previous

statements, the relationship found between soil acidity and site index was positive (Figure 7.4).

This positive relationship is difficult to explain and may be an indirect effect of the variation in

another property. In fact, in the groups characterized as SI21 with Acidity>0.75 cmol(+) L-1

, not

one stand is classified as SI>24. With the exception of the SI>24 and SI<18, characterized by

the Na Sat and K Sat on the first and second branches of the decision tree, there were only 37

134

sites left of intermediate SI classes with a range of acidity between 0.41 and 2.44 cmol(+) L-1

,

revealing a strong correlation with pH, which is probably influencing the solubility of another

nutrient and the N mineralization dynamics. Soil acidity in these plots is relatively high,

corresponding to an A Sat range of 0.95 – 29.28 %, clearly indicating that some plots present

acidity problems and highlighting the weakness of the model as regards this particular aspect;

hence the need for more detailed research to improve the model.

Climate has also been identified as a determining factor for SI in teak (e.g. Vásquez and

Ugalde 1994; Montero 1999). However, the high variability of site index in the limited area of

the present study, where a relatively homogeneous climate can be assumed, suggests that the

influence of soil and other variables, in addition to climate, have an effect on site index.

Moreover, it is likely that SI is influenced by certain soil properties other than those studied in

the present experiment. Hence, in a small region with relatively homogeneous climatic

conditions, a number of variables such as bulk density, water retention capacity or the position

on the slope can result in some plots having more available water than others over the course of

the year. Montes (2012) underlines the need to take soil water dynamics into account in order to

estimate whether Pinus radiata D. Don will respond or not to fertilization. In other words, we

may not find any correlation between nutrition and tree growth if another more limiting variable

is affecting growth, for example, water. However, the climate in the study area is characterized

by high rainfall and the soils are relatively deep with good physical properties which probably

allow them to retain enough water; hence, a slight water deficit is not likely to make a big

difference. Nevertheless, detailed information on climate and physical soil properties in the

study area is lacking.

Due to certain limitations of the present study, it was not possible to establish the considered

values as critical levels sensu stricto, hence the term preliminary critical levels is used. In the

first place, although the relationships of foliar and soil values with MAI H and MAI DBH are

useful as analytical variables, they have an important limitation since these growth variables are

highly dependent on stand age. In addition, several authors have reported increasing or

decreasing tendencies in teak tree growth according to age (e.g. Siddiqui et al. 2009; Fernández-

135

Moya et al. 2013). This would also affect the results of the study as the age of the stands

evaluated ranges from 3 to 8 years. Field experimentation and fertilization trials are considered

the only practical way of deriving critical levels sensu stricto because it is necessary to satisfy

the assumption that all the nutrients, apart from the one being tested, are not limiting (see

Richards and Bevege 1972). Hence, “surveys of existing plantations, even though they can be

very useful for diagnostic purposes, are of little value for establishing critical levels” (Richards

and Bevege 1972). The limitations of the present study reveal the need for a broader study

including fertilization trials in order to determine reliable critical levels sensu stricto for teak

plantations on a regional scale.

136

CAPÍTULO 8

USING MULTIVARIATE ANALYSIS OF SOIL

FERTILITY AS A TOOL FOR FOREST

FERTILIZATION PLANNING

Este capítulo ha sido publicado como:

Fernández-Moya J, Alvarado A, Morales M, San Miguel-Ayanz A, Marchamalo-Sacristán M.

2014. Using multivariate analysis of soil fertility as a tool for forest fertilization planning.

Nutrient Cycling in Agroecosystems 98 (2): 155-167.

(Anexo II)

139

8.1. Introduction

Forest plantation areal extent has globally increased during recent decades, and now it covers

264 · 106 ha, 7% of global forest area, in response to the growing global demand for timber,

pulp, energy and other goods (Evans 2009; FRA 2010). Meanwhile, forest managers have been

increasingly concerned about maintaining high productivity rates through several rotations,

especially in short-rotation plantations, and the relationship between forest nutrition, soil

management and sustainable timber production (e.g., Nambiar 1995; Fox 2000). It has long

been recognized that forest growth depends on the ability of soil to maintain a supply of

required nutrients. However, soil nutrient availability can be modified by management practices,

such as fertilizer use (e.g. Rennie 1955; Miller 1981; Fox 2000). A requirement for fertilization

regimes, to compensate for nutrient export through timber extraction, is the long-standing

specification as indicated by some authors (e.g. Rennie 1955; Worrel and Hampson 1997).

However, as Fölster and Khanna (1997) emphasised, such fertilization provision has been

traditionally neglected. Nowadays, in order to enhance forest productivity, sustain site fertility,

and avoid soil nutrient depletion, fertilization is utilized for intensively managed forests across

the globe (Ballard 1984; Gonçalves et al. 1997).

Assuming that deficient soil nutrients have been identified, fertilization programs should be

designed considering the following aspects: a) what fertilizer to use; b) when to apply it; c) how

much is needed; d) how often to apply it; and, e) by what method to apply it (Ballard 1984;

Bertsch 1998). The current situation in Central America is that fertilization programmes for

forest plantations in most cases have been designed taking into account general rules and a

quick interpretation of soil analyses, based on non-specific critical levels (Bertsch 1998). A

single fertilization recommendation is usually applied to large plantations of several square

kilometres, without taking into account any soil fertility heterogeneity. An important

consequence of the precision agriculture approach was a trend towards heterogeneity of crop

fertilizer application, with modification of formula and rate according to changing requirements

within individual fields, rather than simply considering each field as a whole (Robert 2002).

Such precision farming can be established around (i) site-specific management, e.g.,

140

management focused principally upon soil type heterogeneity within each field, assuming their

microclimate can be considered homogeneous, or (ii) management zones with treatments

specified only at a greater scale, across groups of sites, for cases when budgetary or other

restrictions limit the scope for a wider range of management treatments. A major barrier for site-

specific management is the economic cost of generating a satisfactory soil map (Robert 2002).

Analogously, Fox (2000) observed that “site-specific management is the key to sustaining soil

quality and long-term site productivity” for intensively-managed forest plantations.

The delimitation of ‘stands’ is one of the basic principles of forest management. A ‘stand’ is

regarded as a homogeneous group of trees growing together on a sufficiently uniform site.

Forest management is not as intensive as agriculture can be, and in practice, little consideration

is given to establishing stand-specific nutritional plans. However, forest sites are amenable to

grouping by similarity in their soil fertility, through which managers could delineate nutritional

management areas (groups of stands), and therefore facilitate more efficient and productive

management.

In this study, it was evaluated how effectively multivariate statistical analysis can contribute

to decision-making when used as a tool for analysing soil fertility databases, to classify the

stands of a forest plantation according to their soil fertility, and thereby a specific fertilization

program could be designed for each of the defined groups. The intention is to expose a case

study that illustrates how these analyses can be performed, using data from a specific forest

plantation in Costa Rica. However, the objectives are not to make an interpretation of the results

in terms of nutritional status and quantative fertilization needs of the plantations, evaluate

possible growth responses after the fertilization, or elaborate maps of soil fertility of the

plantations. The aim is just to explore the capabilities of the multivariate techniques and show

the possibility of making groups of the already existing stands according to their soil fertility

similarities, in order to be easily used to improve forest fertilization programs.

141

8.2. Materials and methods

Study area, sites and field sampling

Teak (Tectona grandis L.f.) has been extensively used for forest plantations in Central

America, originally in Costa Rica and Panama (De Camino et al. 2002), and more recently in

Guatemala, El Salvador and Nicaragua. Across the region, teak plantations are intensively

managed in rotations of 20–25 years, usually in carefully selected productive sites, with

commercial volume expected to be around 10 m3 ha

-1 year

-1 (Pandey and Brown 2000; De

Camino et al. 2002). Forest fertilization at establishment has become a common practice for

intensively managed forest plantations in Central America, but fertilization at an intermediate or

even mature age is not a common practice in the region. Notwithstanding the primary

importance of site selection as an issue for teak plantation management, subsequent fertilization

is also necessary. Such amelioration can fulfil the high nutrient demand of teak trees, thereby

maintaining the high nutrient concentrations they exhibit (Drechsel and Zech 1991; Fernández-

Moya et al. 2013), and promoting the productivity and sustainability of production sites (e.g.,

Prasad et al. 1986; Liang et al. 2005; Zhou et al. 2012).

The case study was located on the North Pacific coast of Costa Rica (Figure 8.1), in teak

plantation owned by the Panamerican Woods Ltd. company (hereafter ‘PAW’) which is divided

in two sites: Carrillo (1,040 ha) and Palo Arco (1,488 ha). The climate of the region is classified

as tropical wet forest, following Holdridge’s life zones, with a mean annual rainfall of 2,500

mm, and a dry season of 4–6 months. Most common soils are fertile reddish clayey (Table 8.1),

described as Typic Rhodustalfs mixed with Typic Dystrustepts in Carrillo, and Typic

Haploustalfs mixed with Vertic Haploustepts in Palo Arco, with small clusters of other soils.

Soils are derived from sedimentary limestone and basalt parent material.

The plantations were chosen to be representative of properly-managed teak-planted forests in

Central America. In general, management of these plantations consists of continuous forestry

management activities: fertilization at establishment; weed control; pruning; and thinning (from

approximately 800 trees ha-1

at establishment, to 150–200 trees ha-1

by final felling). The use of

142

clones has become common in recent years. A commercial volume of 100–150 m3 is expected

for this kind of plantation, over a rotation of approximately 20 years.

Through the company’s routine activity, a database was created for the plantations under

study, comprising a total of 195 samples of topsoil (0–20 cm) from across all the different

stands, 75 and 129 from the Carrillo and Palo Arco plantations, respectively. Topsoil (0–20 cm)

nutrient availability estimates are commonly used for forest fertilization planning in Central

America, as fine root absorption is reported to be most active in this soil layer, whether in

plantations of teak, or those of other species (Srivastava et al. 1986; Gonçalves et al. 1997;

Behling 2009). Soil samples were analysed at the Centro de Investigaciones Agronómicas from

the University of Costa Rica (CIA-UCR), to determine the following variables: pH (in water),

exchangeable Ca, Mg, K, P, Fe, Cu, Zn, Mn and acidity. pH was determined in water 10:25;

acidity, Al, Ca and Mg in KCl solution 1M 1:10; P, K, Zn, Fe, Mn and Cu in modified Olsen

solution pH 8,5 (NaHCO3 0,5 N, EDTA 0.01M, Superfloc 127) 1:10. The Effective Cation

Exchange Capacity (ECEC) was calculated as the addition of Ca, Mg, K and acidity

[ECEC=Ca+Mg+K+acidity]. Ca saturation (Ca S.), Mg saturation (Mg S.), K saturation (K S.)

and acidity saturation (A S.) were calculated as the percentage of ECEC relative to each of the

components.

Figure 8.1. Location of the two study sites, Carrillo and Palo Arco, on the north Pacific coast of Costa

Rica (Nicoya Peninsula), comprising two teak (Tectona grandis L.f.) plantations owned by Panamerican

Woods Ltd.

143

Table 8.1. Summary of analysed soil properties for the Panamerican Woods Ltd. teak (Tectona grandis

L.f.) plantations on the north Pacific coast of Costa Rica. Means, standard errors (SE) and coefficients of

variation (CV) are provided. Highlighted values (marked with * and in bold red type) are lower than

adequate reference soil levels (after Bertsch, 1998). The ‘General’ column shows the values when

calculated across all the samples for both plantation sites

Carrillo (n=75) Palo Arco (n=120) General (n=195)

Mean SE CV (%) Mean SE CV (%) Mean SE CV (%)

pH 5.8 0.06 9.7 6.0 0.04 6.4 5.9 0.03 7.9

Ca (cmol (+) L-1) 27 1.11 35.5 26.4 0.74 30.6 26.6 0.62 32.5

Mg (cmol (+) L-1) 7.7 0.44 49.2 8.7 0.35 43.7 8.3 0.27 45.9

K (cmol (+) L-1) 0.2 0.01 63.7 0.2 0.02 84.8 0.2 0.01 77.2

Acidity (cmol (+) L-1) 0.2 0.01 42.4 0.1 0.00 32.3 0.2 0.01 43.3

ECEC (cmol (+) L-1)* 35.1 1.36 33.6 35.4 0.98 30.2 35.3 0.80 31.5

P (mg L-1) 1.5* 0.29 168.3 3.2* 0.43 145.5 2.5* 0.29 159.4

Cu (mg L-1) 7.0 1.21 150.1 12.6 0.61 52.7 10.4 0.62 83.9

Fe (mg L-1) 25.6 2.78 94.2 38.3 3.58 102.3 33.4 2.48 103.7

Mn (mg L-1) 41.4 1.95 40.7 25.3 1.79 77.5 31.5 1.44 63.9

Zn (mg L-1) 15.7 1.99 109.8 3.1 0.25 86.7 7.9 0.89 157.2

A. S. (%) 0.6 0.05 73.0 0.4 0.02 59.6 0.5 0.03 72.1

Ca S. (%) 76.8 0.76 8.6 74.8 0.55 8 75.6 0.45 8.3

Mg S. (%) 22.0 0.75 29.7 24.3 0.56 25.2 23.4 0.46 27.2

K S. (%) 0.6 0.06 88.7 0.5 0.04 88.4 0.5 0.03 88.5

* ECEC = Effective cation exchange capacity.

Multivariate statistical methods

Different multivariate analysis methods were employed for simplifying the data, either

through graphic representation of similarities between plot points (ordination methods), or

through grouping of similar samples into discrete classes (classification techniques) (Oksanen

2010). Both approaches are based on methods to estimate the similarities or dissimilarities

between different objects, based on the values of a set of variables measured on each of the

objects. Selecting the dissimilarity measure is of primary importance to multivariate analysis

(Oksanen 2011). Several distance measures, such as Bray-Curtis, have been considered

appropriate in various ecological community studies, but Euclidean distance is considered to be

the best-disposed dissimilarity measure for this study, as it fulfils the metric properties, is based

upon squared differences, and is dominated by single large differences (Oksanen, 2011). Data

standardization and transformation are critical in the process of selecting between different

methodologies (Kenkel 2006).

144

Principal component analysis (PCA) and non-metric multidimensional scaling (NMDS) are

the most commonly used ordination methods. PCA is based on orthogonal axes, Euclidean

space and linear rotation, with an assumption of normally-distributed data, being analogous to

simple linear regression (Kenkel 2006). NMDS does not require any underlying assumptions of

linearity, and so has emerged as one of the more robust and widely-used techniques, especially

in ecology and related disciplines (Minchin 1987; Kenkel 2006; Oksanen 2011). Conversely,

NMDS does present some disadvantages, in particular: (a) NMDS is unable to interpret the

relative importance of the ordination axes when summarizing the variation of the data; and (b)

NMDS cannot produce a true ordination bi-plot, as variable weights are not determined (Kenkel

2006). While these disadvantages have caused some authors to refrain from advocating adoption

of NMDS (Kenkel 2006), in the context of the objectives and data structure of this study, the

disadvantages were considered to be of negligible importance. Cluster analysis is a customary

classification method which incorporates the calculation of a distance matrix (similar to that

used for ordination methods), from which objects can accordingly be classified. Complete

linkage (or farthest neighbour) hierarchic clustering was considered the best option for our data

and objectives, as this method is based upon maximizing the distance between groups or

clusters (Oksanen 2010).

Data analysis

The topsoil database was used to perform different multivariate analyses of the soil test data

in order to group the sampled soils according to similarities between the measured properties.

The use of different multivariate analysis methods allowed comparing their usefulness for

grouping similar soil samples (Table 8.2). One set of analyses was carried out with the soil test

variables centred using their means, along with an alternative set of analyses that instead used

the soil test critical levels to centre the variables (Bertsch 1998). In both cases, each variable

was standardized using its standard deviation. PCA was performed with the entire dataset,

comprising the 195 samples from both plantation sites. NMDS was also performed with the

general dataset (this analysis is hereafter referred to as the ‘G-NMDS’). Additionally, two

145

NMDS analyses were constructed: (a) one analysis for the 75 samples from the Carrillo

plantation (‘C-NMDS’); and (b) a second analysis for the 120 samples from Palo Arco (‘PA-

NMDS’). Five cluster analyses were carried out using the entire dataset (195 samples from both

plantations), in order to distinguish: (a) two groups, (b) three groups, (c) four groups, (d) five

groups, and (e) six groups. Four additional cluster analyses were computed, two for the Carrillo

and two for the Palo Arco plantations, respectively, in order to distinguish two and three groups

for each plantation.

The coefficient of variation (hereafter ‘CV’) was calculated for each variable in each of the

constructed soil groups, and the average CV for each group was determined as:

CVj = average CVi j

where CVi j is the CV for each of the study variables (i) for each group (j).

The reduction of the CV for each variable in each of the constructed soil groups relative to

the original CV for the 195 samples (CVi general) was estimated as:

ΔCVi j = CVi j / CVi general

The average reduction of the CV of the study variables for each group was calculated as:

ΔCVj = average ΔCVi j (3)

The CV calculations allowed an estimate of the homogeneity for each of the groups

identified; a comparison to be made against the null hypothesis of ‘no-groups’; and to identify

which method resulted in the best grouping.

NMDS and cluster analysis were done using the Vegan library in R (R Development Core

Team, 2011). Euclidean distance was used as the measure of dissimilarity. No rotation was used

for the PCA or the NMDS analyses. For the NMDS analyses, the number of k dimensions was

set to k=2. A Shepard diagram ‘stress-plot’ was constructed as a measure of the goodness of fit

for the NMDS analysis (Oksanen 2011).

146

Table 8.2. Summary of the different multivariate analyses performed in the study

Type of analysis Origin of the data Number of samples Name Number of groups Reference for centering

PCA General 195 G-PCA — average — critical value

NMDS

General 195 G-NMDS — average

— critical value

Carrillo 75 C-NMDS — average

— critical value

Palo Arco 120 PA-NMDS — average

— critical value

Cluster

General 195

G-2 2 average

critical value

G-3 3 average

critical value

G-4 4 average

critical value

G-5 5 average

critical value

G-6 6 average

critical value

Carrillo 75

C-2 2 average

critical value

C-3 3 average

critical value

Palo Arco 120 PA-2 2

average

critical value

PA-3 3 average

critical value

8.3. Results and discussion

No important disparities were found (data not shown) between the results of multivariate

analyses obtained using the mean or the critical value as a reference for centring the data

(Bertsch 1998). Therefore, we hereafter describe only the results of the former analyses, i.e.,

from normally-standardized data that used mean and standard deviation as references. Similarly,

minor differences were observed between the PCA and the NMDS constructed through the

‘general’ analyses, using the data from both plantations (data not shown). The NMDS provided

the best representation of the differences between soil samples, and the Shepard plot showed a

non-metric fit to be better than a linear fit (non-metric fit pseudo R2=0.978; linear fit pseudo

R2=0.936), consolidating our interpretation of NMDS as the better ordination method. Hence,

only NMDS analysis was carried out for the Carrillo and Palo Arco plantation data

independently.

Palo Arco plantation has generally been considered to exhibit higher soil fertility than

Carrillo plantation, to the extent that different nutritional management plans have been designed

for the two plantations. However, the average soil data results for the plantations in Carrillo and

Palo Arco showed similar values (Table 8.1), with the possible exception of the P and Zn.

147

Furthermore, the soil samples from Carrillo could not be differentiated from those of Palo Arco,

when we investigated the similarities between soil samples using the ‘general’ NMDS analyses

(Figure 8.2). This contradiction shows how the traditional methods being used nowadays in

many large forest plantations in Central America can be improved using new techniques and

how this improvement could result in a more appropriate soil and nutrient management in those

ecosystems.

Cluster analyses were used to distinguish the following: two, three, four, five and six groups

of soil samples from the entire dataset in general; and, two and three groups from independent

analyses of Carrillo and Palo Arco data, respectively (Table 8.2). Figure 8.2 represents the ‘G-

NMDS’ general analysis across all 195 samples, plotted in accordance with the plantation field

(Carrillo or PaloArco), while Figure 8.3 does it according to the groups defined by cluster

analysis.

Table 8.3 summarizes the trend in CV that was evident as more groups were differentiated

by cluster analysis: with increasing the number of groups, each group became more

homogeneous, and the CV for each variable diminished. As repeated cluster analyses

progressively distinguished more groups, each corresponding group could then be

diagrammatically isolated within the NMDS (Figure 8.3). Hence, a small increase in operational

cost would allow an improvement of fertilizer efficiency and it would translate into a higher

economic return. However, from a theoretical point of view, at some number of these nutritional

groups, the increase (marginal) in benefit should be equal to the increase (marginal) in cost and

further increase in the number of groups should result in negative increments of the benefits. In

addition to this economical reason, the amount of groups cannot be higher than a reasonable

number in order to be practical for the company managers; groups in excess of this number

would ultimately contribute to generate disproportionate complexity in this approach to forest

management, to the extent that we could anticipate abandonment of such practices. We judged

that a maximum of six groups was an appropriate number of soil groups for a 2,500 ha

plantation.

148

Figure 8.2. Non-metric

multidimensional scaling

(NMDS) for the 195

topsoil samples from the

teak (Tectona grandis

L.f.) plantations owned by

Panamerican Woods Ltd.,

on the north Pacific coast

of Costa Rica: Carrillo (

) and Palo Arco ( ).

Figure 8.3. Graphical representation of (a) two, (b) three, (c) four, (d) five, and (e) six groups,

as defined by cluster analysis, based on the spatial scores of the non-metric multidimensional

scaling (NMDS) for the 195 topsoil samples from the teak (Tectona grandis L.f.) plantations

owned by Panamerican Woods Ltd., on the north Pacific coast of Costa Rica.

149

Table 8.3. Reduction in the average coefficient of variation (CV) for soil fertility variables in each group,

distinguished by the cluster analysis treatments, relative to the null hypothesis (‘no-grouping’, i.e., one

single group of data encompassing both Carrillo and Palo Arco together). Highlighted values marked with

* and in bold blue type show a relative CV reduction of between 20–35 %. Highlighted values marked

with ** and in bold green type show a relative CV reduction of more than 35%.

Multivariate statistical techniques have been widely applied in soil sciences, notably in the

analysis of metal contamination (e.g., Yay et al. 2008), but also in precision agriculture,

delineation of site-specific management zones, and soil classification and mapping (e.g.,

Theocharopoulos et al. 1997; Kalähne et al. 2000; Jaynes et al. 2005; Ortega and Santibáñez

Group Average CV (%) Δ average CV (%) Number of soil

samples in the group

Null hypothesis

(no-grouping) 66.8 ---- 195

Grouping by plantation Carrillo 66.5 -0.4 75

Palo Arco 58.3 -12.7 120

G-2 Group 1 60.2 -10.5 158

Group 2 56.2 -9.7 37

G-3

Group 1 55.8 -14.1 157

Group 2 56.2 -9.7 37

Group 3 --- --- 1

G-4

Group 1 60.2 -10.5 157

Group 2 52.3 -17.4 35

Group 3 51.8 -22.5* 2

Group 4 --- --- 1

G-5

Group 1 60.2 -10.5 157

Group 2 40.2 -40.9** 19

Group 3 38.2 -36.6** 16

Group 4 51.8 -22.5* 2

Group 5 --- --- 1

G-6

Group 1 35.7 -45.5** 5

Group 2 40.2 -40.9** 19

Group 3 55.0 -17.7 152

Group 4 38.2 -36.6** 16

Group 5 51.8 -22.5* 2

Group 6 --- --- 1

C-2 Group 1 -2.8 -2.7 74

Group 2 --- --- 1

C-3

Group 1 -19.2 -16.3 42

Group 2 -17.9 -21.4* 32

Group 3 --- --- 1

PA-2 Group 1 -14.0 -26.8* 110

Group 2 -13.3 -29.0* 10

PA-3

Group 1 --- --- 1

Group 2 -17.4 -31.4* 109

Group 3 -13.3 -29.0* 10

150

2007; Yan et al. 2007; Fu et al. 2010; Arrouays et al. 2011). Fu et al. (2010) identified clustering

as an appropriate analytical technique to delineate soil nutrient management zones, and

therefore it provides an effective basis to establish variable-rate fertilization regimes for

precision agriculture. However, Fu et al. (2010) also noted that clustering methods were

sensitive to the iterative initial value. For the Vegan package in the R-software environment (R

Development Core Team 2011), this issue can be addressed by using the metaMDS function,

which allows establishing several random starts, and selects from similar solutions with smallest

stresses (Oksanen 2011). Ortega and Santibáñez (2007) identified cluster analysis as a better

technique for delineating homogeneous management zones, relative to alternative methods.

Multivariate techniques have also been applied to precision agriculture in association with

geostatistical techniques (e.g., Castrignanò et al. 2005; Morari et al. 2009; Arrouays et al. 2011).

However, in the present context of fertilization management for Central American forest

plantations, nowadays we do not consider analysis from this perspective to be justified, given

the degree of complexity associated with the techniques. Rather, we consider that the proposed

strategy, classifying stands into groups with similar soil properties, affords greater scope for

organizing the already existing stands into management zones, given that it readily facilitates

identification of a limited number of nutritional management groups. As the stands are

considered as a homogeneous unit, no further detail taking into account geostatistics,

geographical location or spatial analysis is considered necessary at this time. The delineation of

intra-field management zones, i.e., zones of uniform management, has been assessed as an

important initial stage in the implementation of site-specific nutrient management (Ortega and

Santibáñez 2007).

In the context of this study, NMDS emerged as a better ordination method than PCA

(Figures 8.2 and 8.3). However, it was important to initially consider both PCA and NMDS for

these analyses, in order to identify the method most appropriate for the data and objectives in

question, as the best option can be anticipated to vary on a case-specific basis (Kenkel 2006).

When we distinguished six groups from the entire dataset in general, the groups were more

homogeneous than those that emerged when independently deriving three groups from the

151

Carrillo data and three from Palo Arco (Table 8.3). As no notable differences were evident

when making comparisons between Carrillo and Palo Arco data (Figure 8.2), analysing these

data independently was not considered a useful basis for further similar analyses.

Relatively high microelement concentrations are typically required in order to maintain an

appropriate nutritional status for trees in teak plantations, and indeed other forest plantations

globally (Gonçalves et al. 1997; Lehto et al. 2010; Zhou et al. 2012; Fernández-Moya et al.

2013). However, little attention has been paid to Zn and B in other studies of teak nutrition.

Tropical soils are usually characterized as highly-weathered, and rich in Fe or Mn, but generally

deficient in Zn, B, Cu and Mo (Barker and Pilbeam 2006). B is typically deficient in soils on a

global scale, and is difficult to evaluate in routine soil fertility analyses (Lehto et al. 2010).

There is therefore still a requirement to implement specific evaluations of B and Zn status for

forest plantations throughout the tropics. The advantages of multivariate analysis techniques are

of particular relevance in this respect. Multivariate analyses can be used to process a large

number of variables, and can therefore readily incorporate the range of micronutrients that

fertilization planning must take into account.

“The amount of fertilizer to be applied to a given species at a particular site will depend on

the level of soil fertility and productivity” (Gonçalves et al. 1997). However, practical

management of any fertilization program established on an explicitly stand-specific basis,

applying a different fertilization formula and dosage to each stand, is generally regarded as

impractical by forest plantation managers. As a contrasting approach, we propose that grouping

stands by similarities in soil fertility represents a more practical strategy, in that it facilitates the

allocation of sites into a manageable range of soil fertility classes. This process of classification

promotes the design of a versatile fertilization regime that is sufficiently proximate to differing

soil requirements across all the sites in question. Analysis of soil data in this context of grouped

samples allows us to carry out soil fertility diagnosis with greater precision, and establish a

basis for improved nutritional and fertilization planning. This is exemplified by the

improvement in precision presented by the results in Tables 8.3 and 8.4.

152

Table 8.4. Average and standard error (SE) of soil attributes for the six stand groups defined for the

Panamerican Woods teak (Tectona grandis L.f.) plantations in Guanacaste (Costa Rica) based on their

similarities in the analysed soil fertility attributes. Highlighted values (marked with * and in bold red

type) are lower than adequate reference soil levels (Bertsch, 1998). Coefficients of variation (%) for each

mean are also provided. Δ CV (%) is the reduction in the coefficient of the variation for the variable when

comparing the proposed grouping against the null hypothesis of one single group that includes all the

stands in the dataset.

Group 1 Group 2 Group 3 Group 4 Group 5 Group 6

(n=5) (n=19) (n=152) (n=16) (n=2) (n=1)

mean

and SE

CV

[%]

ΔCV

[%]

mean

and SE

CV

[%]

ΔCV

[%]

mean

and SE

CV

[%]

ΔCV

[%]

mean

and SE

CV

[%]

ΔCV

[%]

mean

and SE

CV

[%]

ΔCV

[%]

mean

and SE

CV

[%]

ΔCV

[%]

pH 7.0

(0.1) 4 -44

6.1

(0) 3 -61

5.9

(0) 7 -12

5.4*

(0.1) 6 -23

5.1*

(0.1) 2 -70

7.5

(---) --- ---

Ca

[cmol+ L-1]

42.8

(1.72) 9 -71

25.2

(0.93) 16 -51

27.8

(0.59) 26 -19

13.3

(0.76) 23 -29

4.9

(0.56) 16 -52

53.2

(---) --- ---

Mg [cmol+ L-1]

3.6 (0.58)

36 -22 5.3

(0.38) 31 -32

9.5 (0.28)

36 -21 3.3

(0.16) 19 -59

2.3 (0.59)

36 -23 4.3 (---)

--- ---

K

[cmol+ L-1]

0.2

(0.03) 45 -41

0.4

(0.05) 50 -36

0.1*

(0.01) 61 -21

0.2

(0.02) 65 -15

0.2

(<0.01) 0 -100

0.6

(---) --- ---

A [cmol+ L-1]

0.1 (0.02)

31 -29 0.2

(0.01) 27 -39

0.2 (0.01)

45 4 0.2

(0.02) 43 -2

0.2 (0.09)

61 40 0.1 (---)

--- ---

CICE

[cmol+ L-1]

46.7

(1.5) 7 -78

31.0

(1.2) 17 -45

37.6

(0.8) 26 -19

16.9

(0.8) 20 -37

7.6

(1.2) 23 -28

58.2

(---) --- ---

P [mg L-1]

2*

(0.3) 35 -78

5*

(1.3) 114 -28

2*

(0.2) 114 -29

2*

(0.3) 54 -66

2*

(2) 140 -12

41 (---)

--- ---

Cu

[mg L-1]

1

(0.2) 41 -51

11

(1.5) 58 -30

9

(0.5) 67 -20

32

(2) 25 -70

8

(6.7) 119 42

8

(---) --- ---

Fe [mg L-1]

7*

(1.6) 52 -50

32 (3.6)

49 -53 25

(1.5) 73 -29

121 (13)

43 -58 38

(22.8) 85 -18

22 (---)

--- ---

Mn

[mg L-1]

22

(0.6) 6 -91

24

(3.6) 66 4

30

(1.5) 61 -5

58

(5.5) 38 -41

58

(32.8) 80 26

5

(---) --- ---

Zn [mg L-1]

1*

(0.7) 146 -7

4 (0.6)

69 -56 9

(1.2) 160 2

8 (1.8)

91 -42 18

(16.8) 132 -16

3 (---)

--- ---

A. S.

[%]

0.3

(0) 34 -53

0.5

(0) 32 -55

0.4

(0) 52 -28

0.9

(0.1) 51 -29

2.6

(0.8) 41 -43

0.2

(---) --- ---

Ca S. [%]

91.5 (1.2)

3 -60 81.5 (0.7)

4 -51 74.1 (0.4)

7 -13 78.1 (1.2)

6 -32 65.3 (3.2)

7 -15 91.4 (---)

--- ---

Mg S.

[%]

7.9

(1.3) 38 39

16.9

(0.7) 18 -34

25.1

(0.4) 22 -21

20.2

(1.1) 21 -21

29.9

(3) 14 -50

7.4

(---) --- ---

K S. [%]

0.3 (0.1)

47 -47 1.3

(0.1) 49 -45

0.4 (0)

69 -22 0.9

(0.2) 67 -24

2.2 (0.4)

23 -75 1

(---) --- ---

In Table 8.1, which deals with traditional methods for fertilization planning, the soils are

presented only as being P-deficient. Hence, if forest managers only take this into consideration,

they would design a fertilization programme to solve this deficiency (e.g. application of a

phosphorus fertilizer) to the entire plantation area. In comparison, the proposed more detailed

grouping analysis (Table 8.4) indicates that most groups exhibit additional deficiencies. Group 1

shows P, Fe and Zn deficiencies; groups 4 and 5 have low pH values in addition to low values

153

of P; group 3 (representing the majority of the samples) shows low K and P content.

Conversely, group 6 indicates exceptional soil that is extraordinarily high in all nutrients. Only

group 2 still accords with the results for Table 8.1, in that it demonstrates a deficiency only in P

but with some relatively high pH; thus it is the only group which would have a similar

fertilization compared with the initial scenario. On the other hand, a fertilizer formula with P, Fe

and Zn would be applied for the stands in group 1; a common N-P-K formula would probably

be applied to group 3, while some specific P fertilizer would be needed for groups 4 and 5 with

a relatively high basicity index in order to solve the relative low pH values or a common P

fertilizer can be applied with previous liming of those stands. Hence, with the traditional

methods for fertilization planning the majority of the stands would have a hidden nutrient

deficiency that would be lowering the productivity of the plantations, except group 6 (a single

stand) that shows high soil fertility with no need to be fertilized. This single stand could have

been considered as a statistical outlier; however, it has been considered that a whole stand

cannot be removed from the analysis as in the decision-making process done by forest

managers; something needs to be done with every stand, even if it is quite different to the

others.

154

CAPÍTULO 9

IS N-P-K FERTILIZATION OF TEAK

(Tectona grandis L.f.) PLANTATIONS

ALWAYS A GOOD CHOICE?

157

9.1. Introduction

The global importance of short rotation, intensively managed planted forests has increased

over recent years due to the growing need for timber and other goods (Evans 2009). In this kind

of system, nutrient management is a key issue and fertilization plays a double role: a) improving

productivity, and b) compensating nutrient output in order to attain sustainability and maintain

productivity for further rotations. The need to replace nutrients taken up by the growing forest

and removed from the site during timber extraction has long been recognized (Rennie 1955).

However, Fölster and Khanna (1997) state that conventional forest management has shown a

general lack of concern with regard to this problem, although FSC (2004) and several other

authors recommend the application of fertilizer to sustain productivity in short-cycle plantations

(e.g. Rennie 1955; Gonçalves et al. 1997; Worrel and Hampson 1997). This is especially

important in tropical forests, where nutrient dynamics and tree growth take place more rapidly

than in temperate zones.

Teak (Tectona grandis L.f.) has become an important species in the worldwide quality

tropical hardwood sector, with a total planted area of 4.3 ·106 ha, of which 132,780 ha are in

Central America (3%) and 31,500 in Costa Rica (Kollert and Cherubini 2012). As regards teak

nutrition in Central America, special attention should be paid to N and K, together with Ca,

these 3 nutrients being those most absorbed by teak, although P, Zn, B and Mg may also limit

productivity in planted teak forest (Fernández-Moya et al. 2013, 2014). Teak is generally

considered to be a species with high nutrient requirements: deep, well-drained soils with high

chemical fertility (especially Ca), and acidity saturation values lower than 3% are required for

the successful growth of this species (Montero 1999; Oliveira 2003; Alvarado and Fallas 2004;

Mollinedo et al. 2005; Kumar 2011; Alvarado 2012 b; Zhou et al 2012). Although site selection

is a key issue in teak plantation management, subsequent fertilization is also necessary. Such

treatments are aimed at satisfying the high nutrient demand of teak trees, thereby maintaining

the high nutrient concentrations which they exhibit (Drechsel and Zech 1991; Fernández-Moya

et al. 2013).

158

Despite the international importance of teak plantations, very few nutritional and fertilization

studies have been published, offering generally inconsistent conclusions (see Kumar 2011 and

Alvarado 2012 b). However, the application of fertilizer (usually N-P-K formulas) is a common

practice in teak plantations both in Asia (until intermediate ages) and in Central America (only

at establishment). The present study aims to further our understanding of this issue, providing

valuable information from four N-P-K fertilization trials over a 11 year chronosequence in teak

plantations in the Northern lowlands of Costa Rica. The factors that may influence the response

to fertilization in this kind of tropical planted forests are also critically discussed in this paper.

9.2. Materials and methods

Study area

The study area is located in the Northern Atlantic lowlands of Costa Rica, in the San Carlos

region. Four sites were selected within the Expomaderas teak (Tectona grandis L.f.) plantations

(10.75-10.95ºN, 84.50-84.70ºW, 35-60 m asl). The area is classified as tropical wet forest

according to Holdridge’s life zones (Holdridge 1967), with a humid climate characterized by

mean annual rainfall of 2000-4000 mm (up to 8000 in some areas) and around 3 dry months,

when the potential evapotranspiration is higher than the precipitation. The soils in the study area

are low fertility clayey acidic red soils (Ultisols). Due to the absence of data from the

experimental plots themselves, information on the topsoil (0-20 cm) and foliage nutrient

concentration from nearby plots of the same plantations were used as a reference to estimate the

nutritional deficiencies of the plantations (Tables 9.1 and 9.2). More details with regard to the

soil and foliage nutrient concentration data can be found in Fernández-Moya et al. (2013).

The studied plantations are representative of company-managed teak plantations in the

region. Management of these plantations involves continuous silvicultural activities: land

preparation, fertilization and liming during the establishment (at variable dosages and formulas

depending on the company), weed control, pruning and thinning (approximately from 816-1111

trees ha-1

to 150-200 trees ha-1

at final felling). A commercial volume of 100-150 m3 is expected

for this kind of plantation when harvested at the end of the 20-25 year rotation.

159

Table 9.1 Topsoil (0-20 cm) attributes at three of the teak (Tectona grandis L.f.) plantations where the

fertilization trials were established in San Carlos region, North of Costa Rica

Escaleras 8* Banderas 1* Banderas 3*

Critical values

1 yr 6 yr 10 yr

Published*** Experience of the authors

pH 4.39** 4.90** 5.20** 5.5 5.5

OM [%] 2.6 5.1 3.4

Sand [%] 26 33 17

Silt [%] 16 16 18

Clay [%] 58 51 65

Acidity [cmol (+) L-1] 1.15** 1.49** 0.44 0.5 0.5

Ca [cmol (+) L-1] 5.26 2.40** 3.96** 4 10

Mg [cmol (+) L-1] 3.28 0.76** 1.36 1 3

K [cmol (+) L-1] 0.56 0.06** 0.07** 0.2 0.2

ECEC [cmol (+) L-1] 10.24 4.71** 5.83 5 15

AS [%] 11.23** 31.63** 7.55** 3

P [mg L-1] 12 2** 2** 10 5

Zn [mg L-1] 6 1** 1** 3 3

Cu [mg L-1] 6 7 10 1 1

Fe [mg L-1] 200 202 123 10

Mn [mg L-1] 263 8 16 5

OM: Organic matter. ECEC: Effective Cation Exchange Capacity [ECEC=Acidity+Ca+Mg+K]. AS:

Acidity Saturation [AS=Acidity/ECEC].

*Information was obtained from Fernández-Moya et al. (2013) from nearby plots of the same plantations.

** Values marked in red bold type are higher or lower to references considered as adequate.

*** Critical values as reference levels to evaluate soil fertility in Costa Rica, as reported by Bertsch

(1998) for general crops and by Alvarado and Fallas (2004) for AS in teak plantations.

Table 9.2 Foliar nutrient concentration of trees at three of the teak (Tectona grandis L.f.) plantations

where the fertilization trials were established in San Carlos region, North of Costa Rica

Stand Escaleras 8** Banderas 1** Banderas 3** References from literature***

Age (years) 0.8 2.5 6.3 7.3 10.5 12.0

N [%] 2.53 2.12* 1.92* 2.18 1.57* 1.89* 1.52-2.78

Ca [%] 1.40 1.12* 0.94* 1.51 0.90* 1.28* 0.72-2.20

K [%] 0.95 0.87* 0.44* 0.64* 0.83* 0.84* 0.80-2.32

Mg [%] 0.13* 0.19* 0.21* 0.24* 0.36 0.42 0.20-0.37

P [mg kg-1] 0.15 0.12* 0.14 0.12* 0.11* 0.11* 0.14-0.25

S [mg kg-1] 0.08* 0.13 0.08* 0.13 0.08* 0.11* 0.11-0.23

Fe [mg kg-1] 71* 140 70* 69* 116* 96* 58-390

Mn [mg kg-1] 78 54 43 56 71 48 50-112

Cu [mg kg-1] 17 12 14 10* 8* 11 10-25

Zn [mg kg-1] 31* 23* 23* 25* 27* 35 20-50

B [mg kg-1] 25 16* 19 17* 16* 19 15-45

* Values marked in red bold type are higher or lower to references considered as adequate (Fernández-

Moya et al. 2013) for their respective ages for plantations in Central America

** Information was obtained from Fernández-Moya et al. (2013) from nearby plots of the same

plantations

*** Adapted from Drechsel and Zech (1991) and Alvarado (2012 b)

160

Experimental design and field measurements

The same experimental design was replicated over a 11-year chronosequence in four

independent trials at four sites: Escaleras 8 (1 year-old), Escaleras 3 (3 year-old), Banderas 1 (6

year-old) and Banderas 3 (10 year-old). The experimental design established at each site

consisted of four replicates of all 12 treatments defined in Table 3: 1) dosages of N (0, 80, 160

and 220 g tree -1

of fertilizer; i.e. 0, 26.8, 53.6 and 73.7 g N tree -1

), 2) dosages of P (0, 100, 200

and 300 g tree -1

of fertilizer; i.e. 0, 45, 90 and 135 g P2O5 tree -1

), and 3) dosages of K (0, 80,

160 and 220 g tree -1

of fertilizer; i.e. 0, 48, 96 and 132 g K2O tree -1

). The treatments which

involved using one nutrient were fertilized with the highest dose of the other two nutrients

(Table 9.3). Ammonium nitrate (NH4NO3) was used as N fertilizer (33.5% N), triple super

phosphate (Ca(H2PO4)2·H2O) was used as P fertilizer (45% P2O5) and potassium chloride (KCl)

was used as K fertilizer (60% K2O). Fertilizer was applied on the soil surface in a circle around

the selected trees. Due to their high mobility, the dosages of N and K were split into two

applications; half at the beginning of the wet period and half at the peak of the rainy season.

Plot size was: a) 324 m2 (9x36 m) with 36 trees (3x12) and an effective plot (without the

border effect) of 10 trees (1x10) in the 1 year-old trial; b) 432 m2 (9x48 m) with 48 trees (3x16)

and an effective plot of around 10 trees, taking into account the thinning, in the 3 and 6 year-old

trials; and c) 540 m2 (9x60 m) with 60 trees (3x20) and an effective plot of around 10 trees,

taking into account the thinning, in the 10 year-old trial. Only measurements from the effective

plots were used for all the analyses undertaken in the present study.

Plot establishment and first fertilization were carried out between March and April 2007.

Tree growth measurements were taken in June-July 2007 and one year later (August 2008) to

evaluate the effects of the fertilization after one year. Diameter at Breast Height (DBH) and tree

height (H) were measured and tree volume was estimated using the formula developed by

Expomaderas. A detailed summary of the variables of the trees measured in June-July 2007 can

be seen in Table 9.4. Mean Annual Increment (MAI) and Current Annual Increment (CAI) were

calculated. Growth data from nearby Permanent Plots (PP) was used as a control to compare the

data of the best fertilization treatments with the non-fertilized regular plantation practice.

161

Table 9.3 Fertilizer dosage (g tree-1

) for each treatment at the fertilization trial established in teak

(Tectona grandis L.f.) plantations of 1, 3, 6 and 10 year-old in San Carlos region, North of Costa Rica

Fertilizer

Nutrient

Treatments CaCO3* NH4NO3 Ca(H2PO4)2·H2O KCl

N P2O5 K2O

N0 1 0 300 220 0 135 132

N1 1 80 300 220 26.8 135 132

N2 1 160 300 220 53.6 135 132

N3 1 220 300 220 73.7 135 132

P0 1 220 0 220 73.7 0 132

P1 1 220 100 220 73.7 45 132

P2 1 220 200 220 73.7 90 132

P3 1 220 300 220 73.7 135 132

K0 1 220 300 0 73.7 135 0

K1 1 220 300 80 73.7 135 48

K2 1 220 300 160 73.7 135 96

K3 1 220 300 220 73.7 135 132

* Lime dosage was 1 kg tree-1

at younger trees (1 and 3 year-old) and 1 Mg ha-1

in older trees (6 and 10

year-old)

One and three year-old plantations were limed (85 – 95% CaCO3) (1 kg lime tree-1

) and

fertilized at establishment, while 6 and 10 year-old plantations were limed (1 t lime ha-1

) and

fertilized in 2006, one year before the establishment of the fertilization trials. It is not known

whether the latter plantations received any fertilization or liming prior to 2006. These operations

were carried out before the trials were established and were also applied to the Permanent Plots

(hereafter PP) which were used as a control.

Statistical analysis

Increments in DBH, H and volume between June-July 2007 and August 2008 were

calculated. A total of 12 ANOVA tests were performed to analyze the effect of the treatments

(Table 9.3) on the increment in DBH, H and commercial volume (i.e. three response variables)

in the four independent trials at different ages. When the ANOVA tests were significant, a

Tukey-LSD mean differences test was conducted to determine the best fertilization treatment at

each age. All statistical analyses were performed using R (R Development Core Team 2011).

All statistical tests throughout the text are considered significant with α=0.05, if the contrary is

not stated.

162

Table 9.4 Summarized forest stands characteristics for each treatment at the fertilization trial established

in teak (Tectona grandis L.f.) plantations of 1, 3, 6 and 10 years old in San Carlos region, North of Costa

Rica. The information for the four plots established for each treatment at each age is averaged and the

95% confidence interval and number of sampled trees are reported

Treatment

Escaleras 8

(1 year)

Escaleras 3

(3 years)

Banderas 1

(6 years)

Banderas 3

(10 years)

H (m) DBH (cm)

H (m) DBH (cm)

Volume (m3 tree-1)

H (m) DBH (cm)

Volume (m3 tree-1)

H (m) DBH (cm)

Volume (m3 tree-1)

N0

2.6

(±0.2), n=40

4.1

(±0.3), n=38

13.2

(±0.4), n=56

13.6

(±0.5), n=56

0.0533

(±0.0054), n=56

14.8

(±0.7), n=47

16.3

(±0.7), n=47

0.1328

(±0.0146), n=47

17.1

(±0.6), n=44

20.6

(±0.8), n=44

0.24

(±0.0228), n=44

N80

2.8

(±0.2),

n=34

4.1

(±0.2),

n=34

13.1

(±0.3),

n=61

13.2

(±0.5),

n=61

0.0494

(±0.0049),

n=61

15.2

(±0.7),

n=44

16.6

(±0.6),

n=44

0.1373

(±0.0145),

n=44

18

(±0.5),

n=38

21.2

(±0.9),

n=38

0.2575

(±0.0272),

n=38

N160 2.9

(±0.3),

n=40

4.2 (±0.3),

n=39

12.8 (±0.3),

n=58

13.3 (±0.5),

n=58

0.049 (±0.0053),

n=58

13.3 (±0.5),

n=42

15.6 (±0.7),

n=42

0.1185 (±0.0155),

n=42

18.2 (±0.5),

n=41

22.6 (±1.3),

n=41

0.3106 (±0.0502),

n=41

N220

2.8

(±0.2), n=40

4.2

(±0.3), n=40

13.5

(±0.2), n=64

14.1

(±0.4), n=64

0.059

(±0.0047), n=64

14.2

(±0.6), n=46

15.5

(±0.7), n=46

0.1156

(±0.0127), n=46

17.6

(±0.5), n=38

21.8

(±0.9), n=38

0.2755

(±0.0287), n=38

P0

3.1

(±0.2),

n=37

4.4

(±0.3),

n=37

12.4

(±0.4),

n=52

13.1

(±0.6),

n=52

0.0464

(±0.0066),

n=54

14.7

(±0.8),

n=42

15.3

(±0.6),

n=42

0.1095

(±0.0113),

n=42

17.7

(±0.5),

n=39

21.2

(±1),

n=39

0.2594

(±0.0292),

n=39

P100 3

(±0.3),

n=38

4.4 (±0.3),

n=37

12.4 (±0.3),

n=60

12.8 (±0.4),

n=60

0.0433 (±0.0038),

n=65

14.6 (±0.8),

n=42

15.1 (±0.6),

n=42

0.1076 (±0.0117),

n=42

18.4 (±0.3),

n=38

21.6 (±0.8),

n=38

0.2685 (±0.0241),

n=38

P200

2.9

(±0.2), n=40

4.3

(±0.2), n=39

12.4

(±0.3), n=51

12.7

(±0.5), n=51

0.0425

(±0.0043), n=60

13.2

(±0.6), n=50

15.7

(±0.6), n=50

0.1185

(±0.0122), n=50

18.3

(±0.6), n=30

21.9

(±1.1), n=30

0.2798

(±0.0338), n=30

P300

2.9

(±0.2),

n=40

4.2

(±0.3),

n=40

12

(±0.3),

n=60

12.3

(±0.5),

n=60

0.038

(±0.0043),

n=65

13.8

(±0.7),

n=44

15.5

(±0.7),

n=44

0.1157

(±0.0137),

n=44

17.3

(±0.6),

n=38

20.9

(±0.9),

n=38

0.2478

(±0.026),

n=38

K0

2.5

(±0.2), n=40

3.6

(±0.2), n=38

11.2

(±0.3), n=65

12.3

(±0.5), n=65

0.0354

(±0.0043), n=66

16.3

(±0.8), n=39

16.5

(±0.8), n=39

0.1363

(±0.0169), n=39

16.6

(±0.7), n=36

19.5

(±1), n=36

0.2103

(±0.0294), n=36

K80

2.5

(±0.2),

n=39

3.7

(±0.2),

n=37

11.5

(±0.3),

n=63

12.5

(±0.5),

n=63

0.0375

(±0.0045),

n=63

15.5

(±0.7),

n=44

15.9

(±0.6),

n=44

0.1235

(±0.0128),

n=44

17.2

(±0.6),

n=40

20.4

(±0.9),

n=40

0.2364

(±0.0262),

n=40

K160 2.2

(±0.2),

n=40

3.5 (±0.3),

n=38

11.5 (±0.4),

n=58

12.5 (±0.5),

n=58

0.0377 (±0.005),

n=58

15.3 (±0.8),

n=37

16 (±0.6),

n=37

0.1235 (±0.0125),

n=37

16.1 (±0.7),

n=54

19 (±1),

n=54

0.2023 (±0.0267),

n=54

K220

2

(±0.2), n=38

3.3

(±0.2), n=34

11.4

(±0.3), n=63

12.3

(±0.4), n=63

0.0358

(±0.0041), n=63

16

(±0.7), n=48

16.2

(±0.7), n=48

0.1309

(±0.0151), n=48

17

(±0.6), n=38

20.2

(±0.9), n=38

0.2284

(±0.0279), n=38

9.3. Results

Fertilization trial in 1 year old plantation

The fertilization trial in the one year old plantation revealed that the treatments had

significant effects on DBH (F11,455=3.06; p<0.001) and H increments (F11,455=2.84; p=0.001), the

best treatments in both cases being those highest in K (K3), showing slight differences from

163

other treatments (Figure 9.1). No H or volume data were available for the nearby PP to be

compared with the fertilized plots, although K3 treatment plots showed slightly higher DBH

compared with unfertilized ones in the PP (Figure 9.2), which means a higher MAI DBH

(Figure 9.3) and CAI DBH (Figure 9.4).

Figure 9.1 Effects of the fertilization treatments established in teak (Tectona grandis L.f.) plantations of

1, 3, 6 and 10 year-old in San Carlos region (North of Costa Rica) on the increment of height, diameter at

breast height (DBH) and volume over one year of growth (i.e. at 2, 4, 7 and 11 years old respectively).

See the text for more details on the characteristics of the fertilization treatments (Table 9.3) and the

experimental design. Different letters indicate treatments considered different using Tukey test for mean

differences analysis (α=0.05).

164

Figure 9.2 Comparisons of tree height (m), diameter at breast height -DBH- (cm) and commercial

volume (m3) between permanent plots (as control) and the best fertilization treatments established in teak

(Tectona grandis L.f.) plantations of 1, 3, 6 and 10 year-old in San Carlos region (North of Costa Rica).

Study variables mean and confidence interval (α=0.05) are provided for the considered best fertilization

treatments: K3 at age 1-2, N3 at age 2-3, P2 at age 6-7, and K3 at age 10-11. See the text for more details

on the characteristics of the fertilization treatments (Table 9.3) and the experimental design.

Fertilization trial in 3 year old plantation

Fertilization treatments also had an effect on DBH (F11,699=9.87; p<0.001) and volume

(F11,699=3.73; p<0.001) increments in the 3 year-old trial, the best treatments in both cases being

those with the highest N dosages (N1, N2 and N3 for the DBH and N3 for the volume),

although without large differences from the others treatments (Figure 9.1). The treatments also

had an effect on H increment (F11,699=1.99; p=0.027) although there were no differences

between means (Figure 9.1). H, DBH and volume of plots where the fertilization experiment

was established were higher than in the PP (Figure 9.2). The larger size of the fertilized trees

(prior to fertilization) is probably the reason for their lower growth compared with the

165

unfertilized ones in the PP (Figures 9.2, 9.3 and 9.4), as tree growth is usually lower in larger

and/or older trees.

Figure 9.3 Comparisons of tree Mean Annual Increment of height (m), diameter at breast height -DBH-

(cm) and commercial volume (m3) between permanent plots (as control) and the best fertilization

treatments established in teak (Tectona grandis L.f.) plantations of 1, 3, 6 and 10 year-old in San Carlos

region (North of Costa Rica). Study variables mean and confidence interval (α=0.05) are provided for the

considered best fertilization treatments: K3 at age 1-2, N3 at age 2-3, P2 at age 6-7, and K3 at age 10-11.

See the text for more details on the characteristics of the fertilization treatments (Table 9.3) and the

experimental design.

Fertilization trial in 6 year old plantation

The fertilization treatments also had an effect on the DBH (F11,441=2.56; p=0.004), H

(F11,441=7.18; p<0.001) and volume (F11,441=6.87; p<0.001) increments in the 6 year-old trial, the

P2 treatment being the one which gave the best results, although there were no big differences

from the others (Figure 9.1). The plots where the fertilization trials were established had lower

H and similar DBH and commercial volume to the PP (Figure 9.2). After the application of

fertilizer (considering the P2 treatment), tree height increased to that of the PP and both DBH

166

and commercial volume exhibited greater increments than in the unfertilized plots (Figure 9.2).

This response to fertilization can also be observed in the higher MAI DBH and MAI

commercial volume than the unfertilized plots (Fig 9.3) and in the higher CAI H, CAI DBH and

CAI volume exhibited by the fertilized plots compared with the unfertilized ones in the PP

(Figure 9.4).

Fertilization trial in 10 year old plantation

In the 10 year old trial, the fertilization treatments also had an effect on the DBH

(F11,285=1.84; p=0.047), H (F11,284=8.37; p<0.001) and volume increments (F11,285=2.72;

p=0.002), the treatments N0, N2 and N3 being those with highest height increment; K3 with

highest diameter increment and N0 and K3 with highest volume increment, even though the

differences between treatments were low (Figure 9.1). The plots where the fertilization trial was

established had similar or slightly higher tree H, DBH and volume compared with the PP,

although no important effect of fertilization was noticed one year after the experiment (Figure

9.2). DBH and volume increments are similar between fertilized and unfertilized plots, whereas

tree height seems to decrease after the fertilization, although this is probably an indirect effect of

thinning (Figures 9.2, 9.3 and 9.4).

9.4. Discussion

A general positive effect of liming and fertilization is observed in 1 and 6 year-old

plantations, contrasting with the lack of effect in 3 and 10 year-old plantations (Figures 9.2, 9.3

and 9.4). The positive effect of fertilization on the 1 year-old teak plantation is probably more

associated to a residual effect of liming carried out at establishment than to the fertilization

itself, because although the soils show some acidity problems, soil nutrient availability is

adequate (Table 9.1). On the other hand, the foliar nutrient content of a nearby unfertilized 1

year old plantation shows deficiencies in Mg, P, S, Fe and Zn while a 2.5 year-old plantation in

the same area also showed deficiencies in N, Ca, K and B (Table 9.2). The response of young

teak plantations to liming and fertilization have been reported previously by several other (e.g.

167

Singh 1997; Alvarado and Fallas 2004; Kumar 2011; Alvarado 2012 b), including studies under

nursery conditions (Zhou et al. 2012). The positive effect of liming and fertilization in 6 year-

old plantations might also be explained by the generally low soil fertility observed, with

deficiencies of Ca, Mg, K, P and Zn (Table 1), also corresponding with generally low values of

foliar nutrient content (Table 9.2). A positive effect of fertilization on teak growth was also

reported by other authors for plantations of intermediate ages, slightly younger or older than the

6 year-old plantation in this trial (e.g. Prasad et al. 1986, cited in Kumar 2011; Montero 1995;

Alvarado 2012 b).

Figure 9.4 Comparisons of tree Current Annual Increment (CAI) of height (m), diameter at breast height

-DBH- (cm) and commercial volume (m3) between permanent plots (as control) and the best fertilization

treatments established in teak (Tectona grandis L.f.) plantations of 1, 3, 6 and 10 year-old in San Carlos

region (North of Costa Rica). Study variables mean and confidence interval (α=0.05) are provided for the

considered best fertilization treatments: K3 at age 1-2, N3 at age 2-3, P2 at age 6-7, and K3 at age 10-11.

See the text for more details on the characteristics of the fertilization treatments (Table 9.3) and the

experimental design.

168

The lack of fertilization effect in 3 and 10 year-old plantations was unexpected, as soil and

foliar data reveal some important deficiencies in the 10 year-old plantation (Tables 9.1 and 9.2),

and even though there is no available data for the 3 year-old plantation, we can assume

deficiencies exist there too, at least with respect to soil acidity and P, common to nearly all the

other sites (Tables 9.1 and 9.2). The lack of effect in the 3 year-old plantation could be

explained by the larger size of the trees in the plots where the trials were established in

comparison to the PP (Figure 9.2). In the case of the 10 year-old plantation, it is possible that

one year is insufficient to detect any difference in growth between fertilized and unfertilized

plots and that a period of at least two years would be necessary. This two-year response lag has

previously been observed in other fertilization studies, and is associated with the so-called

Steenberg effect (e.g. Blinn and Buckner 1989). Another plausible explanation could be that the

nutritional requirements of trees at this age exceed the dosage of fertilizer used in the trial

(Fernández-Moya et al. 2013, 2014). Contrary to the results of the present study, Prasad et al.

(1986, cited in Kumar 2011) found a positive response to fertilization in 10 year-old plantations.

In addition to this unexpected absence of response to fertilization in the 3 and 10 year-old

plantations, the effect on the 1 and 6 year-old plantations, although positive, was much lower

than expected considering the very low soil fertility and generalized nutrient deficiencies

(Tables 9.1 and 9.2). As discussed below, several factors could be influencing this null or low

response to fertilization given that, if the nutrient status of the stand is not the limiting factor to

growth, no fertilization effect is likely to be observed.

Kumar (2011) highlighted several possible causes which could explain the lack of

fertilization response: a) promoting growth of competing understorey, b) enhance palatability

and herbivore pressure, and c) original high fertility of the sites. None of these explanations tie

in with the observations of this study as weed control was carried out properly, no herbivores

were present in the area and soil fertility was generally low (Table 9.1). Montes et al. (2012)

point to the importance of the water status in determining a response to fertilization in forest

plantations as water stress may be more limiting than nutrient deficit. However, the climate in

the study area is characterized by high rainfall and the soils are relatively deep, with good

169

physical properties which probably allow them to retain enough water; hence, a slight water

deficit is not likely to make a big difference. Nevertheless, detailed information on climate and

physical soil properties in the study area is lacking and should be taken into account both in

forest nutrition management and in further research work.

Another factor which could be affecting the response of teak to fertilization at different ages

in the chronosequence is the thinning regime (e.g. Binkley 1986). As observed in other species,

competition rather than tree nutrition can become the limiting factor, and excessive competition

can mask tree response to fertilization. If the plantation is thinned, the plants are released from

competition and are able to use the extra nutrients from the fertilization to grow (Fôlster and

Khanna 1997; Kumar 2011). The 3 year-old plantation where the fertilization trial was

conducted was unthinned (1111 trees ha-1

), while the density of the PP was 865 and 855 trees

ha-1

at the ages of 3 and 4 years respectively; which may explain the absence of fertilization

response in this trial. However, the 10 year-old plantation in which the fertilized trial was

established had been thinned to a similar density as the control PP.

The previous application of lime and fertilizer, only one year before in the 1, 6 and 10 year-

old plantations and three years before in the 3 year-old plantation, could also be affecting tree

response to fertilization. This application of lime and fertilizer in previous years was also

performed in the permanent plots and may have been sufficient to amend the nutritional

deficiencies of the plantations, which is why only one year later, the 10 year-old plantation did

not respond to fertilization. However, the 1 and 6 year-old plantations did respond to

fertilization only one year after the previous liming and fertilization operations. Possible

explanations for this result are: a) the nutritional deficiencies at these sites were so great that the

treatments applied were insufficient; b) the dosages applied in these operations were inadequate

or for some reason the plants were not able to use the extra fertilization at that time (problems

related to the time and or form of application or high stocking level), c) the growth rates of the

plantations at those stages are so high that the plants are able to use the extra nutrient from the

previous operation as well as that from the second, or d) the observed response is mainly a

residual effect of the previous operations.

170

Finally, the selection of the product added as fertilizer is another factor affecting the

fertilization response. The application of N-P-K alone may be insufficient to address the

nutrient deficit, since other nutrients such as Mg, Zn or B may be limiting tree growth. These

nutrients are generally deficient (Tables 9.1 and 9.2) and have been identified as important

nutrients for teak nutrition (Fernández-Moya et al. 2013, 2014). In addition, even though P

deficit is common in many soils worldwide, it is often not caused by low P content but by low P

availability caused by P immobilization in organic form or P precipitation as Ca phosphates

formed at high soil pH or Fe and Al phosphates in highly weathered acidic soils (e.g.

Gyaneshwar et al. 2002; Khan et al. 2007). Hence, it has been observed that the addition of P

chemical fertilizers is an ineffective practice which does not solve the soil P deficiency in these

systems because it is rapidly immobilized and/or precipitated. Conversely, it has been found that

the application of biofertilizers is far more effective because it solubilizes and/or mineralizes the

immobilized P, resulting in an increase in P availability and higher plant growth (e.g.

Gyaneshwar et al. 2002; Khan et al. 2007). Corryanti et al. (2007) have described how

mycorrhizal activity promotes phosphatase activity in the rhizosphere and root of teak seedlings,

resulting in higher growth. Similarly, Alvarado et al. (2004) collected mycorrhizae throughout

teak plantations in Costa Rica and proposed the inoculation of seedlings as a way to improve P

uptake and enhance productivity, particularly in acid soils.

The results from fertilization studies in teak plantations in Latin America (Alvarado 2012 b)

are similarly inconsistent when compared with the literature revision undertaken by Kumar

(2011), which mainly focuses on Asiatic and African plantations. Several authors (Mothes et al.

1991; Torres et al. 1993; Montero 1995) found a positive effect of fertilization in Panama and

Venezuela, while Hernández et al. (1990) found no effect in El Salvador. Despite this lack of

consistency in the fertilization trials in teak plantations, fertilizer application is usually

recommended and generally applied (Kumar 2011; Alvarado 2012b). The application of

fertilizer, without taking into account the environmental and silvicultural factors discussed

above, or the application of an inadequate product would be of little or no benefit to plantation

growth.

CAPÍTULO 10

DISCUSIÓN GENERAL

173

10.1. Caracterización de la fertilidad del suelo y la nutrición de las plantaciones

La teca se considera como una especie con altos requerimientos nutricionales y, por tanto,

exigente de suelos con alta fertilidad físico-química para su establecimiento: suelos profundos

(profundidad efectiva > 90 cm), bien drenados, con fertilidad química alta (especialmente

requiere abundancia de Ca disponible) y baja acidez (p.ej. Montero 1999; Alvarado y Fallas

2004; Mollinedo et al. 2005; Kumar 2011; Alvarado 2012 b). Sin embargo, se ha observado que

la teca se ha plantado en una gran diversidad de suelos en Africa y en Asia (Ultisoles, Oxisoles,

Vertisoles, Andisoles, Alfisoles, Inceptisoles y Entisoles), con condiciones de fertilidad de

suelos y de acidez muy variables (p.ej. en un rango de pH desde 3,8 a 7,9) (p. ej. Zech y

Drechsel 1991; Drechsel y Zech 1994; Ombina 2008; Adekunle et al. 2011; Kumar 2011). Los

resultados presentados en la presente Tesis confirman que la teca en América Central también

ha sido plantada en esta gran variedad de suelos.

Esta gran variedad de suelos en los que se ha plantado la especie en América Central causa

que en muchos casos se hayan observado deficiencias de algunos nutrientes que pueden estar

afectando a las plantaciones. Así, se ha encontrado que las plantaciones de teca en América

Central se encuentran comúnmente, salvo algunas excepciones, en suelos con problemas de

acidez y con problemas de fertilidad asociados a la baja disponibilidad de P y de K

(especialmente teniendo en cuenta la saturación de K debido al desequilibrio que se produce por

la abundancia generalizada de Ca) (Figura 3.3 y Tablas 3.1 y 3.2).

Estos problemas comunes de fertilidad relativos a la baja disponibilidad de P y de K en el

suelo (capítulo 3) están directamente relacionados con las relativamente bajas concentraciones

foliares de estos elementos (0,88±0,07% K y 0,16±0,04% P) en los árboles muestreados para la

curvas de absorción de nutrientes en plantaciones de teca en América Central (capítulo 4). Estos

valores, aunque se consideran adecuados, están en la parte más baja de los rangos considerados

como adecuados por otros autores: entre 0,80 y 2,32% para el K y entre 0,12 y 0,25% para el P

(Drechsel y Zech, 1991; Boardman et al. 1997). Sin embargo, la metodología de muestreo para

la obtención de esas curvas de absorción puede estar influyendo en estos resultados ya que se

escogieron árboles dominantes y codominantes que presentaban un buen crecimiento y son

174

representativos, por tanto, de plantaciones con un buen estado nutricional. Así, puede que

aunque según la caracterización realizada se observen valores relativamente adecuados de

concentraciones foliares de P y de K, haya abundantes plantaciones que presenten valores por

debajo de los rangos considerados. De hecho, aunque los valores de P se consideran adecuados

según lo mencionado anteriormente, en la Figura 10.1 se puede observar como muchos de los

árboles muestreados presentan valores más bajos que el medio, compensados a nivel estadístico

por unos pocos árboles con valores superiores a 0,20%.

Esta mencionada baja disponibilidad de K y P, común en los suelos plantados con teca en la

región (capítulo 3), y la relativamente baja concentración de estos nutrientes en el follaje

comentada anteriormente (capítulo 4) contrastan con la gran cantidad de estos nutrientes que se

acumulan en la biomasa aérea de las plantaciones de teca a lo largo de una rotación (capítulo 5).

Este contraste puede deberse a varias causas: (1) que los árboles muestreados (dominantes y

codominantes) pueden estarse beneficiando de condiciones particulares de sitio ligeramente

superiores al resto, teniendo acceso a una mayor disponibilidad de nutrientes quizás por un

sistema radicular más desarrollado; (2) que las raíces de la teca (o su simbiosis con algunas

micorrizas) favorezcan la solubilización de estos nutrientes (especialmente el P) precipitados,

fijados o inmovilizados en el suelo, haciéndolos disponibles; y/o (3) la entrada de nutrientes al

sistema (p.ej. Tabla 2.1) por deposición atmosférica puede estar jugando un papel importante en

la dinámica de estos sistemas.

Figura 10.1. Relación entre la

concentración foliar de P (%) y la edad

de los árboles de teca (Tectona grandis

L.f.) muestreados en plantaciones de

América Central según la localización

de las mismas: Guanacaste, Costa Rica

( ); Zona Norte, Costa Rica ( ) y

Panamá ( ). La línea representa la

media de los valores muestreados

(Table 4.2). Más detalles pueden

encontrarse en el capítulo 4 de la

presente tesis.

175

10.2. Implicaciones para la elección de sitio

Aunque, como se ha comentado anteriormente, la teca ha sido establecida en una gran

variedad de suelos, incluyendo algunos con problemas de fertilidad relativamente severos; en la

actualidad se reconoce que el seguimiento estricto de protocolos de selección de sitio es una de

las herramientas selvícolas principales para la obtención de los rendimientos esperados en

plantaciones de teca, especialmente cuando se trata de grandes inversiones de capital por parte

de empresas multinacionales (p.ej. Segura et al. 2013). En ese sentido, aunque en las décadas de

1980 y 1990 se generaron varios protocolos para la evaluación de tierras a este respecto (Keogh

1987; Müller et al. 1998; Núñez 2001), los resultados muestran que éstos no cumplieron su

objetivo (capítulo 3). Por otro lado, se ha observado que el establecimiento de plantaciones de

teca en determinados sitios puede no resultar tan rentable económicamente como el de otras

especies nativas (p.ej. Terminalia amazonia y Swietenia macrophylla) (Griess y Knoke 2011)

Esto podría interpretarse como una mayor adaptación de estas especies a las condiciones

edafoclimáticas de esos sitios, lo cual debe servir de lección tanto para las grandes empresas

como para los pequeños propietarios: si la calidad de estación no es suficientemente buena

como para el establecimiento de la teca, es mejor optar por otra especie o buscar otra

localización establecer la plantación.

Para evitar estos problemas de fertilidad de suelos mencionados anteriormente (capítulo 3),

se han determinado algunos parámetros que sirven como herramientas para evaluar la capacidad

productiva de un sitio determinado, i.e. niveles críticos (capítulo 7). Así, se ha observado que

cuando la Saturación de Na es mayor de 1,1% la probabilidad de que la plantación tenga un

crecimiento lento (Índice de Sitio 19) o muy lento (IS<18) es muy alta (Figura 7.4). De manera

opuesta, si la Saturación de Na es menor de 1,1% y la Saturación de K es mayor de 3,09%, es

muy probable que la plantación tenga un crecimiento muy alto (IS>24) (Figura 7.4). Aunque

estos niveles críticos son todavía preliminares y han de realizarse más investigaciones para

poder conseguir una herramienta predictiva de mayor confiabilidad, estos resultados presentan

dos aspectos importantes: (a) revelan una intolerancia de la teca a suelos salinos que se cree que

no había sido señalada hasta el momento; y (b) confirma que el K es uno de los elementos clave

176

en la nutrición de las plantaciones de teca en la región. En ese sentido, se confirma que es la

dinámica del K en el suelo la que es determinante ya que, como se comentaba también en el

capítulo 1, en el capítulo 7 lo que se observa es una relación directa entre la Saturación de K y el

índice de sitio. Es decir, lo que importa más es el papel del K en el equilibrio catiónico,

especialmente teniendo en cuenta la común abundancia de Ca en esos suelos, y no el valor

absoluto de su disponibilidad.

Otro aspecto clave para la evaluación de un sitio y su elección para el establecimiento de

plantaciones de teca es la acidez del suelo ya que, como se ha comentado en varias ocasiones, la

teca es una especie muy poco tolerante a la acidez (p.ej. Zech y Drechsel 1991; Drechsel y Zech

1994; Wehr et al. 2010; Alvarado 2012b). En ese sentido, Alvarado y Fallas (2004) encontraron

un nivel crítico de 3% de Saturación de acidez, que se considera como umbral para la

evaluación de sitio y, para su enmienda en caso de que sea necesaria. Esto se corresponde con

algunas de las relaciones negativas observadas entre el Incremento Medio Anual en altura y en

diámetro y la acidez del suelo y la saturación de acidez observadas en el capítulo 7 (Figura 7.3).

Por el contrario, el modelo ajustado entre el índice de sitio (IS) y las variables edáficas muestra

una relación positiva entre la acidez del suelo y el índice de sitio (Figura 7.4). Cabe destacar que

aunque el modelo predice un IS de 21 cuando la acidez de suelo es mayor a 0,75 cmol(+) L-1

y

un IS de 19 cuando es menor que este umbral, no se observa ningún rodal con crecimiento muy

alto (IS>24) con valores de acidez más altos que ese umbral. Esta relación positiva es difícil de

explicar y probablemente sea indicativa de otras relaciones indirectas no tenidas en cuenta por el

modelo (p.ej. una influencia del pH en la mineralización y la dinámica del N). En ese sentido, se

considera un síntoma de la necesidad de mejorar el modelo y realizar más investigación

relacionada con el tema para conseguir unas herramientas con un mejor poder predictivo.

10.3. Implicaciones para la sostenibilidad

El contraste entre la gran acumulación de P y de K y su escasa disponibilidad en el suelo que

se ha comentado anteriormente (apartado 10.1) supone un riesgo para la sostenibilidad del

sistema, al salir de éste gran cantidad de nutrientes mediante la extracción de madera y corteza

177

por los aprovechamientos forestales (capítulo 5). Varios autores han alertado de este riesgo de

empobrecimiento de los suelos por la extracción de nutrientes, siendo de hecho este tema uno de

los que han preocupado a la comunidad forestal desde hace más tiempo ya que incumbe a la

nutrición, la productividad y la sostenibilidad de los bosques (p.ej. Rennie 1955; Binkley 1986;

Fölster y Khanna 1997; Worrel y Hampson 1997).

De manera análoga a lo que ocurre en agricultura, tradicionalmente se ha propuesto la

fertilización como medida para compensar esta salida de nutrientes del sistema (p.ej. Rennie

1955; Fölster y Khanna 1997; Worrel y Hampson 1997; FSC 2004). Por otro lado, también se

ha propuesto habitualmente el descortezado de los árboles en campo como un método para

minimizar la salida de nutrientes del sistema, disminuyendo así la intensidad del problema (p.ej.

Rennie 1955; Fölster y Khanna 1997; Ma et al. 1997). No obstante, esta medida no se aplica en

plantaciones forestales en América Central y es considerada como cara y poco rentable.

En ese sentido, en el capítulo 6 se explora otra alternativa para reducir la cantidad de

nutrientes que se extraen del ecosistema, y minimizar así la intensidad del problema, de forma

sencilla y barata: modificar la época de cosecha. La hipótesis en la que se basa esta alternativa,

expuesta en base a los resultados del capítulo 5 y corroborada por los resultados del capítulo 6,

es la siguiente: gran proporción de algunos nutrientes acumulados en el follaje se retranslocan

durante el proceso de senescencia de las hojas (para una revisión ver Aerts 1996), de manera

que tejidos como el tronco y la corteza tendrían mayor concentración durante la época en la que

los árboles defoliados que, en el caso de la teca en América Central, coincide con la época de

verano (Enero-Mayo, aproximadamente) que es precisamente en la que se realizan los

aprovechamientos forestales.

Los resultados que se exponen en el capítulo 6 corroboran esa hipótesis y muestran que

cambiando la época de cosecha (a Septiembre o Diciembre) se puede reducir entre un 24 y un

28% la salida de N asociada a la extracción de madera, un 29% la de P y entre un 14 y un 43%

la de K. Realizar los aprovechamientos en el mes de Septiembre minimizaría la extracción de

nutrientes en la madera. No obstante, esta época coincide con el período de lluvias y puede

suponer problemas logísticos para los gestores de las plantaciones. Por otro lado, empezar las

178

tareas de aprovechamiento en Diciembre, finalizada la época de lluvias pero antes de que se

produzca la senescencia foliar, supone una gran reducción de la extracción de nutrientes sin

suponer problemas adicionales a los gestores.

En ese sentido, precisamente el P y el K son los nutrientes que más problemas de

sostenibilidad presentaban (capítulo 5) por lo que la medida supone especial relevancia.

Además, en el caso del N, aunque anteriormente no ha sido considerado como un nutriente de

especial riesgo, esto se debe a que no existen datos de disponibilidad del elemento en el suelo

que permitiesen evaluar su sostenibilidad. No obstante, las grandes cantidades del elemento que

se acumulan en la biomasa aérea que se extrae (junto con el Ca es el principal elemento

acumulado) (capítulo 5) hacen que una reducción en su salida del sistema se considere muy

beneficiosa (capítulo 6).

10.4. Implicaciones para la interpretación de análisis foliares

La realización de análisis foliares es una práctica común en sistemas forestales con una

gestión relativamente intensa ya que la concentración foliar refleja de forma precisa la cantidad

de nutrientes que está aprovechando la planta, teniendo en cuenta a la vez la disponibilidad de

ellos en el suelo y la capacidad de la planta para absorberlos en un función de otros factores

(ambientales, selvícolas …) (p.ej. Mead 1984; Dreschel y Zech 1991; West 2006). Pese a las

ventajas que esta práctica supone comparada con los análisis de suelos (ver apartado 2.2.3), la

dificultad que a veces conlleva su interpretación hace que sea una herramienta infrautilizada, al

menos si comparamos las plantaciones de teca en la región centroamericana con otras

plantaciones agrícolas.

En el capítulo 4 se ofrecen unos valores característicos de plantaciones de teca con un buen

estado nutricional, que se pueden usar como referencia para la interpretación de análisis foliares.

Valores por debajo de éstos podrían estar indicando deficiencias. Aún así, los resultados de la

Tesis evidencian la necesidad de continuar con esta línea de investigación para mejorar estas

herramientas de diagnóstico. En ese sentido, la concentración foliar media de K (0.88±0.07%)

para las plantaciones consideradas como características de la región coincide con la parte más

179

baja del rango señalado como adecuado por otros autores (capítulo 4) (Drechsel y Zech 1991;

Boardman et al., 1997). Sin embargo, valores más bajos (0.66±0.03%) encontrados en

plantaciones de teca en Panamá no afectaron significativamente a su productividad (capítulo 7).

Por otro lado, el rango de referencia señalado como adecuado en el capítulo 4 para la

concentración foliar de P (0.16±0.04%), similar al reportado por otros autores en otras regiones

geográficas (Drechsel y Zech 1991; Zech y Drechsel 1991), se ve confirmado por los resultados

que se muestran en el capítulo 7. Así, se propone un umbral de 0,125% como nivel crítico de la

concentración foliar de P en plantaciones de teca de América Central. Plantaciones con

concentraciones foliares de P por encima de este umbral presentan en general un muy buen

crecimiento (IS>24), mientras que las que presentan valores más bajos que este nivel crítico

tienen, en general, crecimientos más bajos (IS<18) (Figura 7.2).

10.5. Implicaciones para el diseño de planes de fertilización

Una vez detectado algún problema relativo a la nutrición de las plantas o a la fertilidad de los

suelos, se suele aplicar alguna enmienda o fertilización para solucionarlo. Los capítulos 8 y 9

tratan, precisamente, de mejorar la eficiencia en el diseño de planes de fertilización en

plantaciones forestales.

La fertilización es una práctica común en las plantaciones de teca en la región (p.ej.

Alvarado 2012 b; Alvarado y Mata 2013). Habitualmente se aplican entre 50 y 150 g por planta

de una fórmula N-P-K (p.ej. 10-30-10 ó 12-24-12) o 5-15 g de un fertilizante de liberación lenta

(p.ej. osmocote) durante el establecimiento de la plantación. Aunque, generalmente, se hacen

uno o varios muestreos de suelos (e incluso foliares) antes de realizar estas aplicaciones, éstos

son frecuentemente infrautilizados y se suele recurrir, pese a contar con esa información, a

recomendaciones generales de fertilización. Además, en plantaciones grandes, de varios

kilómetros cuadrados no se suele tener en cuenta la variabilidad espacial de la fertilidad de

suelos y es común la aplicación de la misma fórmula y la misma dosis a toda la superficie.

El capítulo 8 explora la capacidad del uso de técnicas estadísticas de análisis multivariante

como herramienta para mejorar el diseño de planes de fertilización en plantaciones con una

180

superficie relativamente grande. Estas técnicas permiten trabajar con bases de datos de

abundantes muestras de suelos, con muchas variables por cada muestra, de manera que se

pueden agrupar los rodales en base a sus similitudes en cuanto a la fertilidad del suelo. Así, se

pueden ajustar planes de fertilización más eficientes para cada uno de los grupos homogéneos

de rodales definidos.

Por otro lado, en el capítulo 9 se analiza el efecto de la fertilización sobre el crecimiento de

plantaciones de teca utilizando un estudio de caso que sirve para evaluar la eficiencia de estas

prácticas en la región. Pese a que los suelos de la zona de estudio presentan una fertilidad

generalmente baja y los análisis foliares de plantaciones de la zona muestran varias deficiencias

nutricionales, la aplicación de fertilizantes proporciona resultados contradictorios. Estos

resultados coinciden con la enorme variación del efecto de la fertilización que se puede observar

entre los resultados publicados por distintos autores (para una revisión ver Kumar 2011;

Alvarado 2012 b) (capítulo 9).

Como se discute ampliamente en el capítulo 9 (y en los puntos 2.2.5 y 2.2.6 del marco

teórico), además de las deficiencias mostradas por análisis foliares y de suelos comentados

anteriormente, hay varios factores que influyen en la respuesta de las plantaciones forestales a la

fertilización que deben ser tenidos en cuenta para diseñar planes eficientes de fertilización.

Dado que la competencia de la vegetación acompañante, la presión de herbívoros y el estrés

hídrico no son factores que tengan relevancia en el estudio de caso llevado a cabo, éste parece

indicar que hay dos factores que pueden estar afectando principalmente al diseño de los planes

de fertilización en plantaciones de teca en América Central: (a) el efecto de la competencia

intra-específica (i.e., la densidad de las plantaciones) y (b) la aplicación de fertilizantes con una

incorrecta formulación o una dosis insuficiente.

La densidad excesiva y la competencia intra-específica son factores fundamentales que

afectan al crecimiento de las plantaciones y que pueden convertirse en limitantes, siendo incluso

más restrictivos que la deficiencia de agua y nutrientes. Así, si una plantación tiene una

densidad apropiada, es capaz de aprovechar los nutrientes extras que recibe tras la aplicación de

fertilizantes. Por el contrario, si la densidad es excesiva los árboles no tienen la capacidad de

181

crecer más y el fertilizante aplicado normalmente se pierde por lixiviación al no ser

aprovechado por éstos (p.ej. Binkley 1986; Fôlster y Khanna 1997; Kumar 2011). Por esta

razón, resulta fundamental programar las actividades de fertilización con el resto de prácticas

selvícolas, aparte de para mejorar el coste y el control logístico de la gestión integral de la

plantación, para lograr efectos sinérgicos entre las distintas prácticas.

La aplicación de dosis bajas de N-P-K1 puede no ser suficiente para mejorar

significativamente la nutrición de las plantaciones. En primer lugar, en plantaciones de edad

avanzada (como es el caso de los rodales de 10 años analizados en el capítulo 9) puede que la

dosis aplicada fuese muy baja comparada con las necesidades de los árboles. En segundo lugar,

la aplicación sólo de N-P-K puede resultar ineficiente si (como es el caso de los rodales

analizados en el capítulo 9) existen deficiencias generalizadas de otros nutrientes, destacando

algunos como Mg, Zn y B (capítulo 9). Además, como se ha comentado varias veces a lo largo

de la tesis (capítulos 2, 3 y 7), puede que la aplicación de P mineral no sea efectiva en suelos

ácidos o calizos, abundantes en la región, ya que el elemento precipita como fosfatos de Ca de

Al o de Fe. En estos casos, resultaría más conveniente la micorrización o la aplicación de otro

tipo de productos (i.e. biofertilizantes) que consiguiesen solubilizar el P para que estuviese

disponible para ser absorbido por las plantas (p.ej. Gyaneshwar et al. 2002; Alvarado et al.

2004; Corryanti et al. 2007; Khan et al. 2007).

1 Se entiende por dosis bajas de N-P-K las aplicadas en el estudio de caso mostrado en el capítulo 9 que consisten

en: entre 27 y 74 g N árbol-1, entre 45 y 135 g P2O5 árbol-1 y entre 48 y 132 g K2O árbol-1, con densidades entre los

1111 árboles ha-1 (marco de plantación 3x3 m) y 300 árboles ha-1 en las parcelas con más edad en las que se han

realizado más raleos

182

CAPÍTULO 11

CONCLUSIONES [en español]

185

(1) Aunque se suele asumir que la teca requiere suelos fértiles, nuestros resultados muestran

que esta especie se ha plantado en una gran variedad de suelos, algunos con escasa fertilidad

química. En las plantaciones de teca de América Central son habituales las deficiencias de K y P

además de algunos problemas de acidez.

(2) Los problemas actuales de fertilidad se originan, principalmente, por la mala selección

de sitio. En la actualidad, las empresas prestan una gran atención al cumplimiento de protocolos

de evaluación y elección de sitio. Los pequeños propietarios, que no pueden permitírselo y que

ya suelen tener una parcela de su propiedad, pueden optar por otra especie más adaptada a sus

condiciones edáficas, consiguiendo una mayor productividad.

(3) La baja disponibilidad de P y K en el suelo es causante de las moderadamente bajas

concentraciones foliares de estos elementos en plantaciones de teca características de la región.

(4) Mediante el empleo de modelos presentamos unos valores de concentraciones foliares

de nutrientes característicos de plantaciones de teca con un buen estado nutricional, que pueden

ser usados para la interpretación de análisis foliares y el diagnóstico de deficiencias.

(5) Se presentan modelos de nutrientes acumulados en la biomasa de plantaciones de teca

que permiten estimar la extracción de nutrientes del sistema a través del aprovechamiento de la

madera (claras o cortas finales), lo que hace posible mejorar su eficiencia y sustentabilidad.

(6) Los nutrientes que más se acumulan en la biomasa aérea de plantaciones de teca en

América Central son N, K y Ca. Además, P y K adquieren mayor relevancia, ya que su salida

del sistema por la extracción de madera y su escasa disponibilidad en los suelos hacen que se

presente un importante desequilibrio que pone en riesgo la sostenibilidad del sistema.

(7) Cambiar la época de corta de la teca, de la actual (en Enero-Mayo, época seca con

árboles defoliados) a Septiembre o Diciembre, puede reducir entre un 24 y un 28% la salida de

N asociada a la extracción de madera, un 29% la de P y entre un 14 y un 43% la de K. Al

contrario de lo que ocurre en Septiembre (época de lluvias), comenzar los aprovechamientos en

Diciembre (una vez finalizadas las lluvias pero antes de la senescencia foliar) no presenta

inconvenientes logísticos para los gestores y supone una importante mejora en la sostenibilidad.

186

(8) Se propone un nivel crítico de 0,125% para la concentración foliar de P en plantaciones

de teca de América Central. Concentraciones por encima de este umbral se asocian a un muy

buen crecimiento, mientras que las más bajas muestran deficiencias que lo limitan.

(9) La teca presenta una tolerancia muy baja a suelos salinos, tendencia que no había sido

señalada hasta el momento, siendo muy alta la probabilidad de que la plantación tenga un

crecimiento lento o muy lento cuando la Saturación de Na es mayor de 1,1%.

(10) Se confirma que K es un elemento clave en la nutrición de las plantaciones de teca en la

región centroamericana. Se propone un nivel crítico provisional de 3,09% para la Saturación de

K, por encima del cual es muy probable que la plantación tenga un crecimiento muy alto.

(11) Las técnicas estadísticas de análisis multivariante pueden ser usadas como herramienta

para agrupar los rodales en base a sus similitudes de fertilidad del suelo. Esto permite ajustar

planes de fertilización más eficientes para cada uno de los grupos. Diversificar la gestión en

función de las diferencias edáficas es un primer paso hacia la selvicultura de precisión.

(12) Aunque los análisis foliares y de suelos indiquen la existencia de deficiencias

nutricionales, la fertilización de las plantaciones no siempre proporciona resultados positivos si

no se diseña teniendo en cuenta los factores que influyen en ella. La densidad de las

plantaciones es fundamental, y es imprescindible diseñar los planes de fertilización teniendo en

cuenta la programación de clareos y claras para buscar sinergias entre tratamientos. La habitual

formulación de dosis bajas de fertilizantes tipo N-P-K puede no resultar eficiente en bastantes

plantaciones donde son necesarias dosis más altas y, sobre todo, la aplicación de otros productos

que incluyan otros nutrientes (p.ej. Mg, Zn y/o B) o que mejoren la disponibilidad de nutrientes

que ya se encuentran en el suelo pero inmovilizados (p.ej. micorrizas, biofertilizantes…).

(13) Las líneas de investigación desarrolladas en la presente Tesis Doctoral indican

claramente la necesidad de ser continuadas para conseguir mejores herramientas de diagnóstico

del estado nutricional de las plantaciones, mejores modelos que permitan estimar la calidad de

sitio y predecir la respuesta a la fertilización en función de variables edafoclimáticas y, en

general, mayor conocimiento sobre el sistema que permita una mejor y más efectiva gestión de

la nutrición y de la fertilidad de los suelos en plantaciones de teca en la región.

CAPÍTULO 12

CONCLUSIONS [in English]

189

(1) Even though high soil fertility is assumed as a requirement for establishing teak

plantations, our results show that the species has been planted in a wide variety of soils,

including some of them with low chemical fertility. Teak plantations in Central America usually

have P and K deficiencies and some acidity problems.

(2) Soil fertility problems are mainly originated by the poor site selection performed.

Nowadays, forest companies use to invest in a good site selection with standardized protocols

for land acquisition. Small land-owners, who cannot afford this and who use to have already a

farm, may prefer to choose another species which would be more adapted to the soil properties.

(3) Low P and K soil availability originates moderately low foliar concentration of these

nutrients in characteristic teak plantations in Central America.

(4) We report statistical models representing foliar nutrient concentration values considered

as adequate, which can be used for foliar analysis interpretation and nutrient deficiencies

diagnosis in teak plantations in Central America.

(5) Nutrient accumulation in aerial biomass is also modeled, which allow forest managers

to estimate nutrient extraction by wood removal (either by thinning or final felling), helping to

improve nutrient efficiency and sustainability.

(6) The nutrients with highest accumulation in teak biomass are N, K and Ca. In addition, P

and K become more relevant because their low soil availability cause an important unbalance

which suppose a risk to system sustainability.

(7) Modifying the harvesting time, from the current situation (January-May, dry season) to

September or December, can reduce between 24 to 28% the N output associated with wood

extraction, 29% the P and between 14 to 43% the K. Harvesting in September (rainy season)

may have some logistic problems. On the other hand, starting the harvest in December (after the

rainy period but before leaf senescence) does not have any practical inconvenience and it

suppose an important improvement for the system sustainability.

(8) A critical level of 0.125% is proposed for foliar P concentration for teak plantations in

Central America. Foliar concentration above this value is associated with very high growth,

while lower values indicate a deficiency which may be limiting the teak growth.

190

(9) Teak shows a very low tolerance to soil salinity, which has not been previously reported

by any other author. Thus, when Na Saturation is higher than 1.1%, teak growth can be

anticipated to be poor or very poor.

(10) Potassium is confirmed as a key element regarding teak nutrition in Central America. A

critical level of 3.09% is proposed for K Saturation. Teak plantations in soils with higher

contents may be anticipated to have very high growth.

(11) Multivariate statistical analyses can be used as tools to group forest stands based on

their soil fertility similarities. This may allow forest managers to design more efficient

fertilization plans for each of these groups. Diversifying forest management according to soil

variability is regarded as a first step towards precision forestry.

(12) Even though foliar and soil analyses would point out the existence of nutritional

deficiencies, fertilization may not always would have positive results if it is not designed taking

into account the different factors which affect it. Plantation’s density is a key factor in this sense

and it should be mandatory to design the fertilization plans taking into account the thinning

scheduling in order to look for synergies between different treatments. The usual low dosage of

N-P-K fertilizers may not be efficient either, as higher dosages may be necessary in many

plantations or perhaps other products might be a better choice. To that sense, fertilizers with

more nutrients (e.g. Mg, Zn and/or B) or products which increase availability of elements which

are immobilized in the soil (e.g. biofertilizers or mycorrhizas) may be better choices in many

plantations.

(13) The research lines presented in this Thesis need to be continued in order to build better

diagnosis tools to be used to evaluate the nutritional status of the plantations, fit better models

which allow forest managers to estimate sites quality and predict fertilization response based on

edaphic and climatic variables and, generally, acquire more knowledge about the system which

would allow a better and more effective management of tree nutrition and soil fertility in teak

plantations in the region.

CAPÍTULO 13

BIBLIOGRAFÍA

193

Adekunle V, Alo A, Adekayode F. 2011. Yields and nutrient pools in soils cultivated with

Tectona grandis and Gmelina arborea in Nigerian rainforest ecosystem. Journal of the Saudi

Society of Agricultural Sciences 10:127-135.

Adu-Bredu S, Foua Tape Bi A, Bouillet JP, Kouame Me M, Yamoah Kyei S, Saint-André L.

2008. An explicit stem profile model for forked and un-forked teak (Tectona grandis) trees

in West Africa. Forest Ecology and Management 255 (1-3): 2189–2203.

Aerts R. 1996. Nutrient resorption from senescing leaves of perennials: are there general

patterns? Journal of Ecology 84: 597-608.

Alvarado A. 2012 a. Diagnóstico de la nutrición en plantaciones forestales. En: Nutrición y

fertilización forestal en regiones tropicales (Alvarado A, Raigosa J eds). Asociación

Costarricense de las Ciencias del Suelo, San José, Costa Rica, pp. 93-120.

Alvarado A. 2012 b. Nutrición y fertilización de Tectona grandis Linn. f. En: Nutrición y

fertilización forestal en regiones tropicales (Alvarado A, Raigosa J eds). Asociación

Costarricense de las Ciencias del Suelo, San José, Costa Rica, pp. 313-340.

Alvarado A. 2012 c. Reciclaje de nutrimentos en plantaciones y bosques tropicales. En:

Nutrición y fertilización forestal en regiones tropicales (Alvarado A, Raigosa J eds).

Asociación Costarricense de las Ciencias del Suelo, San José, Costa Rica, pp. 51-96.

Alvarado A. 2012 d. Requerimiento, fertilización y manejo de nutrimentos en plantaciones

forestales. En: Nutrición y fertilización forestal en regiones tropicales (Alvarado A, Raigosa

J eds). Asociación Costarricense de las Ciencias del Suelo, San José, Costa Rica, pp. 125-

139.

Alvarado A, Chavarría M, Guerrero R, Boniche J, Navarro JR. 2004. Características edáficas y

presencia de micorrizas en plantaciones de teca (Tectona grandis L.f.) en Costa Rica.

Agronomía Costarricense 28(1): 89-100.

Alvarado A, Fallas JL. 2004. La saturación de acidez y el encalado sobre el crecimiento de la

teca (Tectona grandis L.f.) en suelos ácidos de Costa Rica. Agronomía Costarricense 28(1):

81-87.

194

Alvarado A, Herrera B. 2012. Caracterización de las tierras para uso forestal. En: Nutrición y

fertilización forestal en regiones tropicales (Alvarado A, Raigosa J eds). Asociación

Costarricense de las Ciencias del Suelo, San José, Costa Rica, pp. 25-49.

Alvarado A, Mata R. 2013. Condiciones de sitio y la silvicultura de la teca. En: Las

plantaciones de teca en América Latina: Mitos y realidades (De Camino R, Morales JP eds).

Centro Agronómico Tropical de Investigación y Enseñanza, Turrialba, Costa Rica, pp. 54-

83.

Álvarez-Álvarez P, Afif-Khouri E, Cámara-Obregón A, Castedo-Dorado F, Barrio-Anta M.

2011. Effects of foliar nutrients and environmental factors on site productivity in Pinus

pinaster Ait. stands in Asturias (NW Spain). Annals of Forest Science 68:497–509

Andrews J, Siccama TG, Vogt KA. 1999. The effect of soil nutrient availability on

retranslocation of Ca, Mg and K from senescing sapwood in Atlantic white cedar. Plant and

Soil 208: 117-123.

Arguedas M, Arias D, Mata R, Herrera W, Calvo J. 2009. Síndrome del decaimiento lento de la

teca (SDLT) en Costa Rica. Kurú, Revista Forestal Centroamericana 6:2

Arias D, Calvo-Alvarado J, de Ritchter D., Dohrenbusch A. 2011. Productivity, aboveground

biomass, nutrient uptake and carbon content in fast-growing tree plantations of native and

introduced species in Southern Region of Costa Rica. Biomass and Bioenergy 35: 1779-

1788.

Arias G. 2004. Análisis del impacto económico y social de las plantaciones forestales en Costa

Rica. FUNDECOR, San José, Costa Rica.

Arrouays D, Saby NPA, Thioulouse J, Jolivet C, Boulonne L, Ratié C. 2011. Large trends in

French topsoil characteristics are revealed by spatially constrained multivariate analysis.

Geoderma 161: 107–114.

Ballard R. 1984. Fertilization of plantations. En: Nutrition of Plantation Forests (Bowen GD,

Nambiar EKS eds). Academic Press, London, pp 327–360.

Barker AV, Pilbeam DJ. 2006. Handbook of Plant Nutrition. CRC Press, USA.

195

Behling M. 2009. Nutrição, partição de biomassa e crescimento de povoamentos de teca em

Tangará da Serra-MT. Ph. D. Thesis. Universidade Federal de Viçosa, Brasil.

Bermejo I, Cañellas I, San Miguel A. 2004. Growth and yield models for teak plantations in

Costa Rica. Forest Ecology and Management189: 97-110.

Bertsch F, Bejarano JA, Corrales M. 2005. Correlación entre las soluciones extractoras KCl-

Olsen modificado y Mehlich 3, usadas en los laboratorios de suelos de Costa Rica.

Agronomía Costarricense 29 (3): 137-142.

Bertsch F. 1998. La fertilidad de los suelos y su manejo. Asociación Costarricense de la Ciencia

del Suelo, San José, Costa Rica.

Binkley D. 1986. Forest nutrition management. Wiley-Interscience publication. New York.

Binkley D, Stape JL, Ryan MG, Barnard HR, Fownes J. 2002. Age-related decline in forest

ecosystem growth: an individual-tree, stand-structure hypothesis. Ecosystems 5:58-67.

Blinn CR, Buckner ER. 1989. Normal foliar nutrient levels in North American Forest Trees.

Station Bulletin 590-1989. Minessota Agricultural Experimental Station. University of

Minessota

Boardman R, Cromer RN, Lambert MJ, Webb MJ. 1997. Forest plantations. En: Plant analysis,

an interpretation manual (Reuter DJ, Robinson JB, eds). CSIRO Publishing, Australia, pp.

505-561.

Brauman K.A., Daily G.C., Duarte T.K. 2007. The nature and value of ecosystem services: an

overview highlighting hydrologic services. Annual Review of Environment and Resources

32: 67-98.

Briceño J, Pacheco R. 1984. Métodos analíticos para el análisis de suelos y plantas. Centro de

Investigaciones Agronómicas, Universidad de Costa Rica, San José, Costa Rica.

Brockerhoff EG, Jactel H, Parrotta JA, Quine CP, Sayer J, Hawksworth D. 2008. Plantation

Forests and Biodiversity: Oxymoron or Opportunity? Biodiversity and Conservation 17(5):

925-951.

Bruijnzeel LAS. 2004. Hydrological functions of tropical forests: not seeing the soil for the

trees? Agriculture, Ecosystems and Environment 104: 185–228.

196

Bruijnzeel LAS. 1991. Nutrient input-output budgets of tropical forest ecosystems: a review.

Journal of Tropical Ecology 7: 1-24.

Buol S, Hole FD, McCraken RJ. 1989. Soil genesis and classification. Iowa State University

Press, Ames, USA.

Buongiorno J. 2013. Assessing the Impact of Planted Forests on Global Forest Economy. III

International Congress on Planted Forests. France-Ireland-Portugal.

Cabalceta G. 1995. Correlación de soluciones extractoras de fósforo en suelos de Costa Rica.

Agronomía Costarricense 19: 29-37.

Canadell JG, Raupach MR. 2008. Managing Forests for Climate Change Mitigation. Science

320 (5882): 1456-1457.

Carle J. 2013. Planted Forests: Wood Production and Enhanced Trade Opportunities through

Certification, Verification and Governance Instruments. III International Congress on

Planted Forests. France-Ireland-Portugal.

Castrignanò A, Cherubini C, Giasi CI, Castore M, Di Mucci G, Molinari M. 2005. Using

multivariate geostatistics for describing spatial relationships among some soil properties.

ISTRO Conference, Brno, Czech Republic.

Corryanti, Soedarsono J, Radjagukguk B, Widyastuti SM. 2007. Phosphatase activity in the

rhizosphere and root of mycorrhizal teak seedlings with three levels of NPK fertilization.

Biodiversitas 8 (2): 204-209.

De Camino R, Alfaro MM, Sage LFM. 2002. Teak (Tectona grandis) in Central America. FAO

Plantations Working Papers FP/19, Rome.

De Camino R, Morales JP. 2013. Las plantaciones de teca en América Latina: Mitos y

realidades. Centro Agronómico Tropical de Investigación y Enseñanza, Turrialba, Costa

Rica.

Díaz-Romeu R, Hunter A. 1978. Metodología de muestreo de suelos, análisis químico de suelos

y tejido vegetal e investigación en invernadero. CATIE, Turrialba, Costa Rica.

Drechsel P, Zech W. 1991. Foliar nutrient levels of broad-leaved tropical trees: a tabular review.

Plant and Soil 131 (1): 29-46.

197

Drechsel P, Zech W. 1994. DRIS evaluation of teak (Tectona grandis L.f.) mineral nutrition and

effects of nutrition and site quality on the teak growth in West Africa. Forest Ecology and

Management 70 (1-3):121-133.

EFIATLANTIC. 2013. Planted forests are a vital resource for future green economies. III

International Congress on Planted Forests. France-Ireland-Portugal

Egnell G. 2011. Is the productivity decline in Norway spruce following whole-tree harvesting in

the final felling in boreal Sweden permanent or temporary? Forest Ecology and Management

261: 148-153.

Evans J, Turnbull JW. 2004. Plantation Forestry in the Tropics. Oxford University Press, United

States.

Evans J. 2009. Planted Forests: Uses, Impacts and Sustainability. CAB International, FAO,

Rome.

FAO. 2009. Situación de los bosques en el mundo. FAO, Roma.

Fernández Moya J. 2013. Efecto de las plantaciones de teca (Tectona grandis L. f.) sobre las

propiedades hidrológicas de los suelos en Puerto Carrillo, Costa Rica. Tesis MSc.

Universidad de Costa Rica. San José, Costa Rica

Fernández-Moya J, Alvarado A, San Miguel-Ayanz A, Marchamalo-Sacristán M. 2011. Uso

sostenible de nutrimentos en plantaciones de teca (Tectona grandis L.f.) en América Central:

implicaciones para la certificación. Conferencia Forestal Internacional “Bosques plantados

de teca - un recurso forestal emergente a nivel global”. CATIE-TEAKNET-FAO, Turrialba,

Costa Rica.

Fernández-Moya J, Murillo R, Portuguez E, Fallas JL, Ríos V, Kottman F, Verjans JM, Mata R,

Alvarado A. 2013. Nutrient concentrations age dynamics of teak (Tectona grandis L.f.)

plantations in Central America. Forest Systems 22(1): 123-133.

Fernández-Moya J, Murillo R, Portuguez E, Fallas JL, Ríos V, Kottman F, Verjans JM, Mata R,

Alvarado A. 2014. Nutrient accumulation and export in teak (Tectona grandis L.f.)

plantations of Central America. Aceptado en iForest - Biogeosciences and Forestry.

198

Fife DN, Nambiar EKS, Saur E. 2008. Retranslocation of foliar nutrients in evergreen tree

species planted in a Mediterranean environment. Tree Physiology 28: 187-196.

Fölster H, Khanna PK. 1997. Dynamics of nutrient supply in plantation soils. En: Management

of Soil, Nutrients and Water in Tropical Plantation Forests (Nambiar EKS, Brown AG eds).

Australian Centre for International Agricultural Research, Australia, pp. 339-379.

FONAFIFO. 2010. Servicios Ambientales. FONAFIFO, San José, Costa Rica. Disponible en

[www.fonafifo.com/paginas_espanol/servicios_ambientales/servicios_ambientales.htm].

Forsythe W. 1975. Física de Suelos. Manual de Laboratorio. IICA, Costa Rica.

Fox J, Weisberg S. 2011. An R companion to applied regression. SAGE, California, USA.

Fox TR. 2000. Sustained productivity in intensively managed forest plantations. Forest Ecology

and Management 138: 187–202.

Fox TR, Miller BW, Rubilar R, Stape JL, Albaugh TJ. 2011. Phosphorus Nutrition of Forest

Plantations: The Role of Inorganic and Organic Phosphorus. Soil Biology 100: 317-338.

FRA. 2010. Global Forest Resources Assessment 2010. FAO, Rome.

FSC. 2004. Perspectivas sobre plantaciones: Desafío para el manejo de plantaciones. Forest

Stewardship Council A.C. Bonn, Germany.

Fu Q, Wang Z, Jiang Q. 2010. Delineating soil nutrient management zones based on fuzzy

clustering optimized by PSO. Mathematical and Computer Modelling 51: 1299–1305.

Gandullo JM. 2000. Climatología y Ciencia del suelo. FUCOVASA, Madrid, España.

Gholz HL, Lima WP. 1997. The ecophysiological basis for productivity in the tropics. En:

Management of Soil, Nutrients and Water in Tropical Plantation Forests (Nambiar EKS,

Brown AG, eds). Australian Centre for International Agricultural Research, Australia, pp.

339-379.

Gonçalves JLM, Barros NF, Nambiar EKS, Novais RF. 1997. Soil and stand management for

short-rotation plantations. En: Management of Soil, Nutrients and Water in Tropical

Plantation Forests (Nambiar EKS, Brown AG eds). Australian Centre for International

Agricultural Research, Australia, pp. 379–418.

199

Gower ST, McMurtrie RE, Murty D. 1996. Aboveground net primary production decline with

stand age: potential causes. Tree 11 (9):378-382.

Griess VC, Knoke T. 2011. Can native tree species plantations in Panama compete with teak

plantations? An economic estimation. New Forests 41:13-39.

Gyaneshwar P, Naresh Kumar G, L. Parekh LJ, Poole PS. 2002. Role of soil microorganisms in

improving P nutrition of plants. Developments in Plant and Soil Sciences 95: 133-143.

Hartemink AE. 2004. Soils of the tropics. Geoderma 123: 373-375.

Hedin LO, Brookshire ENJ, Menge DNL, Barron AR. 2009. The nitrogen paradox in tropical

forest ecosystems. Annual Review of Ecology, Evolution, and Systematics 40:613-635.

Helmisaari H, Hanssen KH, Jacobson S, Kukkola M, Luiro J, Saarsalmi A, Tamminen P, Tveite

B. 2011. Logging residue removal after thinning in Nordic boreal forests: Long-term impact

on tree growth. Forest Ecology and Management 261, 1919-1927.

Helmisaari HS, Siltala T. 1989. Variation in nutrient concentrations of Pinus sylvestris stems.

Scandinavian Journal of Forest Research 4 (1-4): 443-451.

Hernández A, Franco H, Zambrana H, García C. 1990. Fertilización de Tectona grandis L.f. en

la región occidental de El Salvador. Centro de Recursos Naturales, San Salvador, El

Salvador.

Holdridge LR. 1967. Life zone ecology. Tropical Science Center. San José, Costa Rica

Hornbeck JW, Martin CW, Pierce RS, Bormann FH, Likens GE, Eaton JS. 1986. Clearcutting

Northern hardwoods: Effects on hhydrrrologic and ion budgets. Forest Science 32(3): 667-

686.

Horneck DA, Miller RO. 1998. Determination of total nitrogen in plant tissue. En: Handbook of

reference methods for plant analysis (Kalra YP ed). Soil and Plant Analysis Council Inc. and

CRC Press, pp. 75-83.

Ilstedt U, Malmer A, Verbeeten E, Murdiyarso D. 2007. The effect of afforestation on water

infiltration in the tropics: A systematic review and meta-analysis. Forest Ecology and

Management 251: 45–51.

200

Jackson RB, Jobbágy EG, Avissar R, Roy SB, Barrett DJ, Cook CW, Farley KA, le Maitre DC,

McCarl BA, Murray BC. 2005. Trading Water for Carbon with Biological Carbon

Sequestration. Science 310: 1944-1947.

Jacobson S, Kukkola M, Mälkönen E, Tveit B. 2000. Impact of whole-tree harvesting and

compensatory fertilization on growth of coniferous thinning stands. Forest Ecology and

Management 129: 41-51.

Jayaraman K, Bhat KV. 2011. Proceedings of the International Training Programme

“Innovations in the Management of Planted Teak Forests”. TEAKNET/Kerala Forest

Research Institute, Peechi, India.

Jaynes DB, Colvin TS, Kaspar TC. 2005. Identifying potential soybean management zones from

multi-year yield data. Computers and Electronics in Agriculture 46: 309–327.

Johnson EA, Miyanishi K. 2008. Testing the assumptions of chronosequences in succession.

Ecology Letters 11 (5):419-431.

Kalähne R, Amin M, Sobottka J, Sauerbrey R. 2000. Methods of multivariate data analysis

applied to the investigation of fen soils. Analytica Chimica Acta 420: 205–216.

Kamprath EJ. 1984. Crop response to lime on soils in the tropics. En: Soil Acidity and Liming

(Adams F ed). American Society of Agronomy,Wisconsin. pp. 349-369.

Kaul ON, Sharma DC, Tandon VN, Srivastava PPL. 1979. Organic matter and plant nutrients in

a teak (Tectona grandis) plantation. Indian Forester 105 (8): 573-582.

Kenkel NC. 2006. On selecting an appropriate multivariate analysis. Canadian Journal of Plant

Science 86: 663–676.

Keogh R. 1987. The care and management of teak (Tectona grandis L f) plantations.

Universidad Nacional de Costa Rica. Heredia, Costa Rica.

Khan MS, Zaidi A, Wani PA. 2007. Role of phosphate-solubilizing microorganisms in

sustainable agriculture - a review. Agronomy for Sustainable Development 27: 29–43.

Kollert W, Cherubini L. 2012. Teak resources and market assessment 2010. FAO Planted

Forests and Trees Working Paper FP/47/E, Rome.

201

Kumar K. 2014. Temporal patterns of storage and flux of N and P in young teak plantations of

tropical moist deciduous forest, India. Journal of Forestry Research 25 (1): 75-86.

Kumar BM. 2011. Soil management in teak plantations. En: Proceedings of the International

Training Programme “Innovations in the Management of Planted Teak Forests” (Jayaraman

K, Bhat KV eds). TEAKNET/Kerala Forest Research Institute, Peechi, India, pp 24-24-30.

Kumar JIN, Kumar RN, Kumar BR, Sajish PR. 2009. Quantification of nutrient content in the

aboveground biomass of teak plantation in a tropical dry deciduous forest of Udaipur, India.

Journal of Forests Science 55 (6): 251-256.

Kumar PS. 2009. Nutrient dynamics of teak plantations and their impact on soil productivity –

A case study from India. XIII World Forestry Congress, Argentina.

Laclau JP, Bouillet JP, Ranger J, Joffres R, Gouma R, Saya A. 2001. Dynamics of Nutrient

Translocation in Stemwood across an Age Series of a Eucalyptus Hybrid. Annals of Botany

88: 1079-1092.

Laclau JP, Deleporte P, Ranger J, Bouillet JP, Kazotti G. 2003. Nutrient Dynamics throughout

the Rotation of Eucalyptus Clonal Stands in Congo. Annals of Botany 91: 879-892.

Lamb D. 1977. Relationships between growth and foliar nutrient concentrations in Eucalyptus

deglupta. Plant and Soil 47 (2): 495-508.

Lathwell DJ, Grove TL. 1986. Soil-plant relationships in the tropics. Annual Review of

Ecology, Evolution, and Systematics 17: 1-16.

Lattimore B, Smith CT, Titus BD, Stupak I, Egnell G. 2009. Environmental factors in woodfuel

production: Opportunities, risks, and criteria and indicators for sustainable practices.

Biomass and Bioenergy 33: 1321-1342.

Lehto T, Ruuhola T, Dell B. 2010. Boron in forest trees and forest ecosystems. Forest Ecology

and Management 260: 2053-2069.

Liang KN, Pan YF, Liu WM. 2005. Multi-factors fertilization trial of Tectona grandis in

seedling stage. Forest Research 188(5): 535–540.

Lugo AE. 1992. Comparison of tropical tree plantations with secondary forests of similar age.

Ecological Monographs 62: 1–41.

202

Ma X, Heal KV, Liu A, Jarvis PG. 2007. Nutrient cycling and distribution in different-aged

plantations of Chinese fir in southern China. Forest Ecology Management 243: 61-74.

Marland G, Schlamadinger B. 1997. Forests for Carbon Sequestration or Fossil Fuel

Substitution? A Sensitivity Analysis. Biomass and Bioenergy 13: 389-397.

Mead DJ. 1984. Diagnosis of nutrient deficiencies in plantations. En: Nutrition of Plantation

Forests (Bowen GD, Nambiar EKS eds). Academic Press, London, pp. 259-292.

Meerts P. 2002 Mineral nutrient concentrations in sapwood and heartwood: a literature review.

Annals of Forest Science 59, 713-722.

Merino A, Balboa MA, Rodríguez-Soalleiro R, Álvarez González JG. 2005. Nutrient exports

under different harvesting regimes in fast-growing forest plantations in southern Europe.

Forest Ecology and Management 207 (3): 325–339.

Miller HG. 1981. Forest fertilization: some guiding concepts. Forestry 54 (2):157-167.

Miller HG. 1984. Dynamics of nutrient cycling in plantation ecosystems. En: Nutrition of

Plantation Forests (Bowen GD, Nambiar EKS eds). Academic Press, London, pp. 53-79.

Miller HG. 1995. The influence of stand development on nutrient demand, growth and

allocation. Plant and Soil 168-169: 225-232.

Minchin PR. 1987. An evaluation of the relative robustness of techniques for ecological

ordination. Vegetatio 69: 89–107

Molina E. 2012. Características y manejo de los fertilizantes. En: Nutrición y fertilización

forestal en regiones tropicales (Alvarado A, Raigosa J eds). Asociación Costarricense de las

Ciencias del Suelo, San José, Costa Rica, pp. 141-162.

Molina E, Alvarado A. 2012. Manejo de la acidez y el encalado del suelo. En: Nutrición y

fertilización forestal en regiones tropicales (Alvarado A, Raigosa J eds). Asociación

Costarricense de las Ciencias del Suelo, San José, Costa Rica, pp. 163-181.

Mollinedo M, Ugalde L, Alvarado A, Verjans JM, Rudy LC. 2005. Relación suelo-árbol y

factores de sitio en plantaciones jóvenes de teca (Tectona grandis) en la zona Oeste de la

cuenca del canal de Panamá. Agronomía Costarricense 29: 67-75.

203

Montero M. 1995. Dinámica de crecimiento de la teca (Tectona grandis) bajo fertilización en El

Limón de Chupampa, Herrera, Panamá. En: Memorias del Seminario Técnico sobre

Fertilización Forestal, Santiago, Panamá. CATIE/INRENARE, pp. 17-29.

Montero M. 1999. Factores de sitio que influyen en el crecimiento de Tectona grandis L.f. y

Bombacopsis quinata (Jacq.) Dugand, en Costa Rica. Tesis MSc. Universidad Austral de

Chile / CATIE.

Montes C, Ojeda H, Barría B, Burgos P, Allen HL. 2012. Combining soil and environmental

information to predict growth response to Nitrogen fertilizer in adult stands of radiate pine in

Chile. IUFRO Conference “Nutrient Dynamics of Planted Forests”, Vancouver, Washington,

USA.

Morari F, Castrignanò A, Pagliarin C. 2009. Application of multivariate geostatistics in

delineating management zones within a gravelly vineyard using geo-electrical sensors.

Computers and Electronics in Agriculture 68: 97–107.

Mothes M, Cuevas e, Franco W. 1991. Limitación nutricional de fósforo en plantaciones de teca

(Tectona grandis) en los llanos Occidentales venezolanos. Revista Facultad Agronomía

(Maracay) 17: 309-315.

Murillo O, Wright J, Monteuuis O, Montenegro F. 2013. Mejoramiento genético de la teca en

América Latina. En: Las plantaciones de teca en América Latina: Mitos y realidades (De

Camino R, Morales JP eds). Centro Agronómico Tropical de Investigación y Enseñanza,

Turrialba, Costa Rica, pp. 86-111.

Müller S, Núñez J, Ramírez L. 1998. Indicadores para el uso de la tierra: el caso de la cuenca

del río Reventado, Costa Rica. IICA. San José, Costa Rica.

Nambiar EKS. 1995. Sustained productivity of plantations: science and practice. Bosque 16 (1):

3-8.

Nambiar EKS, Brown AG (eds). 1997. Management of Soil, Nutrients and Water in Tropical

Plantation Forests. Australian Centre for International Agricultural Research, Australia.

204

Negi JDS, Bahuguna VK, Sharma DC. 1990. Biomass production and distribution of nutrients

in 20 years old teak (Tectona grandis) and gamar (Gmelina arborea) plantation in Tripura.

Indian Forester 116: 682-685.

Negi MS, Tandon VN, Rawat HS. 1995. Biomasss and nutrient distribution in young teak

(Tectona grandis Linn. f.) plantation in Tarai Region of Uttar Pradesh. Indian Forester

121(6): 455-464.

Nieuwenhuyse A, Hengsdijk H, Bouman B, Schipper RA, Jansen H. 2000. Can forestry be a

competitive land use option? Model simulations from humid tropical Costa Rica. Forest

Ecology and Management 137: 23-40.

Núñez J. 2001. Criterios edáficos en la selección de sitios para teca. San José, Costa Rica.

Nwoboshi LC. 1984. Growth and nutrient requirements in a teak plantation age series in

Nigeria. II. Nutrient accumulation and minimum annual requirements. Forest science 30 (1):

35-40.

Oksanen J. 2010. Cluster Analysis: Tutorial with R. University of Oulu, Finland.

Oksanen J. 2011. Multivariate Analysis of Ecological Communities in R: vegan tutorial.

University of Oulu, Finland.

Ola-Adams BA. 1993. Effects of spacing on biomass distribution and nutrient content of

Tectona grandis Linn. f. (teak) and Terminalia superba Engl. and Diels. (afara) in South-

Western Nigeria. Forest Ecology Management 58: 299-319.

Oliveira JRV. 2003. Sistema para cálculo de balanço nutricional e recomendação de calagem de

povoamentos de teça- NUTRITECA. Tesis MSc. Viçosa Federal University. Viçosa, Brasil.

Ombina CA. 2008. Soil characterisation for teak (Tectona grandis) plantations in the Nzara

district of Southern Sudan. Tesis MSc. Stellenbosch University. Stellenbosch, South Africa.

Ortega RA, Santibáñez OA. 2007. Determination of management zones in corn (Zea mays L.)

based on soil fertility. Computers and Electronics in Agriculture 58: 49–59.

Pandey D, Brown C. 2000. Teak: a global overview. Unasylva 51 (2): 1-15.

Pérez LD, Kanninen M. 2003. Aboveground biomass of Tectona grandis plantations in Costa

Rica. Journal of Tropical Forest Sciences 15 (1): 199-213.

205

Pérez D, Kanninen M. 2005 a. Effect of thinning on stem form and wood characteristics of teak

(Tectona grandis) in a humid tropical site in Costa Rica. Silva Fennica 39(2): 217–225.

Pérez D, Kanninen M. 2005 b. Stand growth scenarios for Tectona grandis plantations in Costa

Rica. Forest Ecology and Management 210 (1-3): 425–441.

Peri PL, Gargaglione V, Martínez-Pastur G. 2008. Above- and belowground nutrients storage

and biomass accumulation in marginal Nothofagus antarctica forests in Southern Patagonia.

Forest Ecology Management 255: 2502-2511.

Poels RLH. 1994. Nutrient balance studies to determine the sustainability of management

systems of natural and plantation forests in Costa Rica. CATIE/AUW/MAG, San José, Costa

Rica.

Prasad R, Sah AKY, Bhandari AS. 1986. Fertilizer trial in ten and twenty years old teak

plantations in Nadhya Pradesh. Journal of Tropical Forestry 2(1): 47–52.

Qiong Z, Xing-yu L, De-hui Z. 2011. Aboveground biomass and nutrient allocation in an age-

sequence of Larix olgensis plantations. Journal of Forest Resources 22 (1): 71-76.

R Development Core Team. 2011. R: A Language and Environment for Statistical Computing.

R Foundation for Statistical Computing. Vienna, Austria.

Ranger J, Marques R, Colin-Belgrand M, Flammang N, Gelhaye D. 1995. The dynamics of

biomass and nutrient accumulation in a Douglas-fir (Pseudotsuga menziesii Franco) stand

studied using a chronosequence approach. Forest Ecology Management 72: 167-183.

Raupach M. 1967. Soil and fertilizer requirements for forests of Pinus radiata. Advances in

Agronomy 19: 307–353.

Rennie PJ. 1955. The uptake of nutrients by mature forest growth. Plant and Soil 8 (1): 49-95.

Richards BN, Bevege DI. 1972. Principles and practice of foliar analysis as a basis for crop-

logging in pine plantations: 1. Basic considerations. Plant and Soil 36: 109-119.

Robert PC. 2002. Precision agriculture: a challenge for crop nutrition management. Plant and

Soil 247: 143–149.

Ryan MG, Binkley D, Fownes JH. 1997. Age-related decline in forest productivity: pattern and

process. Advances in Ecological Research 27: 213-262.

206

Sadler EJ, Karlen DL. 1995. Aerial dry matter and nutrient accumulation comparisons among

five soybean experiments. Communications on Soil Science and Plant Analysis 26 (19&20):

3145-3163.

Saint-Andre L, Laclau JP, Deleporte P, Ranger J, Gouma R, Saya A, Joffre R. 2002 A Generic

Model to Describe the Dynamics of Nutrient Concentrations within Stemwood across an

Age Series of a Eucalyptus Hybrid. Annals of Botany 90: 65-76.

Sánchez PA, Logan TJ. 1992. Myths and Science about the chemistry and fertility of soils in the

tropics. En: Myths and science of soils in the tropics. SSSA Special publication no. 29. Soil

Science Society of America and American Society of Agronomy, Madison, Wisconsin,

USA, pp. 35-46.

SAS Institute Inc. 2002. SAS/STAT®9.0 User’s Guide. Cary NC: SAS Institute Inc.

Saur E, Nambiar EKS, Fife DN. 2000 Foliar nutrient retranslocation in Eucalyptus globulus.

Tree Physiology 20: 1105-1112.

Schaetzl R, Anderson S. 2005. Soils: Genesis and Geomorphology. Cambridge University

Press, New York.

Segura JM, Alvarado A, Vaides EE. 2013. Clasificación de tierras aptas para el establecimiento

de plantaciones forestales de teca (Tectona grandis L f.) en Petén, Guatemala. VIII Congreso

Forestal Centroamericano. San Pedro Sula, Honduras.

Serrada R. 2008. Apuntes de Selvicultura. FUCOVASA, Madrid, España.

Siddiqui MT, Ali S, Adnan M. 2007. Root nutrient concentrations in teak (Tectona grandis L.f.)

plantations as influenced by fertilization and age. Canadian Journal of Pure and Applied

Sciences 1 (1): 45-52.

Siddiqui MT, Shah AH, Yaqoob S. 2009. Chronosequence and crown strata effects on foliar

nutrient concentrations in teak plantations. Pakistan Journal of Botany 41 (6): 3023-3034.

Singh A. 2004. Effect of fertilization on N and P resorption efficiency of selected leguminous

and nonleguminous tropical trees planted on coal mine spoil. Journal of the Indian Institute

of Science 84: 173–182.

207

Smethurst PJ. 2010. Forest fertilization: Trends in knowledge and practice compared to

agriculture. Plant and Soil 335: 83–100.

Smethurst PJ, Nambiar EKS. 1989. Role of weeds in the management of nitrogen in a young

Pinus radiata plantation. New Forests 3 (3): 203-224.

Sociedad Española de Ciencias Forestales. 2005. Diccionario forestall. Ediciones Mundi-Prensa,

Madrid, España.

Soil Survey Staff. 2010. Keys to Soil Taxonomy. 11th Ed. United Stated Department of

Agriculture (USDA).

Srivastava SK, Singh KP, Upadhyay RS. 1986. Fine root growth dynamics in teak (Tectona

grandis Linn F.). Canadian Journal of Forest Resesearch 16: 1360-1364.

Theocharopoulos SP, Petrakis PV, Trikatsoula A. 1997. Multivariate analysis of soil grid data as

a soil classification and mapping tool: the case study of a homogeneous plain in Vagia,

Viotia, Greece. Geoderma 77: 63–79.

Thiele H. 2008. Variables edáficas que afectan el crecimiento de la teca (Tectona grandis L.f.)

en la vertiente del Pacífico de Costa Rica. Tesis M Sc. Universidad de Costa Rica, San José,

Costa Rica.

Torres DA, del Valle JI, Restrepo G. 2012. Site index for teak in Colombia. Journal of Forestry

Research 23(3): 405–411.

Torres S, Márquez O, Hernández R, Franco W. 1993. Respuesta inicial de crecimiento a la

fosforita en teca en los Llanos Occidentales de Venezuela. Turrialba 43(2): 113-118.

Totey N. 1992. Fertilizer management in forest plantations. En: Non-traditional sectors for

fertilizer use (Tandon HLS ed). Fertilizer Development and Consultation Organization, New

Delhi, India, pp. 107-128.

Turner J, Lambert MJ. 2008. Nutrient cycling in age sequences of two Eucalyptus plantation

species. Forest Ecology and Management 255: 1701-1712.

van den Driessche R. 1984. Nutrient storage, retranslocation and relationship of stress to

nutrition. En: Nutrition of plantation forests (Bowen GD, Nambiar EKS eds). Academic

Press, London, pp, 181 - 209.

208

van Dijk AIJM, Keenan RJ. 2007. Planted forests and water in perspective. Forest Ecology and

Management 251: 1-9.

Vásquez W, Ugalde LA. 1994. Rendimiento y calidad de sitio para Gmelina arborea, Tectona

grandis, Bombacopsis quinata y Pinus caribaea en Guanacaste, Costa Rica. Proyecto

Madeleña CATIE/IDA/FAO/HOLANDA. Turrialba, Costa Rica.

Vitousek PM. 1984. Litterfall, nutrient cycling, and nutrient limitation in tropical forests.

Ecology 65: 285-298.

Wehr JB, Smith T, Blamey P, Menzies N. 2010. Aluminum sensitivity and optimum Ca and pH

requirement of teak (Tectona grandis) clones used for forestry plantations in Australia. XIX

Congress of Soil Science, Brisbane, Autralia, pp. 18 – 21.

West PW. 2006. Growing Plantation Forests. Springer, Germany.

Whitehead D. 2011. Forests as carbon sinks—benefits and consequences. Tree Physiology 31:

893–902.

Woods PV, Nambiar EKS, Smethurst PJ. 1992. Effect of annual weeds on water and nitrogen

availability to Pinus radiata trees in a young plantation. Forest Ecology and Management

48: 145-163.

Worrel R, Hampson A. 1997. The influence of some forest operations on the sustainable

management of forest soils – a review. Forestry 70 (1): 61-85.

Yamada M, Toma T, Hiratsuka M, Morikawa Y. 2004. Biomass and potential nutrient removal

by harvesting in short-rotation plantations. En: Site management and productivity in tropical

plantation forests: Proceedings of workshops in Congo July 2001 and China February 2003

(Nambiar EKS, Ranger J, Tiarks A, Toma T eds). CIFOR, Bogor, Indonesia.

Yan L, Zhou S, Feng L, Hong-Yi L. 2007. Delineation of site-specific management zones using

fuzzy clustering analysis in a coastal saline land. Computers and Electronics in Agriculture

56: 174–186.

Yay OD, Alagha O, Tuncel G. 2008. Multivariate statistics to investigate metal contamination

in surface soil. Journal of Environmental Management 86: 581–594.

209

Yuan Z, Liu W, Niu S, Wan S. 2007. Plant Nitrogen dynamics and Nitrogen-use strategies

under altered Nitrogen seasonality and competition. Annals of Botany 100: 821-830.

Yuan ZY, Chen HYH. 2010. Changes in nitrogen resorption of trembling aspen (Populus

tremuloides) with stand development. Plant and Soil 327: 121-129.

Zech W, Dreschel P. 1991. Relationship between growth, mineral nutrition and site factors of

teak (Tectona grandis) plantations in the rainforest zone of Liberia. Forest Ecology and

Management 41 (3-4): 221-235.

Zhou Z, Liang K, Xu D, Zhang Y, Huang G, Ma H. 2012. Effects of calcium, boron and

nitrogen fertilization on the growth of teak (Tectona grandis) seedlings and chemical

property of acidic soil substrate. New Forests 43: 231-243.

210

ANEXO I

Fernández-Moya J, Murillo R, Portuguez E, Fallas JL, Ríos V, Kottman F, Verjans JM,

Mata R, Alvarado A. 2013 Nutrient concentration age dynamics of teak (Tectona

grandis L.f.) plantations in Central America. Forest Systems 22 (1): 123-133.

Introduction

Teak (Tectona grandis L.f.) has been planted ex-tensively in Central America, acquiring socio-econo-mical relevance due to its productivity (Pandey andBrown, 2000; FAO, 2002). Reference foliar nutrient con-centrations have been summarized for teak (Drechseland Zech, 1991; Boardman et al., 1997), and a prelimi-nary Diagnosis and Recommendation IntegratedSystem (DRIS) has been developed for West Africanplanted teak forests (Drechsel and Zech, 1994). How-ever, appropriate knowledge regarding teak nutrition

is still required for a better management of the planta-tions to attain high productivity and sustainability.

The concentrations of nutrients in tissues dependmainly on species, environmental factors (climate andsoil availability) and plantation management. Whethercomparing different species or within a single species,genetic requirements, root distribution and age (deve-lopmental stage) are usually the most important factorsaffecting nutrient absorption. At early growth stages,tree nutrition is considered crucial to sustain high growthrates and rapid expansion of the crown and roots. Afterthe crown is fully developed, if seedling nutrition hasbeen adequate, tree requirements during the remainderof the rotation are assumed to be satisfied by environ-mental inputs, nutrient recycling and nutrient translo-

Nutrient concentration age dynamics of teak (Tectona grandis L.f.)plantations in Central America

J. Fernandez-Moya1,2*, R. Murillo2,3, E. Portuguez2, J. L. Fallas2,4, V. Rios2,5, F. Kottman2,5, J. M. Verjans2,6, R. Mata2 and A. Alvarado2

1 Departamento de Silvopascicultura. ETSI Montes. Technical University of Madrid (UPM). Ciudad Universitaria, s/n. 28040 Madrid, Spain

2 Centro de Investigaciones Agronómicas. Universidad de Costa Roca (CIA-UCR). Costa Rica3 Instituto de Investigación y Servicios Forestales. Universidad Nacional (INISEFOR-UNA). Costa Rica

4 GFA Consulting Group. Costa Rica5 Panamerican Woods, S.A. Costa Rica

6 Panaforest. Panamá

Abstract

Aim of study: Appropriate knowledge regarding teak (Tectona grandis L.f.) nutrition is required for a better mana-gement of the plantations to attain high productivity and sustainability. This study aims to answer the following ques-tions: How can it be determined if a teak tree suffers a nutrient deficiency before it shows symptoms? Are nutrientconcentration decreases in older trees associated with age-related declines in forest productivity?

Area of study: Costa Rica and Panama.Material and methods: Nutrient concentration in different tree tissues (bole, bark, branches and foliage) were me-

asured at different ages using false-time-series in 28 teak plantations.Research highlights: Foliar N concentration decreases from 2.28 in year 1 to 1.76% in year 19. Foliar Mg concen-

tration increases from 0.23 in year 1 to 0.34% in year 19. The foliar concentrations of the other nutrients are assumedto be constant with tree age: 1.33% Ca, 0.88% K, 0.16% P, 0.12% S, 130 mg kg–1 Fe, 43 mg kg–1 Mn, 11 mg kg–1 Cu,32 mg kg–1 Zn and 20 mg kg–1 B. The nutrient concentration values showed can be taken as a reference to evaluate thenutritional status of similar teak plantations in the region. The concentrations of K, Mg and N could be associated withdeclines in teak plantation productivity as the plantation becomes older. Whether age-related changes in nutrient con-centrations are a cause or a consequence of age-related declines in productivity is an issue for future research with theaim of achieving higher growth rates throughout the rotation period.

Key words: age-related decline in productivity; forest nutrition; nutrient bole concentration; nutrient foliar con-centration; resorption.

* Corresponding author: jesusfmoya@gmail.comReceived: 31-07-12. Accepted: 04-02-13.

Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria (INIA) Forest Systems 2013 22(1), 123-133 Available online at www.inia.es/forestsystems ISSN: 2171-5068http://dx.doi.org/10.5424/fs/2013221-03386 eISSN: 2171-9845

cation (Miller, 1981, 1984, 1995). Nutrients in foliageare considered to represent 20-40% of the total amountof nutrients in a stand, and lower nutrient concentra-tions found in the tree bole (e.g. Miller, 1995). Kumaret al. (2009) found teak reproductive parts to containthe highest nutrient concentrations, while twigs andfoliage also showed high values; the Ca and B concen-trations in bark are higher than in other tissues.

Foliar concentration is considered a useful para-meter to evaluate the nutritional status of a stand and asa reference to evaluate plantation fertilizer recommen-dations because (a) its variation is highly dependenton site and soil parameters; (b) it reflects the currentnutrient supply; (c) it allows diagnosis of nutritionaldeficiencies when they are not severe enough to causevisually observable symptoms and, thus, allows actionto be taken before the effects on productivity are signi-ficant; and (d) deficiency symptoms are confused easilywith other effects when visual guidelines are used(Mead, 1984; Drechsel and Zech, 1991; Barker andPilbeam, 2006; Lehto et al., 2010).

The nutrient concentrations presented in this paperfor dominant and co-dominant teak trees, which weredeveloped for a variety of soils and environmental con-ditions in Central America (Costa Rica and Panama), shouldbe taken as reference values for evaluating the nutri-tional status of similar teak plantations in the region.

Material and methods

Study sites

Three teak (Tectona grandis L.f.) plantations werestudied in Central America: two in Costa Rica (Guana-

caste and northern region) and one in Panama (PanamaCanal watershed) (Fig. 1). The three areas are classi-fied as tropical wet forest according to Holdridge’s lifezones, with similar mean annual rainfall (2,500-3,100mm), although in Guanacaste the dry season is longerthan at the other two sites. The soils of the study areasare also similar, although the northern region of CostaRica is less fertile than the other sites and it has highersoil acidity (Table 1).

The studied stands were chosen to be representativeof properly managed teak plantations in Central Ame-rica. In general, management of these plantations con-sists on continuous silvicultural activities: weed control,pruning, thinning regimen (approximately from 800-1,000 trees ha–1 at establishment to 150-200 trees ha–1

at final felling) and fertilization during the establish-ment. The use of clones is common in recent years, sothey were not sampled in the study. An expectedcommercial volume of 100-150 m3 is expected for thiskind of plantations in approximately 20 years rotation.

Field sampling and design

False time series (chronosequences) method wasused to analyze nutrient concentration dynamics ofteak trees from 1 to 19 years of age. Despite of the cri-tiques of this method (Johnson and Miyanishi, 2008),it was considered to be valid as all the studied standsare considered to be similar in environmental condi-tions (soil and climate) and management practices.

A total of 28 stands were analyzed, seven in Panama,12 in the northern region of Costa Rica and nine inGuanacaste (Costa Rica). In order to analyze a maxi-mum yield research experiment (Bertsch, 1998; Alva-

124 J. Fernández-Moya et al. / Forest Systems (2013) 22(1), 123-133

0 100 20050Km

80°

80°

10° 10°

Caribbean sea

Pacific ocean

COSTA RICA

PANAMA

Costa Rican Northern region study area

Guanacastestudy area

Panama watershedstudy area

Nicaragua

Panama

Costa Rica

Honduras

Pacific ocean

Caribbean sea

Guatemala

Central America

Mexico N

S

EW

Figure 1. Locations of the study teak (Tectona grandis L.f.) plantations: Guanacaste (Costa Rica); northern region (Costa Rica)and Panama Canal Watershed (Panama).

rado, 2012), dominant and co-dominant trees were se-lected: (a) without visible symptoms of diseases ornutritional deficiencies and (b) that were representativeof the best-performing trees of the plantations, assu-ming optimal nutrition and a full expression of geneticpotential. By analyzing nutrient concentration in themost productive soils without soils deficiencies (andin dominant or co-dominant trees in a site), the maxi-mum species requirements are assessed; hence, if theconsidered minimum inputs for these high-fertilitysites are applied in sites of lower fertility where tree nu-trient uptake would be lower, the productivity of the plan-tation is still achieved (Bertsch, 1998; Alvarado, 2012).

In stands younger than ten years of age, two treeswere sampled per stand, whereas only one tree wassampled in older stands. Trees were felled and treecomponents (bole, bole’s bark, foliage and primary andsecondary branches) were analyzed for nutrient con-centration. Tissue samples (1 kg per tissue per tree)were collected and analyzed at the “Centro de Investi-gaciones Agronómicas” of the University of Costa Rica(hereafter CIA-UCR) to determine nutrient concen-trations (N, P, Ca, Mg, K, S, Fe, Mn, Cu, Zn and B,hereafter referred to as nutrients) after samples weredried and water content was assessed. Dry combustionwas used to measure the N concentration, and wetdigestion and atomic spectrometry were used to extract

and determine other nutrients (Bertsch, 1998). Primaryand secondary branches were weighted averaged andare reported as “branches”. Foliage samples werecollected as representative homogeneous mixture ofall the foliage of the sampled trees. All the field workwas performed during July-September, at the treesoptimal nutritional status during the period of maxi-mum growth activity, to avoid effects of seasonality.In order to estimate soil nutrient availability, topsoilsamples were collected (0-20 cm). Soil informationwas only available for 23 of the 28 sampled stands(Table 1). Soil samples were analyzed at CIA-UCR todetermine: pH (in water), Ca, Mg, K, P, Fe, Cu, Zn,Mn, exchangeable acidity and Al, following the KCl-modified Olsen method, as described by Díaz-Romeuand Hunter (1978). Organic matter was determined bythe wet combustion method of Walkey and Black, asdescribed by Briceño and Pacheco (1984). Soil texturewas determined using the modified Bouyoucos method,as described by Forsythe (1975).

Statistical analysis

Generalized linear mixed models (GLMMs) wereused to study the relationships between the concen-tration of nutrients (N, P, Ca, Mg, K, S, Fe, Mn, Cu,

Nutrient concentration age dynamics of teak plantations in Central America 125

Table 1. Summary of soil properties at the different study areas; means and coefficients of variation (in parentheses) areprovided. Soil information was only available for 23 of the 28 sampled stands

Northern region, Guanacaste, Canal ZoneTotal

Costa Rica Costa Rica Panama(n = 23)

(n = 11) (n = 9) (n = 3)

pH 5.11 (6) 5.90 (6) 6.70 (12) 5.63 (12)Acidity [cmol(+)–1] 0.70* (5) 0.31 (30) 0.15 (33) 0.48 (81)Ca [cmol(+) L–1] 4.45 (44) 21.36 (28) 20.97 (38) 13.22 (74)Mg [cmol(+) L–1] 1.46 (47) 6.89 (54) 5.25 (64) 4.08 (89)K [cmol(+) L–1] 0.13* (109) 0.33 (87) 0.36 (82) 0.24 (101)ECEC [cmol(+) L–1] 6.74 (31) 28.90 (32) 26.72 (40) 18.02 (71)AS (%) 11.96* (84) 1.22 (58) 0.65 (55) 6.28* (139)P (mg L–1) 3* (114) 3* (146) 2* (0) 3* (124)Zn (mg L–1) 2* (84) 3 (58) 3 (107) 2* (77)Cu (mg L–1) 8 (19) 11 (85) 4 (83) 9 (71)Fe (mg L–1) 165 (23) 37 (81) 65 (154) 102 (75)Mn (mg L–1) 43 (171) 38 (82) 19 (101) 38 (142)Organic matter (%) 4.6 (27) 3.8 (28) 4.6 (16) 4.3 (27)Sand (%) 24.9 (23) 23.4 (56) 29.0 (31) 24.8 (38)Silt (%) 18.4 (13) 36.9 (42) 36.8 (43) 28.0 (51)Clay (%) 56.7 (11) 39.7 (22) 34.3 (50) 47.1 (27)

ECEC: effective cations exchange capacity. AS: acidity saturation. * Values outside the adequate reference soil levels (Bertsch,1998).

Zn and B) in each tissue (bole, bark, bole and bark, bran-ches, foliage and total) and tree age. The use of GLMMswas required, as most of the study variables did notapproach the normal distribution hypothesized in tra-ditional models. Appendix 1 summarizes the completelist of the 44 response variables analyzed and the distribu-tion approached by the data used to construct the GLMMs.The exponentially distributed variables were modeledusing a Gamma distribution approach with α = 1.

To evaluate the best fitted model for each study va-riable, a total of 83 different models were constructed,selecting the one with lowest deviance. Appendix 2summarizes a complete list of the models constructedfor each study variable. Three groups of models wereconstructed: (1) a null model considering only anintercept [yi = b0]; (2) a model considering an interceptin addition to age as an explanatory variable [(yi)λ =b1 · age + b0]; and (3) a model without an intercept[(yi)λ = b1 · age ]. For groups (2) and (3), 41 differentpower link functions [g(µ) = µλ] were tested for eachone, with λ varying between λ = 2 to λ = –2 and aλgap = 0.1. When no model including age as a parameterwas statistically significant, or when the data did notfollow any of the studied distribution functions (Appen-dix 1), the resulting model included only an intercept

representing the mean of the variable, and no age effectwas taken into account.

The sampled stands in each study area were spatiallycorrelated. The spatial correlation was taken intoaccount by including a random effect for the study area,modeling the working correlation matrix with a first-order autoregressive structure. The goodness-of-fit ofthe models was estimated by measuring the percentagedifference between the deviance of the model and thedeviance of a model with no covariates (hereafter re-ferred to as efficiency, EF), which is a pseudo-R2 mea-sure reported for GLMMs. All statistical analyses wereperformed using SAS 9.0 (SAS Institute Inc., 2002).All statistical tests throughout the text are consideredsignificant with α = 0.05.

Results

N concentrations decreased with age in all tissues(Fig. 2, Table 2); the N concentration was higher in thefoliage (1.7-2.3%) than in the other tissues. Ca concen-trations did not show any relationship with age in anytissue (Fig. 3, Table 2) but were highest in the bark(1.9%), followed by the foliage (1.3%) and finally the

126 J. Fernández-Moya et al. / Forest Systems (2013) 22(1), 123-133

Figure 2. Tissue N concentration (%) related to tree age (years). Points represent trees sampled at three different locations: Gua-nacaste, Costa Rica (●); northern region, Costa Rica (�) and Panama (�). Lines represent fitted models (Table 2).

N(%

) bol

e

N(%

) bar

kN

(%) F

ollia

ge

N(%

) bra

nche

s

0 5 10 15 20 0 5 10 15 20

0 5 10 15 20 0 5 10 15 20

0.60

0.40

0.20

0.0

0.80

0.60

0.40

0.20

0.00

3.00

2.00

1.00

0.00

1.00

0.75

0.50

0.25

0.00

bole (0.1%). K concentrations decreased with age inthe bole, bark and branches but were constant in thefoliage (Fig. 4, Table 2). K-bark (0.6-1.6%) was higherthan in the other tissues during the early growth years,but after 14 years, K-foliage was highest (0.88%). Mg-bole and Mg-branches decreased with age, whereasMg-foliage increased and Mg-bark was constant (Fig. 5,Table 2). The Mg-foliage (0.23-0.34%) was highest,although Mg-bark was also high (0.23%) comparedwith Mg-bole and Mg-branches. P concentrations didnot show any relationship with age in any tissue (Table 2).P-foliage (0.16%) was higher than in the other tissues(< 0.08%). S-bark decreased with age but remainedconstant in the other tissues (Table 2). In most tissues,the macronutrient concentrations generally followedthe trend N > Ca = K >> Mg > P > S.

Fe-bark decreased with age but remained constantin the other tissues (Table 2); in young trees, Fe-bark(28-304 mg kg–1) was higher than in other tissues,

although after 14 years, Fe-branches (163 mg kg–1) washighest. Mn-bark decreased with age but remainedconstant in the other tissues (Table 2). Cu-branchesdecreased with age but remained constant in the othertissues (Table 2). Zn-branches decreased with age butremained constant in the other tissues (Table 2). Zn-foliage (32 mg kg–1) was higher than the Zn con-centration in the other tissues, although the Zn-bark(30 mg kg–1) was also high as was Zn-branches inyounger trees (9-28 mg kg–1). B concentrations did notshow any relationship with age in any tissue (Table 2);B-bark (31 mg kg–1) was highest, although B-foliage(20 mg kg–1) was also high compared to B-branches(11 mg kg–1) and B-bole (3 mg kg–1). In general, themicronutrient concentrations in the tissues followedthe trend Fe >> Mn > Zn = B > Cu.

Ca was found to be the most concentrated nutrientin the tree bark; its concentration in the bark remainedconstant with age, which was also found for Mg, P, S,

Nutrient concentration age dynamics of teak plantations in Central America 127

Table 2. Regressions between tissues’ nutrient concentration and tree age (years) in 1 to 19 years old teak (Tectona grandisL.f.) plantations in Costa Rica and Panama. Below specified models are in the form [y = (b0 + b1 · age)1/λ], where the responsevariables (y) are the nutrients concentration at tree tissues. When no model including age as a parameter was statisticallysignificant, the model only included an intercept (b0) representing the mean of the variable

TissuesMacronutrient

b0b0 b1

b1 λλ EF Micronutrientb0

b0 b1b1 λλ EF

(%) (std. error) (std. error) (%) (mg kg–1) (std. error) (std. error) (%)

Foliage N 0.1845 0.0170 0.0073 0.0004 –2 34 Fe 129.6087 22.9492Ca 1.3363 0.1062 Mn 42.5507 1.7941K 0.8759 0.0736 Cu 11.0761 0.4361Mg 0.0501 0.0151 0.0033 0.0003 2 90 Zn 31.9978 3.7041P 0.1589 0.0198 B 19.6158 0.6613S 0.1181 0.0047

Bark N 1.4506 0.1306 0.1258 0.0246 –2 36 Fe 97,791.3100 1,9837.4200 –5,106.0 1,043.8660 2 29Ca 1.9098 0.2409 Mn 0.0004 0.0002 0.0002 0.0001 –2 43K 1.2218 0.0234 –0.0205 0.0042 0.4 30 Cu 3.4615 0.3050Mg 0.2349 0.0072 Zn 29.8471 4.5945P 0.0772 0.0112 B 30.6635 1.1832S 32.0732 1.6898 2.2538 0.3823 –1.4 23

Bole N 6.1237 0.9564 0.6100 0.1253 –2 26 Fe 72.4639 23.0074Ca 0.1093 0.0057 Mn 1.2500 0.3113K 0.6179 0.0197 –0.0229 0.0029 0.5 72 Cu 2.0907 0.2691Mg 18.9991 3.2787 7.2370 1.5985 –1.7 30 Zn 10.3356 4.4177P 0.0645 0.0162 B 2.7143 0.2763S 0.0446 0.0098

Branches N 2.4636 0.4475 0.2449 0.0420 –2 36 Fe 162.7561 8.7137Ca 0.9122 0.0544 Mn 13.9337 1.2417K 0.4282 0.0458 Cu –0.0060 0.0026 0.0085 0.0010 –2 48Mg 14.5100 3.9542 5.9444 0.3472 –2 39 Zn 3.3847 0.1286 –0.0648 0.0088 0 28P 0.0751 0.0221 B 11.1654 0.1765S 0.0673 0.0064

EF (%): model efficiency, pseudo R2 estimate for Generalized Linear Mixed Models.

128 J. Fernández-Moya et al. / Forest Systems (2013) 22(1), 123-133

Figure 3. Tissue Ca concentration (%) related to tree age (years). Points represent trees sampled at three different locations: Gua-nacaste, Costa Rica (●); northern region, Costa Rica (�) and Panama (�). Lines represent fitted models (Table 2).

0.20

0.15

0.10

0.05

0.00

1.50

1.00

0.50

0.00

2.50

2.00

1.50

1.00

0.50

0.00

4.00

3.00

2.00

1.00

0.000 5 10 15 20 0 5 10 15 20

0 5 10 15 20 0 5 10 15 20

Ca(%

) bol

eCa

(%) b

ranc

hes

Ca(%

) fol

iage

Ca(%

) bar

k

Figure 4.Tissue K concentration (%) related to tree age (years). Points represent trees sampled at three different locations: Gua-nacaste, Costa Rica (●); northern region, Costa Rica (�) and Panama (�). Lines represent fitted models (Table 2).

0.50

0.40

0.30

0.20

0.10

0.0

2.00

1.50

1.00

0.50

0.00

1.50

1.00

0.50

0.00

2.50

2.00

1.50

1.00

0.50

0.000 5 10 15 20 0 5 10 15 20

0 5 10 15 20 0 5 10 15 20

K(%

) bol

e

K(%

) bar

kK

(%) f

olia

ge

K(%

) bra

nche

s

Cu, Zn and B. In contrast, the N, K, Fe and Mn con-centrations in the bark declined with tree age. N wasthe most concentrated nutrient in the tree bole and haddecreasing concentrations with age. K-bole was alsohigh in young trees but decreased sharply with age,with very low values at the end of the rotation. Mg-bole declined slowly with age due to the low valuesfound in young trees. The relatively low concentrationsof Ca, P, S, Fe, Mn, Cu, Zn and B in the bole showedno relationship with tree age. K was the most concen-trated nutrient in young tree branches but decreasedsharply with tree age. Ca-branches was highest in treesolder than approximately 3 years. The concentrationsof other elements in the tree branches, including N,Mg, Cu and Zn, also declined with age, whereas thoseof Ca, P, S, Fe, Mn and B did not show any relationshipwith tree age. N was the most concentrated nutrient inthe foliage and decreased with age, while the otherelements did not show any relationship with age, withthe exception of Mg, which increased with tree age.

Discussion

In West African planted teak forests, Drechsel andZech (1994) observed decreases in foliar N and P con-

centrations with tree age; Montero (1999) also reportedfoliar N, P and K concentrations decrease in CostaRica. Similarly, Siddiqui et al. (2009) found foliar N,P, K and Zn concentration decrease with age, whereasCa and Mn increase. Siddiqui et al. (2007) also re-ported a significant reduction in nutrient concentra-tions in teak roots with trees age. These previously re-ported patterns do not correspond to the dynamicsfound in the present study, in which the only foliarnutrient concentration found to decline with tree agewas N (Fig. 2, Table 2), whereas there was a tendencyfor the Mg concentration to increase with tree age (Fig. 5,Table 2).

N concentrations decreased with age in all tissues(Fig. 2), especially in the foliage (2.3 to 1.7%), as hasbeen previously reported for teak (Montero, 1999) andis considered to be a general trend in plant nutrition(Barker and Pilbeam, 2006; Yuan et al., 2007). HighN concentrations in young trees could be due to thelarge requirements of plants at this fast-growing stage,as N is usually related to plant growth (Fölster andKhanna, 1997; Barker and Pilbeam, 2006). However,as greater soil N availability leads to higher plant Nconcentrations, the high N concentration at the be-ginning of the rotation could also be explained by thelarge amount of N available from the soil at this stage,

Nutrient concentration age dynamics of teak plantations in Central America 129

Figure 5. Tissue Mg concentration (%) related to tree age (years). Points represent trees sampled at three different locations: Gua-nacaste, Costa Rica (●); northern region, Costa Rica (�) and Panama (�). Lines represent fitted models (Table 2).

0.20

0.15

0.10

0.05

0.00

0.30

0.20

0.10

0.00

0.50

0.40

0.30

0.20

0.10

0.00

0.50

0.40

0.30

0.20

0.10

0.000 5 10 15 20 0 5 10 15 20

0 5 10 15 20 0 5 10 15 20

Mg

(%) b

ole

Mg

(%) b

ark

Mg

(%) f

olia

ge

Mg

(%) b

ranc

hes

which could be supplied either by large amounts oforganic residues combined with high mineralizationrates and/or by fertilization. Thus, a plant would absorblarge amounts of N, store it as a reservoir and use itlater during the rotation by translocation from onetissue to another (Miller, 1984; Yuan et al., 2007; Yuanand Chan, 2010). Hence, the application of N fertilizerat this stage could be futile or could cause even greaterlosses due to leaching, resulting in contamination andeconomic losses (Fölster and Khanna, 1997).

The decreases in N concentration observed in seve-ral tissues with tree age (Fig. 2) could be explained by(a) decreasing plant requirements as plants age anddecline in productivity and require less N to supportthese lower growth rates; (b) a growth dilution effect,as plant biomass increases with age and usually tendsto allocate more structural and storage materials con-taining little N (Yuan et al., 2007); or (c) a decline inthe soil N supply, which would result in lower N uptakeand greater nutrient translocation using the nutrientsstored during younger years. However, declines in Nconcentration with age are also considered one of thecauses of age-related declines in forest productivitybecause (i) N is usually considered to be the limitingnutrient in forest ecosystems, particularly in young tro-pical soils (Hedin et al., 2009); (ii) the N minerali-zation rate in older forest soils is lower than in youngerstands, causing a diminishing soil N supply in olderstands; and (iii) plant N supply and forest nutrition aregenerally related to the photosynthesis rate, so a de-crease in plant N supply would cause lower net primaryproduction (Gower et al., 1996; Ryan et al., 1997;Binkley et al., 2002). Understanding whether N con-centration declines are a cause or a consequence ofplanted forest productivity declines is an importantissue that should be addressed in further research. Oneway to resolve this issue could be to establish a ferti-lization experiment in mature planted teak forests(combined with a thinning program), monitoring theN concentration before and after fertilization and eva-luating whether an increase in the N concentrationwould eventually result in an increase in growth rates.

High Ca concentrations in teak bark have also beenreported by other authors (Nwoboshi, 1984; Totey,1992; Negi et al., 1995). The Ca concentration in teakbark tended to increase with tree age in the trees sam-pled in Guanacaste (Costa Rica) and Panama, but wecould not fit a sound statistical model reflecting thistendency, as some trees in the northern region of CostaRica presented low Ca bark concentrations (Fig. 3).

These low concentrations probably reflect Ca deficien-cies or certain disorders, as the soils in this regionexhibited lower values of available soil Ca and highacidity (Table 1). However, the foliar Ca concentrationof the trees in this region was adequate and comparableto that of trees from other regions (Fig. 3). If Ca defi-ciency occurs it would mainly affect new leaves (Barkerand Pilbeam, 2006), so trees with lower bark Ca con-centrations could have suffered a Ca deficiency in thepast but then recovered, thus exhibiting adequate nu-tritional status at sampling time. This phenomenoncould be explained by lime application at intermediatetree ages.

Zech and Drechsel (1991) consider values of 0.96-1.21 for the Ca:K ratio in foliage as adequate for healthyteak trees; the average ratio value found in our studywas 1.53, reflecting a nutrient imbalance involving aCa excess and/or a K deficiency. It could also reflecta difference between African planted teak forests andthe investigated Central American teak stands in termsof environmental, management or even genetic diffe-rences. The foliar K concentration (0.88 ± 0.07%) fellat the lower end of the range (0.80-2.32%) consideredas adequate (Drechsel and Zech, 1991; Boardman etal., 1997), higher than the values reported by Negi etal. (1990) in India (0.83%) and Benin (0.29%). K is amobile nutrient that plays key roles in photosynthesisand CO2 assimilation (Barker and Pilbeam, 2006) andexerts a regulatory effect on stomatal movement andtranspiration rates; thus, foliar K requirements wouldbe expected to increase with tree age because modi-fying transpiration rates to control increased hydraulicresistance and sustaining a higher photosynthesis rateare two of the key physiological mechanisms under-lying the plant response to the aforementioned age-related decline in plant productivity (Gower et al.,1996; Ryan et al., 1997; Binkley et al., 2002). How-ever, the foliar K concentration showed no relationshipwith tree age, in spite of the foliar K decreases asso-ciated with age in the sampled trees reported by Montero(1999). The decreasing K concentrations in the bole,bark and branches observed with increasing tree age(Fig. 4) are probably related to the constant foliar Kconcentration, as the increasing foliar K requirementsare probably supplied by the K in those tissues bynutrient translocation.

An increase in the foliar Mg concentration with treeage, as found in this study, has also been reported byMontero (1999) in teak plantations in Costa Rica. Thisincrease could be related to the decline in the Mg con-

130 J. Fernández-Moya et al. / Forest Systems (2013) 22(1), 123-133

centrations in the bole and branches because Mg isprobably translocating from these tissues to leaves. Fo-liar Mg may increase with age to meet the physiolo-gical demands of older trees; these physiological needsinclude the following: a) to sustain high photosynthesiseff iciency; b) to partially inhibit excess photophos-phorylation; and c) to regulate leaf stomatal conduc-tance (Gower et al., 1996; Gholz and Lima, 1997; Ryanet al., 1997; Binkley et al., 2002; Barker and Pilbeam,2006). However, some of the foliar Mg concentrationsfound in young trees (Fig. 5) are considered low rela-tive to the adequate reference range values (0.20-0.37%) proposed for teak in the literature (Drechseland Zech, 1991; Boardman et al., 1997); in fact, duringthe dry season in the northern region of Costa Rica,symptoms of foliar Mg deficiency are common. Hence,the translocation of Mg from the bole and branchesand the increase in the foliar Mg concentration can beconsidered as mechanisms to achieve an adequate Mglevel to ensure plant productivity.

The P concentration was higher in the foliage thanin the other tissues analyzed, as previously reported byNwoboshi (1984), showing no tendency associatedwith tree age. The average foliar P concentration liesat the lower limit of the reference range (0.14-0.25%)for adequate values reported in the literature for teak(Drechsel and Zech, 1991; Boardman et al., 1997),although many of the foliage samples showed valueslower than this reference. S was found to be concen-trated mainly in the teak foliage, showing values in thelower end of the reference range (0.11-0.23%) consi-dered as adequate for teak (Drechsel and Zech, 1991;Boardman et al., 1997).

The tissue Fe and Mn concentrations showed highvariability, probably due to sample contamination withsoil during f ieldwork; however, prior to statisticalanalysis, the values determined as too high were consi-dered as outliers and removed from the dataset. Thiscontamination was most noticeable in the bole, thebranches and especially the bark, where the Fe and Mnconcentration declines with tree age were probablycaused by a decrease in the proportion of contaminatedvs. properly collected samples as the biomass increa-sed. The foliar Fe (58-390 mg kg–1) and Cu (10-25 mgkg–1) concentrations were within the ranges consideredadequate for teak, while the foliar Mn (50-112 mg kg–1)values were below it (Drechsel and Zech, 1991; Board-man et al., 1997).

The foliar Zn concentration lies within the range(20-50 mg kg–1) considered as adequate by other authors

(Drechsel and Zech, 1991; Boardman et al., 1997),although Zn is usually deficient in plants growing inhighly weathered soils (Barker and Pilbeam, 2006),such as the ones in our study areas. The overall averageamong the locations showed low available soil Zn; thislow average was influenced by the low values found inthe northern region of Costa Rica, as the soil Zn valuesin Guanacaste and Panama are considered adequate(Table 1). Montero (1999) reported an increase in thefoliar Zn concentration with age, in contrast with thelack of a relationship found in this study, where thevariability of the data was high. The foliage B concen-tration lies within the range (15-45 mg kg–1) consideredadequate by other authors (Drechsel and Zech, 1991;Boardman et al., 1997), which is higher than the requi-rements reported for other species (Lehto et al., 2010).Of all tested tissues, the highest B concentration wasin the bark, as has been reported for other tree species(Lehto et al., 2010).

Generally, relatively high values of microelementsare required to maintain an appropriate nutritional statusin teak and to ensure forest productivity and sustaina-bility, although little attention has been paid to thisissue in other studies of teak nutrition (Nwoboshi,1984; Totey, 1992; Negi et al., 1995; Kumar et al., 2009).Tropical soils are usually characterized as highly wea-thered soils that are rich in Fe or Mn but generally defi-cient in Zn, B, Cu, and Mo (Barker and Pilbeam, 2006).B is commonly deficient in soils throughout the worldand is difficult to evaluate in routine soil fertility ana-lyses (Lehto et al., 2010). Hence, special care shouldbe taken to evaluate the B and Zn status in plantedforests throughout the tropics.

Tissue nutrient concentrations, especially those forfoliage, are considered to be a management tool forevaluating the nutritional status of planted trees (Mead,1984; Drechsel and Zech, 1991; Barker and Pilbeam,2006; Lehto et al., 2010). Foliar concentrations havebeen reported to be remarkably useful for this purposebecause they are sensitive indicators of nutritional defi-ciencies due to their direct relation with productivity,as foliage is where photosynthesis takes place (Mead,1984). Table 2 summarizes the models and values thatwe put forth for consideration as adequate concentra-tion reference levels to be used in nutrient managementof planted teak forests in Central America. By selectingdominant and co-dominant trees within well-managedand highly productive plantations, we sampled treeswith an appropriate nutritional status and highernutrient requirements than average, so if the planta-

Nutrient concentration age dynamics of teak plantations in Central America 131

tions are managed to ensure the aforementioned levels,and if fertilizers are added accordingly, the trees wouldhave good nutritional status. Hence, the nutrientconcentration values found in this study can be takenas a reference to evaluate the nutritional status of simi-lar teak plantations in the region, i.e., as teak nutritionguidelines for Central America.

Decreases in N concentration with tree age areconsidered to be either a cause or a consequence of thedecline in productivity associated with increasing treeage; the K and Mg concentrations could also be relatedto this phenomenon, which is a key issue in appliedecology. Future research about these relationshipsshould be performed with the aim of achieving highergrowth rates throughout the rotation period, whichwould allow shorter cycles to be used.

Acknowledgments

The authors would like to acknowledge the colla-boration of Ecoforest (Panama), S.A., Inversiones Agro-forestales Ltd., Panamerican Woods Ltd., and Expoma-deras Ltd. in the sampling and cost of analysis for thesoil and tissue samples. The authors also thank the per-sonnel of the Natural Resources Laboratory at CIA/ UCR for their help while sampling soils and tissues inthe f ield and Richard Anderson (Green MilleniumLtd.) and Adam Collins for their comments on themanuscript and especially the English language correc-tions. The present work was done within the frame ofthe MACOSACEN project, financed by PCI-AECID.

References

Alvarado A, 2012. Diagnóstico de la nutrición en planta-ciones forestales. In: Nutrición y fertilización forestal enregiones tropicales (Alvarado A, Raigosa J, eds). Asocia-ción Costarricense de las Ciencias del Suelo, San José,Costa Rica. pp: 93-120.

Barker AV, Pilbeam DJ, 2006. Handbook of Plant Nutrition.CRC Press, USA.

Bertsch F, 1998. La fertilidad de los suelos y su manejo. Aso-ciación Costarricense de la Ciencia del Suelo, San José,Costa Rica.

Binkley D, Stape JL, Ryan MG, Barnard HR, Fownes J,2002. Age-related decline in forest ecosystem growth: anindividual-tree, stand-structure hypothesis. Ecosystems5: 58-67.

Boardman R, Cromer RN, Lambert MJ, Webb MJ, 1997.Forest plantations. In: Plant analysis, an interpretation

manual (Reuter DJ, Robinson JB, eds). CSIRO Publishing,Australia. pp: 505-561.

Briceño J, Pacheco R, 1984. Métodos analíticos para el aná-lisis de suelos y plantas. UCR-CIA, San José, Costa Rica.

Díaz-Romeu R, Hunter A, 1978. Metodología de muestreode suelos, análisis químico de suelos y tejido vegetal e inves-tigación en invernadero. CATIE, Turrialba, Costa Rica.

Drechsel P, Zech W, 1991. Foliar nutrient levels of broad-leaved tropical trees: a tabular review. Plant and Soil131(1): 29-46.

Drechsel P, Zech W, 1994. DRIS evaluation of teak (Tectonagrandis L.f.) mineral nutrition and effects of nutrition andsite quality on the teak growth in West Africa. For EcolManag 70(1-3): 121-133.

FAO, 2002. Teak (Tectona grandis) in Central America. FAO,Rome.

Forsythe W, 1975. Física de suelos. Manual de laboratorio.IICA, Costa Rica.

Fölster H, Khanna PK, 1997. Dynamics of nutrient supplyin plantation soils. In: Management of soil, nutrients andwater in tropical plantation forests (Nambiar EKS, BrownAG, eds). Australian Centre for International AgriculturalResearch, Australia. pp: 339-379.

Gholz HL, Lima WP, 1997. The ecophysiological basis forproductivity in the tropics. In: Management of soil, nu-trients and water in tropical plantation forests (NambiarEKS, Brown AG, eds). Australian Centre for InternationalAgricultural Research, Australia. pp: 339-379.

Gower ST, McMurtrie RE, Murty D, 1996. Aboveground netprimary production decline with stand age: potential cau-ses. Tree 11(9): 378-382.

Hedin LO, Brookshire ENJ, Menge DNL, Barron AR, 2009.The nitrogen paradox in tropical forest ecosystems. AnnRev Ecol Evol and Systematics 40: 613-635.

Johnson EA, Miyanishi K, 2008. Testing the assumptions of chronosequences in succession. Ecol Letters 11(5):419-431.

Kumar JIN, Kumar RN, Kumar BR, Sajish PR, 2009.Quantification of nutrient content in the aboveground bio-mass of teak plantation in a tropical dry deciduous forestof Udaipur, India. J For Sci 55(6): 251-256.

Lehto T, Ruuhola T, Dell B, 2010. Boron in forest trees andforest ecosystems. For Ecol Manag 260: 2053-2069.

Mead DJ, 1984. Diagnosis of nutrient deficiencies in plan-tations. In: Nutrition of plantation forests (Bowen GD,Nambiar EKS, eds). Academic Press, Great Britain. pp:259-292.

Miller HG, 1981. Forest fertilization: some guiding con-cepts. For 54(2): 157-167.

Miller HG, 1984. Dynamics of nutrient cycling in plantationecosystems. In: Nutrition of plantation forests (BowenGD, Nambiar EKS, eds). Academic Press, Great Britain.pp: 53-79.

Miller HG, 1995. The influence of stand development onnutrient demand, growth and allocation. Plant and Soil168-169: 225-232.

Montero M, 1999. Factores de sitio que influyen en elcrecimiento de Tectona grandis L.f. y Bombacopsis qui-

132 J. Fernández-Moya et al. / Forest Systems (2013) 22(1), 123-133

nata (Jacq.) Dugand, en Costa Rica. MSc thesis. Uni-versidad Austral de Chile/CATIE.

Negi JDS, Bahuguna VK, Sharma DC, 1990. Biomassproduction and distribution of nutrients in 20 years oldteak (Tectona grandis) and gamar (Gmelina arborea)plantation in Tripura. Indian Forester 116: 682-685.

Negi MS, Tandon VN, Rawat HS, 1995. Biomasss and nu-trient distribution in young teak (Tectona grandis Linn.f.) plantation in Tarai Region of Uttar Pradesh. IndianForester 121(6): 455-464.

Nwoboshi LC, 1984. Growth and nutrient requirements in ateak plantation age series in Nigeria. II. Nutrientaccumulation and minimum annual requirements. For Sci30(1): 35-40.

Pandey D, Brown C, 2000. Teak: a global overview. Unasylva51(2): 1-15.

Ryan MG, Binkley D, Fownes JH, 1997. Age-related declinein forest productivity: pattern and process. Adv Ecol Res27: 213-262.

SAS Institute Inc, 2002. SAS/STAT® 9.0 User’s Guide. CaryNC: SAS Institute Inc.

Siddiqui MT, Ali S, Adnan M, 2007. Root nutrient concen-trations in teak (Tectona grandis L.f.) plantations asinfluenced by fertilization and age. Can J Pure Appl Sci1(1): 45-52.

Siddiqui MT, Shah AH, Yaqoob S, 2009. Chronosequenceand crown strata effects on foliar nutrient concentrationsin teak plantations. Pak J Bot 41(6): 3023-3034.

Totey N, 1992. Fertilizer management in forest plantations.In: Non-traditional sectors for fertilizer use (Tandon HLS,ed). Fertilizer Development and Consultation Organiza-tion, New Delhi, India. pp: 107-128.

Yuan Z, Liu W, Niu S, Wan S, 2007. Plant nitrogen dynamicsand nitrogen-use strategies under altered Nitrogen seaso-nality and competition. Ann Bot 100: 821-830.

Yuan ZY, Chen HYH, 2010. Changes in nitrogen resorptionof trembling aspen (Populus tremuloides) with stand deve-lopment. Plant Soil 327: 121-129.

Zech W, Dreschel P, 1991. Relationship between growth,mineral nutrition and site factors of teak (Tectona grandis)plantations in the rainforest zone of Liberia. For EcolManag 4(3-4): 221-235.

Nutrient concentration age dynamics of teak plantations in Central America 133

ANEXO II

Fernández-Moya J, Alvarado A, Morales M, San Miguel-Ayanz A, Marchamalo-

Sacristán M. 2014. Using multivariate analysis of soil fertility as a tool for forest

fertilization planning. Nutrient Cycling in Agroecosystems 98 (2): 155-167.

ORIGINAL ARTICLE

Using multivariate analysis of soil fertility as a tool for forestfertilization planning

Jesus Fernandez-Moya • Alfredo Alvarado •

Manuel Morales • Alfonso San Miguel-Ayanz •

Miguel Marchamalo-Sacristan

Received: 10 October 2013 /Accepted: 22 January 2014 / Published online: 29 January 2014

� Springer Science+Business Media Dordrecht 2014

Abstract The design of fertilization plans to cover

large areas is complex, due to the considerable number

of soil samples and soil fertility variables that must be

taken into account. Classifying forest stands in groups

according to their soil fertility (i.e. in nutrient man-

agement areas) can be very helpful to this respect and

it is considered to be a first step in what has been called

precision forestry. For this paper, we explore the

capability of multivariate analyses of topsoil data to be

used as tools for evaluating and classifying soil

fertility. A case study from a teak (Tectona grandis

L.f.) plantation in Costa Rica was used to evaluate and

illustrate how to use multivariate analysis with these

aims. A topsoil (0–20 cm) database with soil test

results assembled by Panamerican Woods Ltd. was

used. Different multivariate techniques [Principal

Component Analysis, Non-metric Multidimensional

Scaling (NMDS), Cluster analysis] were performed

and compared. Cluster analysis resulted as an appro-

priate tool for grouping soil samples into soil fertility

classes. Therefore, it is considered as a promising tool

which would help to design a fertilization program to

meet the specific needs of each group of stands with

relatively homogeneous soil fertility properties.

NMDS is also a suitable complementary tool to

graphically explore the similarities within groups and

the differences between them. The application of

procedures similar to those being reported may help to

optimize the design of nutritional and fertilization

plans across large forest plantations, by using multi-

variate analysis to establish fertilization regimes that

are appropriate to groups of stands of more homoge-

neous soil fertility.

Keywords Forest nutrition � Planted forests �Soil fertility � Tectona grandis � Site-specificmanagement � Nutrition management areas

Introduction

Forest plantation areal extent has globally increased

during recent decades, and now it covers

264 9 106 ha, 7 % of global forest area, in response

to the growing global demand for timber, pulp, energy

J. Fernandez-Moya (&) � A. San Miguel-Ayanz

Dpto. Silvopascicultura, E.T.S.I. Montes, Universidad

Politecnica de Madrid (UPM), Ciudad Universitaria s/n,

28040 Madrid, Spain

e-mail: jesusfmoya@gmail.com

J. Fernandez-Moya � A. AlvaradoCentro de Investigaciones Agronomicas (CIA),

Universidad de Costa Rica (UCR), San Jose, Costa Rica

M. Morales

Panamerican Woods S.A., P.O. Box 7842-1000, San Jose,

Costa Rica

M. Marchamalo-Sacristan

Dpto. Ingenierıa y Morfologıa del Terreno, E.T.S.I.

Caminos, Canales y Puertos, Universidad Politecnica de

Madrid (UPM), Ciudad Universitaria s/n, 28040 Madrid,

Spain

123

Nutr Cycl Agroecosyst (2014) 98:155–167

DOI 10.1007/s10705-014-9603-3

Author's personal copy

and other goods (Evans 2009; FRA 2010). Meanwhile,

forest managers have been increasingly concerned

about maintaining high productivity rates through

several rotations, especially in short-rotation planta-

tions, and the relationship between forest nutrition,

soil management and sustainable timber production

(e.g., Nambiar 1995; Fox 2000). It has long been

recognized that forest growth depends on the ability of

soil to maintain a supply of required nutrients.

However, soil nutrient availability can be modified

by management practices, such as fertilizer use (e.g.

Rennie 1955; Miller 1981; Fox 2000). A requirement

for fertilization regimes, to compensate for nutrient

export through timber extraction, is the long-standing

specification as indicated by some authors (e.g. Rennie

1955; Worrel and Hampson 1997). However, as

Folster and Khanna (1997) emphasised, such fertil-

ization provision has been traditionally neglected.

Nowadays, in order to enhance forest productivity,

sustain site fertility, and avoid soil nutrient depletion,

fertilization is utilized for intensively managed forests

across the globe (Ballard 1984; Goncalves et al. 1997).

Assuming that deficient soil nutrients have been

identified, fertilization programs should be designed

considering the following aspects: (a) what fertilizer to

use; (b) when to apply it; (c) how much is needed;

(d) how often to apply it; and, (e) by what method to

apply it (Ballard 1984; Bertsch 1998). The current

situation in Central America is that fertilization

programmes for forest plantations in most cases have

been designed taking into account general rules and a

quick interpretation of soil analyses, based on non-

specific critical levels (Bertsch 1998). A single

fertilization recommendation is usually applied to

large plantations of several square kilometres, without

taking into account any soil fertility heterogeneity. An

important consequence of the precision agriculture

approach was a trend towards heterogeneity of crop

fertilizer application, with modification of formula and

rate according to changing requirements within indi-

vidual fields, rather than simply considering each field

as a whole (Robert 2002). Such precision farming can

be established around (1) site-specific management,

e.g., management focused principally upon soil type

heterogeneity within each field, assuming their micro-

climate can be considered homogeneous, or (2)

management zones with treatments specified only at

a greater scale, across groups of sites, for cases when

budgetary or other restrictions limit the scope for a

wider range of management treatments. A major

barrier for site-specific management is the economic

cost of generating a satisfactory soil map (Robert

2002). Analogously, Fox (2000) observed that ‘‘site-

specific management is the key to sustaining soil

quality and long-term site productivity’’ for inten-

sively-managed forest plantations. The delimitation of

‘stands’ is one of the basic principles of forest

management. A ‘stand’ is regarded as a homogeneous

group of trees growing together on a sufficiently

uniform site. Forest management is not as intensive as

agriculture can be, and in practice, little consideration

is given to establishing stand-specific nutritional

plans. However, forest sites are amenable to grouping

by similarity in their soil fertility, through which

managers could delineate nutritional management

areas (groups of stands), and therefore facilitate more

efficient and productive management.

In this study, it was evaluated how effectively

multivariate statistical analysis can contribute to

decision-making when used as a tool for analysing

soil fertility databases, to classify the stands of a forest

plantation according to their soil fertility, and thereby

a specific fertilization program could be designed for

each of the defined groups. The intention is to expose a

case study that illustrates how these analyses can be

performed, using data from a specific forest plantation

in Costa Rica. However, the objectives are not to make

an interpretation of the results in terms of nutritional

status and quantitative fertilization needs of the

plantations, evaluate possible growth responses after

the fertilization, or elaborate maps of soil fertility of

the plantations. The aim is just to explore the

capabilities of the multivariate techniques and show

the possibility of making groups of the already

existing stands according to their soil fertility simi-

larities, in order to be easily used to improve forest

fertilization programs.

Materials and methods

Study area, sites and field sampling

Teak (Tectona grandis L.f.) has been extensively used

for forest plantations in Central America, originally in

Costa Rica and Panama (De Camino et al. 2002), and

more recently in Guatemala, El Salvador and Nicara-

gua. Across the region, teak plantations are intensively

156 Nutr Cycl Agroecosyst (2014) 98:155–167

123

Author's personal copy

managed in rotations of 20–25 years, usually in

carefully selected productive sites, with commercial

volume expected to be around 10 m3 ha-1 year-1

(Pandey and Brown 2000; De Camino et al. 2002).

Forest fertilization at establishment has become a

common practice for intensively managed forest

plantations in Central America, but fertilization at an

intermediate or even mature age is not a common

practice in the region. Notwithstanding the primary

importance of site selection as an issue for teak

plantation management, subsequent fertilization is

also necessary. Such amelioration can fulfil the high

nutrient demand of teak trees, thereby maintaining the

high nutrient concentrations they exhibit (Drechsel

and Zech 1991; Fernandez-Moya et al. 2013), and

promoting the productivity and sustainability of

production sites (e.g., Prasad et al. 1986; Liang et al.

2005; Zhou et al. 2012).

The case study was located on the North Pacific

coast of Costa Rica (Fig. 1), in a company managed

teak plantation1 which is divided in two sites: Carrillo

(1,040 ha) and Palo Arco (1,488 ha). The climate of

the region is classified as tropical wet forest, following

Holdridge’s life zones, with a mean annual rainfall of

2,500 mm, and a dry season of 4–6 months. Most

common soils are fertile reddish clayey (Table 1),

described as Typic Rhodustalfs mixed with Typic

Dystrustepts in Carrillo, and Typic Haploustalfs

mixed with Vertic Haploustepts in Palo Arco, with

small clusters of other soils. Soils are derived from

sedimentary limestone and basalt parent material.

The plantations were chosen to be representative of

properly-managed teak-planted forests in Central

America. In general, management of these plantations

consists of continuous forestry management activities:

fertilization at establishment; weed control; pruning;

and thinning (from approximately 800 trees ha-1 at

establishment to 150–200 trees ha-1 by final felling).

The use of clones has become common in recent years.

An expected commercial volume of 100–150 m3 is

expected for this kind of plantation, over a rotation of

approximately 20 years.

Through the company’s routine activity, a database

was created for the plantations under study, compris-

ing a total of 195 samples of topsoil (0–20 cm) from

across all the different stands, 75 and 129 from the

Carrillo and Palo Arco plantations, respectively.

Topsoil (0–20 cm) nutrient availability estimates are

commonly used for forest fertilization planning in

Central America, as fine root absorption is reported to

be most active in this soil layer, whether in plantations

of teak, or those of other species (Srivastava et al.

1986; Goncalves et al. 1997; Behling 2009). Soil

samples were analysed at the Centro de Investigaci-

ones Agronomicas from the University of Costa Rica

(CIA-UCR), to determine the following variables: pH

(in water), exchangeable Ca, Mg, K, P, Fe, Cu, Zn, Mn

and acidity. pH was determined in water 10:25;

Fig. 1 Location of the two study sites, Carrillo and Palo Arco, on the north Pacific coast of Costa Rica (Nicoya Peninsula), comprising

two teak (Tectona grandis L.f.) plantations owned by Panamerican Woods Ltd.

1 The teak plantations used as a study case in this work are both

owned by the Panamerican Woods Ltd. company (hereafter

‘PAW’).

Nutr Cycl Agroecosyst (2014) 98:155–167 157

123

personal copy

acidity, Al, Ca andMg in KCl solution 1 M 1:10; P, K,

Zn, Fe, Mn and Cu in modified Olsen solution pH 8,5

(NaHCO3 0.5 N, EDTA 0.01 M, Superfloc 127) 1:10.

The effective cation exchange capacity (ECEC) was

calculated as the addition of Ca, Mg, K and acidity

[ECEC = Ca ? Mg ? K ? acidity]. Ca saturation

(Ca S.), Mg saturation (Mg S.), K saturation (K S.)

and acidity saturation (A. S.) were calculated as the

percentage of ECEC relative to each of the

components.

Multivariate statistical methods

Different multivariate analysis methods were

employed for simplifying the data, either through

graphic representation of similarities between plot

points (ordination methods), or through grouping of

similar samples into discrete classes (classification

techniques) (Oksanen 2010). Both approaches are

based on methods to estimate the similarities or

dissimilarities between different objects, based on the

values of a set of variables measured on each of the

objects. Selecting the dissimilarity measure is of

primary importance to multivariate analysis (Oksanen

2011). Several distance measures, such as Bray–

Curtis, have been considered appropriate in various

ecological community studies, but Euclidean distance

is considered to be the best-disposed dissimilarity

measure for this study, as it fulfils the metric

properties, is based upon squared differences, and is

dominated by single large differences (Oksanen

2011). Data standardization and transformation are

critical in the process of selecting between different

methodologies (Kenkel 2006).

Principal component analysis (PCA) and non-

metric multidimensional scaling (NMDS) are the

most commonly used ordination methods. PCA is

based on orthogonal axes, Euclidean space and linear

rotation, with an assumption of normally-distributed

data, being analogous to simple linear regression

(Kenkel 2006). NMDS does not require any underly-

ing assumptions of linearity, and so has emerged as

one of the more robust and widely-used techniques,

especially in ecology and related disciplines

Table 1 Summary of analysed soil properties for the Panamerican Woods Ltd. teak (Tectona grandis L.f.) plantations on the north

Pacific coast of Costa Rica

Carrillo (n = 75) Palo Arco (n = 120) General (n = 195)

Mean SE CV (%) Mean SE CV (%) Mean SE CV (%)

pH 5.8 0.06 9.7 6.0 0.04 6.4 5.9 0.03 7.9

Ca [cmol (?) L-1] 27 1.11 35.5 26.4 0.74 30.6 26.6 0.62 32.5

Mg [cmol (?) L-1] 7.7 0.44 49.2 8.7 0.35 43.7 8.3 0.27 45.9

K [cmol (?) L-1] 0.2 0.01 63.7 0.2 0.02 84.8 0.2 0.01 77.2

Acidity [cmol (?) L-1] 0.2 0.01 42.4 0.1 0.00 32.3 0.2 0.01 43.3

ECEC [cmol (?) L-1]a 35.1 1.36 33.6 35.4 0.98 30.2 35.3 0.80 31.5

P (mg L-1) 1.5* 0.29 168.3 3.2* 0.43 145.5 2.5* 0.29 159.4

Cu (mg L-1) 7.0 1.21 150.1 12.6 0.61 52.7 10.4 0.62 83.9

Fe (mg L-1) 25.6 2.78 94.2 38.3 3.58 102.3 33.4 2.48 103.7

Mn (mg L-1) 41.4 1.95 40.7 25.3 1.79 77.5 31.5 1.44 63.9

Zn (mg L-1) 15.7 1.99 109.8 3.1 0.25 86.7 7.9 0.89 157.2

A. S. (%) 0.6 0.05 73.0 0.4 0.02 59.6 0.5 0.03 72.1

Ca S. (%) 76.8 0.76 8.6 74.8 0.55 8 75.6 0.45 8.3

Mg S. (%) 22.0 0.75 29.7 24.3 0.56 25.2 23.4 0.46 27.2

K S. (%) 0.6 0.06 88.7 0.5 0.04 88.4 0.5 0.03 88.5

Means, standard errors (SE) and coefficients of variation (CV) are provided. Values marked with * are lower than adequate reference

soil levels (after Bertsch 1998). The ‘General’ column shows the values when calculated across all the samples for both plantation

sitesa ECEC effective cation exchange capacity

158 Nutr Cycl Agroecosyst (2014) 98:155–167

123

Author's personal copy

(Minchin 1987; Kenkel 2006; Oksanen 2011). Con-

versely, NMDS does present some disadvantages, in

particular: (a) NMDS is unable to interpret the relative

importance of the ordination axes when summarizing

the variation of the data; and (b) NMDS cannot

produce a true ordination bi-plot, as variable weights

are not determined (Kenkel 2006). While these

disadvantages have caused some authors to refrain

from advocating adoption of NMDS (Kenkel 2006), in

the context of the objectives and data structure of this

study, the disadvantages were considered to be of

negligible importance. Cluster analysis is a customary

classification method which incorporates the calcula-

tion of a distance matrix (similar to that used for

ordination methods), from which objects can accord-

ingly be classified. Complete linkage (or farthest

neighbour) hierarchic clustering was considered the

best option for our data and objectives, as this method

is based upon maximizing the distance between

groups or clusters (Oksanen 2010).

Data analysis

The topsoil database was used to perform different

multivariate analyses of the soil test data in order to

group the sampled soils according to similarities

between the measured properties. The use of different

multivariate analysis methods allowed comparing

their usefulness for grouping similar soil samples

(Table 2). One set of analyses was carried out with the

soil test variables centred using their means, along

with an alternative set of analyses that instead used the

soil test critical levels to centre the variables (Bertsch

1998). In both cases, each variable was standardized

using its standard deviation. PCA was performed with

the entire dataset, comprising the 195 samples from

both plantation sites. NMDS was also performed with

the general dataset (this analysis is hereafter referred

to as the ‘G-NMDS’). Additionally, two NMDS

analyses were constructed: (a) one analysis for the

75 samples from the Carrillo plantation (‘C-NMDS’);

and (b) a second analysis for the 120 samples from

Palo Arco (‘PA-NMDS’). Five cluster analyses were

carried out using the entire dataset (195 samples from

both plantations), in order to distinguish: (a) two

groups, (b) three groups, (c) four groups, (d) five

groups, and (e) six groups. Four additional cluster

analyses were computed, two for the Carrillo and two

for the Palo Arco plantations, respectively, in order to

distinguish two and three groups for each plantation.

The coefficient of variation (hereafter ‘CV’) was

calculated for each variable in each of the constructed

soil groups, and the average CV for each group was

determined as:

CVj ¼ averageCVij

where CVi j is the CV for each of the study variables

(i) for each group (j).

The reduction of the CV for each variable in each of

the constructed soil groups relative to the original CV

for the 195 samples (CVi general) was estimated as:

DCVij ¼ CVij=CVi general

The average reduction of the CV of the study

variables for each group was calculated as:

DCVj ¼ averageDCVij ð1ÞThe CV calculations allowed an estimate of the

homogeneity for each of the groups identified; a

comparison to be made against the null hypothesis of

‘no-groups’; and to identify which method resulted in

the best grouping.

NMDS and cluster analysis were done using the

Vegan library in R (R Development Core Team 2011).

Euclidean distance was used as the measure of

dissimilarity. No rotation was used for the PCA or

the NMDS analyses. For the NMDS analyses, the

number of k dimensions was set to k = 2. A Shepard

diagram ‘stress-plot’ was constructed as a measure of

the goodness of fit for the NMDS analysis (Oksanen

2011).

Results and discussion

No important disparities were found (data not shown)

between the results of multivariate analyses obtained

using the mean or the critical value as a reference for

centring the data (Bertsch 1998). Therefore, we

hereafter describe only the results of the former

analyses, i.e., from normally-standardized data that

used mean and standard deviation as references.

Similarly, minor differences were observed between

the PCA and the NMDS constructed through the

‘general’ analyses, using the data from both planta-

tions (data not shown). The NMDS provided the best

representation of the differences between soil samples,

Nutr Cycl Agroecosyst (2014) 98:155–167 159

123

Author's personal copy

and the Shepard plot showed a non-metric fit to be

better than a linear fit (non-metric fit pseudo

R2 = 0.978; linear fit pseudo R2 = 0.936), consoli-

dating our interpretation of NMDS as the better

ordination method. Hence, only NMDS analysis was

carried out for the Carrillo and Palo Arco plantation

data independently.

Palo Arco plantation has generally been considered

to exhibit higher soil fertility than Carrillo plantation,

to the extent that different nutritional management

plans have been designed for the two plantations.

However, the average soil data results for the planta-

tions in Carrillo and Palo Arco showed similar values

(Table 1), with the possible exception of the P and Zn.

Furthermore, the soil samples from Carrillo could not

be differentiated from those of Palo Arco, when we

investigated the similarities between soil samples

using the ‘general’ NMDS analyses (Fig. 2). This

contradiction shows how the traditional methods being

used nowadays in many large forest plantations in

Central America can be improved using new tech-

niques and how this improvement could result in a

more appropriate soil and nutrient management in

those ecosystems.

Cluster analyses were used to distinguish the

following: two, three, four, five and six groups of soil

samples from the entire dataset in general; and, two

and three groups from independent analyses of

Carrillo and Palo Arco data, respectively (Table 2).

Figure 2 represents the ‘G-NMDS’ general analysis

across all 195 samples, plotted in accordance with the

plantation field (Carrillo or Palo Arco), while Fig. 3

does it according to the groups defined by cluster

analysis. Table 3 summarizes the trend in CV that was

evident as more groups were differentiated by cluster

analysis: with increasing the number of groups, each

Table 2 Summary of the

different multivariate

analyses performed in the

study

Type of

analysis

Origin of the

data

Number of

samples

Name Number of

groups

Reference for

centering

PCA General 195 G-PCA – Average

– Critical value

NMDS General 195 G-NMDS – Average

– Critical value

Carrillo 75 C-NMDS – Average

– Critical value

Palo Arco 120 PA-NMDS – Average

– Critical value

Cluster General 195 G-2 2 Average

Critical value

G-3 3 Average

Critical value

G-4 4 Average

Critical value

G-5 5 Average

Critical value

G-6 6 Average

Critical value

Carrillo 75 C-2 2 Average

Critical value

C-3 3 Average

Critical value

Palo Arco 120 PA-2 2 Average

Critical value

PA-3 3 Average

Critical value

160 Nutr Cycl Agroecosyst (2014) 98:155–167

123

personal copy

group became more homogeneous, and the CV for

each variable diminished. As repeated cluster analyses

progressively distinguished more groups, each corre-

sponding group could then be diagrammatically

isolated within the NMDS (Fig. 3). Hence, a small

increase in operational cost would allow an improve-

ment of fertilizer efficiency and it would translate into

a higher economic return. However, from a theoretical

point of view, at some number of these nutritional

groups, the increase (marginal) in benefit should be

equal to the increase (marginal) in cost and further

increase in the number of groups should result in

negative increments of the benefits. In addition to this

economical reason, the amount of groups cannot be

higher than a reasonable number in order to be

practical for the company managers; groups in excess

of this number would ultimately contribute to generate

disproportionate complexity in this approach to forest

management, to the extent that we could anticipate

abandonment of such practices. We judged that a

maximum of six groups was an appropriate number of

soil groups for a 2,500 ha plantation.

Multivariate statistical techniques have been

widely applied in soil sciences, notably in the analysis

of metal contamination (e.g., Yay et al. 2008), but also

in precision agriculture, delineation of site-specific

management zones, and soil classification and map-

ping (e.g., Theocharopoulos et al. 1997; Kalahne et al.

2000; Jaynes et al. 2005; Ortega and Santibanez 2007;

Yan et al. 2007; Fu et al. 2010; Arrouays et al. 2011).

Fu et al. (2010) identified clustering as an appropriate

analytical technique to delineate soil nutrient

management zones, and therefore it provides an

effective basis to establish variable-rate fertilization

regimes for precision agriculture. However, Fu et al.

(2010) also noted that clustering methods were

sensitive to the iterative initial value. For the Vegan

package in the R-software environment (R Develop-

ment Core Team 2011), this issue can be addressed by

using the metaMDS function, which allows establish-

ing several random starts, and selects from similar

solutions with smallest stresses (Oksanen 2011).

Ortega and Santibanez (2007) identified cluster ana-

lysis as a better technique for delineating homoge-

neous management zones, relative to alternative

methods. Multivariate techniques have also been

applied to precision agriculture in association with

geostatistical techniques (e.g., Castrignano et al. 2005;

Morari et al. 2009; Arrouays et al. 2011). However, in

the present context of fertilization management for

Central American forest plantations, nowadays we do

not consider analysis from this perspective to be

justified, given the degree of complexity associated

with the techniques. Rather, we consider that the

proposed strategy, classifying stands into groups with

similar soil properties, affords greater scope for

organizing the already existing stands into manage-

ment zones, given that it readily facilitates identifica-

tion of a limited number of nutritional management

groups. As the stands are considered as a homoge-

neous unit, no further detail taking into account

geostatistics, geographical location or spatial analysis

is considered necessary at this time. The delineation of

intra-field management zones, i.e., zones of uniform

Fig. 2 Non-metric

multidimensional scaling

for the 195 topsoil samples

from the teak (Tectona

grandis L.f.) plantations

owned by Panamerican

Woods Ltd., on the north

Pacific coast of Costa Rica:

Carrillo (times symbol) and

Palo Arco (circle)

Nutr Cycl Agroecosyst (2014) 98:155–167 161

123

Author's personal copy

management, has been assessed as an important initial

stage in the implementation of site-specific nutrient

management (Ortega and Santibanez 2007).

In the context of this study, NMDS emerged as a

better ordination method than PCA (Figs. 2, 3). How-

ever, it was important to initially consider both PCAand

Fig. 3 Graphical representation of a two, b three, c four, d five,

and e six groups, as defined by cluster analysis, based on the

spatial scores of the NMDS for the 195 topsoil samples from the

teak (Tectona grandis L.f.) plantations owned by Panamerican

Woods Ltd., on the north Pacific coast of Costa Rica

162 Nutr Cycl Agroecosyst (2014) 98:155–167

123

Author's personal copy

NMDS for these analyses, in order to identify the

method most appropriate for the data and objectives in

question, as the best option can be anticipated to vary on

a case-specific basis (Kenkel 2006). When we distin-

guished six groups from the entire dataset in general, the

groups were more homogeneous than those that

emerged when independently deriving three groups

from the Carrillo data and three from Palo Arco

(Table 3). As no notable differences were evident when

making comparisons between Carrillo and Palo Arco

data (Fig. 2), analysing these data independently was

not considered a useful basis for further similar analyses.

Relatively high microelement concentrations are

typically required in order to maintain an appropriate

nutritional status for trees in teak plantations, and

indeed other forest plantations globally (Goncalves

et al. 1997; Lehto et al. 2010; Zhou et al. 2012;

Fernandez-Moya et al. 2013). However, little attention

has been paid to Zn and B in other studies of teak

nutrition. Tropical soils are usually characterized as

Table 3 Reduction in the

average coefficient of

variation (CV) for soil

fertility variables in each

group, distinguished by the

cluster analysis treatments,

relative to the null

hypothesis (‘no-grouping’,

i.e., one single group of data

encompassing both Carrillo

and Palo Arco together)

Values marked with * show

a relative CV reduction of

between 20 and 35 %.

Values marked with **

show a relative CV

reduction of more than

35 %

Group Average

CV (%)

D average

CV (%)

Number of soil

samples in the group

Null hypothesis (no-

grouping)

66.8 – 195

Grouping by plantation Carrillo 66.5 -0.4 75

Palo Arco 58.3 -12.7 120

G-2 Group 1 60.2 -10.5 158

Group 2 56.2 -9.7 37

G-3 Group 1 55.8 -14.1 157

Group 2 56.2 -9.7 37

Group 3 – – 1

G-4 Group 1 60.2 -10.5 157

Group 2 52.3 -17.4 35

Group 3 51.8 -22.5* 2

Group 4 – – 1

G-5 Group 1 60.2 -10.5 157

Group 2 40.2 -40.9** 19

Group 3 38.2 -36.6** 16

Group 4 51.8 -22.5* 2

Group 5 – – 1

G-6 Group 1 35.7 -45.5** 5

Group 2 40.2 -40.9** 19

Group 3 55.0 -17.7 152

Group 4 38.2 -36.6** 16

Group 5 51.8 -22.5* 2

Group 6 – – 1

C-2 Group 1 -2.8 -2.7 74

Group 2 – – 1

C-3 Group 1 -19.2 -16.3 42

Group 2 -17.9 -21.4* 32

Group 3 – – 1

PA-2 Group 1 -14.0 -26.8* 110

Group 2 -13.3 -29.0* 10

PA-3 Group 1 – – 1

Group 2 -17.4 -31.4* 109

Group 3 -13.3 -29.0* 10

Nutr Cycl Agroecosyst (2014) 98:155–167 163

123

Author's personal copy

Table

4Averageandstandarderror(SE)ofsoilattributesforthesixstandgroupsdefined

forthePanam

erican

Woodsteak

(TectonagrandisL.f.)plantationsin

Guanacaste

(CostaRica)

based

ontheirsimilaritiesin

theanalysedsoilfertilityattributes

Group1(n

=5)

Group2(n

=19)

Group3(n

=152)

Group4(n

=16)

Group5(n

=2)

Group6(n

=1)

Mean

andSE

CV

(%)

DCV

(%)

Mean

andSE

CV

(%)

DCV

(%)

Mean

andSE

CV

(%)

DCV

(%)

Mean

andSE

CV

(%)

DCV

(%)

Mean

andSE

CV

(%)

DCV

(%)

Mean

andSE

CV

(%)

DCV

(%)

pH

7.0 (0

.1)

4-44

6.1

(0)

3-61

5.9

(0)

7-12

5.4 (0

.1)*

6-23

5.1 (0

.1)*

2-70

7.5

(–)

––

Ca (cmol?

L-1)

42.8 (1.72)

9-71

25.2 (0.93)

16

-51

27.8 (0.59)

26

-19

13.3 (0.76)

23

-29

4.9 (0

.56)

16

-52

53.2

(–)

––

Mg (cmol?

L-1)

3.6 (0

.58)

36

-22

5.3 (0

.38)

31

-32

9.5 (0

.28)

36

-21

3.3 (0

.16)

19

-59

2.3 (0

.59)

36

-23

4.3

(–)

––

K

(cmol?

L-1)

0.2 (0

.03)

45

-41

0.4 (0

.05)

50

-36

0.1 (0

.01)*

61

-21

0.2 (0

.02)

65

-15

0.2 (\

0.01)

0-100

0.6

(–)

––

A

(cmol?

L-1)

0.1 (0

.02)

31

-29

0.2 (0

.01)

27

-39

0.2 (0

.01)

45

40.2 (0

.02)

43

-2

0.2 (0

.09)

61

40

0.1

(–)

––

CICE

(cmol?

L-1)

46.7 (1.5)

7-78

31.0 (1.2)

17

-45

37.6 (0.8)

26

-19

16.9 (0.8)

20

-37

7.6

(1.2)

23

-28

58.2

(–)

––

P(m

gL-1)

2(0.3)*

35

-78

5(1.3)*

114

-28

2(0.2)*

114

-29

2(0.3)*

54

-66

2(2)*

140

-12

41(–)

––

Cu(m

gL-1)

1(0.2)

41

-51

11(1.5)

58

-30

9(0.5)

67

-20

32(2)

25

-70

8(6.7)

119

42

8(–)

––

Fe(m

gL-1)

7(1.6)*

52

-50

32(3.6)

49

-53

25(1.5)

73

-29

121 (13)

43

-58

38(22.8)

85

-18

22(–)

––

Mn(m

gL-1)

22(0.6)

6-91

24(3.6)

66

430(1.5)

61

-5

58(5.5)

38

-41

58(32.8)

80

26

5(–)

––

Zn(m

gL-1)

1(0.7)*

146

-7

4(0.6)

69

-56

9(1.2)

160

28(1.8)

91

-42

18(16.8)

132

-16

3(–)

––

AS.(%

)0.3

(0)

34

-53

0.5

(0)

32

-55

0.4

(0)

52

-28

0.9 (0

.1)

51

-29

2.6

(0.8)

41

-43

0.2

(–)

––

CaS.(%

)91.5 (1.2)

3-60

81.5 (0.7)

4-51

74.1 (0.4)

7-13

78.1 (1.2)

6-32

65.3 (3.2)

7-15

91.4

(–)

––

MgS.(%

)7.9 (1

.3)

38

39

16.9 (0.7)

18

-34

25.1 (0.4)

22

-21

20.2 (1.1)

21

-21

29.9

(3)

14

-50

7.4

(–)

––

KS.(%

)0.3 (0

.1)

47

-47

1.3 (0

.1)

49

-45

0.4

(0)

69

-22

0.9 (0

.2)

67

-24

2.2

(0.4)

23

-75

1(–)

––

Values

marked

with*arelower

than

adequatereference

soillevels(Bertsch

1998).Coefficientsofvariation(%

)foreach

meanarealso

provided.DCV(%

)isthereductionin

the

coefficientofthevariationforthevariable

when

comparingtheproposedgroupingagainst

thenullhypothesis

ofonesingle

groupthat

includes

allthestandsin

thedataset

164 Nutr Cycl Agroecosyst (2014) 98:155–167

123

Author's personal copy

highly-weathered, and rich in Fe or Mn, but generally

deficient in Zn, B, Cu and Mo (Barker and Pilbeam

2006). B is typically deficient in soils on a global scale,

and is difficult to evaluate in routine soil fertility

analyses (Lehto et al. 2010). There is therefore still a

requirement to implement specific evaluations of B

and Zn status for forest plantations throughout the

tropics. The advantages of multivariate analysis

techniques are of particular relevance in this respect.

Multivariate analyses can be used to process a large

number of variables, and can therefore readily incor-

porate the range of micronutrients that fertilization

planning must take into account.

‘‘The amount of fertilizer to be applied to a given

species at a particular site will depend on the level of

soil fertility and productivity’’ (Goncalves et al. 1997).

However, practical management of any fertilization

program established on an explicitly stand-specific

basis, applying a different fertilization formula and

dosage to each stand, is generally regarded as

impractical by forest plantation managers. As a

contrasting approach, we propose that grouping stands

by similarities in soil fertility represents a more

practical strategy, in that it facilitates the allocation

of sites into a manageable range of soil fertility

classes. This process of classification promotes the

design of a versatile fertilization regime that is

sufficiently proximate to differing soil requirements

across all the sites in question.

Analysis of soil data in the context of grouped

samples allows us to carry out soil fertility diagnosis

with greater precision, and establish a basis for

improved nutritional and fertilization planning. This

is exemplified by the improvement in precision

presented by the results in Tables 3 and 4. In Table 1,

which deals with traditional methods for fertilization

planning, the soils are presented only as being P-defi-

cient. Hence, if forest managers only take this into

consideration, they would design a fertilization

programme to solve this deficiency (e.g. application

of a phosphorus fertilizer) to the entire plantation area.

In comparison, the proposed more detailed grouping

analysis (Table 4) indicates that most groups exhibit

additional deficiencies. Group 1 shows P, Fe and Zn

deficiencies; groups 4 and 5 have low pH values in

addition to low values of P; group 3 (representing the

majority of the samples) shows low K and P content.

Conversely, group 6 indicates exceptional soil that is

extraordinarily high in all nutrients. Only group 2 still

accords with the results for Table 1, in that it demon-

strates a deficiency only in P but with some relatively

high pH; thus it is the only group which would have a

similar fertilization compared with the initial scenario.

On the other hand, a fertilizer formula with P, Fe and

Zn would be applied for the stands in group 1; a

commonN–P–K formulawould probably be applied to

group 3, while some specific P fertilizer would be

needed for groups 4 and 5 with a relatively high

basicity index in order to solve the relative low pH

values or a common P fertilizer can be applied with

previous liming of those stands. Hence, with the

traditional methods for fertilization planning the

majority of the stands would have a hidden nutrient

deficiency that would be lowering the productivity of

the plantations, except group 6 (a single stand) that

shows high soil fertility with no need to be fertilized.

This single stand could have been considered as a

statistical outlier; however, it has been considered that

a whole stand cannot be removed from the analysis as

in the decision-making process done by forest manag-

ers; something needs to be done with every stand, even

if it is quite different to the others.

Conclusions

Cluster analysis can be applied as an appropriate tool

for grouping forest stands according to their soil

fertility status and, consequently, for designing fertil-

izer use programs appropriate to the disparate require-

ments across differing groups of stands, where each

group exhibits relatively homogeneous soil fertility

properties. NonMultidimensional Scaling represents a

useful complementary tool for graphically exploring

the similarities of soil fertility within groups of forest

stands, and the differences between those stands.

Multivariate analysis provides techniques to clas-

sify soil groups by integrating a large number of soil

fertility variables, such as micronutrient concentration

values, from across a large number of soil samples. By

designing forest fertilization plans for groups of

stands, where each group comprises stands with

homogeneous soil fertility properties, fertilizer appli-

cation can therefore be implemented with much

greater efficiency and productivity.

Acknowledgments The authors would like to acknowledge

the collaboration of Panamerican Woods Ltd. in regard to

Nutr Cycl Agroecosyst (2014) 98:155–167 165

123

Author's personal copy

providing access to the soil database. The authors also thank

Paul Robertson, Adam Collins and the personnel of the Natural

Resources Laboratory at CIA (UCR) for their help and

comments about English language, manuscript format and for

their assistance about the contents of this paper. One of the

authors, M. Morales, is an employee of a forest plantation

company (Panamerican Woods) but none of the authors show

any kind of conflict of interests. The present paper was

conducted under the MACOSACEN project, financed by PCI-

AECID. The authors also thank to two anonymous reviewers for

their comments to the present manuscript.

References

Arrouays D, Saby NPA, Thioulouse J, Jolivet C, Boulonne L,

Ratie C (2011) Large trends in French topsoil character-

istics are revealed by spatially constrained multivariate

analysis. Geoderma 161:107–114

Ballard R (1984) Fertilization of plantations. In: Bowen GD,

Nambiar EKS (eds) Nutrition of plantation forests. Aca-

demic Press, London, pp 327–360

Barker AV, Pilbeam DJ (2006) Handbook of plant nutrition.

CRC Press, Boca Raton, FL

Behling M (2009) Nutricao, particao de biomassa e crescimento

de povoamentos de Teca em Tangara da Serra-MT. PhD

thesis. Universidade Federal de Vicosa, Brasil

Bertsch F (1998) La fertilidad de los suelos y su manejo. Aso-

ciacion Costarricense de la Ciencia del Suelo, San Jose

Castrignano A, Cherubini C, Giasi CI, Castore M, Di Mucci G,

Molinari M (2005) Using multivariate geostatistics for

describing spatial relationships among some soil proper-

ties. In: ISTRO conference, Brno, Czech Republic

De Camino RV, Alfaro MM, Sage LFM (2002) Teak (Tectona

grandis) in Central America. Forest plantations working

paper 19. Forest Resources Development Service, Forest

Resources Division, FAO, Rome

Drechsel P, Zech W (1991) Foliar nutrient levels of broad-

leaved tropical trees: a tabular review. Plant Soil

131(1):29–46

Evans J (2009) Planted forests: uses. Impacts and sustainability.

CAB International, FAO, Rome

Fernandez-Moya J, Murillo R, Portuguez E, Fallas JL, Rıos V,

KottmanF,Verjans JM,MataR,AlvaradoA (2013)Nutrient

concentrations age dynamics of teak (Tectona grandis L.f.)

plantations in Central America. For Syst 22(1):123–133

Folster H, Khanna PK (1997) Dynamics of nutrient supply in

plantation soils. In: Nambiar EKS, Brown AG (eds) Man-

agement of soil, nutrients and water in tropical plantation

forests. Australian Centre for International Agricultural

Research, Australia, pp 339–379

Fox TR (2000) Sustained productivity in intensively managed

forest plantations. For Ecol Manag 138:187–202

FRA (2010) Global forest resources assessment 2010. FAO,

Rome

Fu Q, Wang Z, Jiang Q (2010) Delineating soil nutrient man-

agement zones based on fuzzy clustering optimized by

PSO. Math Comput Model 51:1299–1305

Goncalves JLM, Barros NF, Nambiar EKS, Novais RF (1997)

Soil and stand management for short-rotation plantations.

In: Nambiar EKS, Brown AG (eds) Management of soil,

nutrients and water in tropical plantation forests. Australian

Centre for International Agricultural Research, Australia,

pp 379–418

Jaynes DB, Colvin TS, Kaspar TC (2005) Identifying potential

soybean management zones from multi-year yield data.

Comput Electron Agric 46:309–327

Kalahne R, AminM, Sobottka J, Sauerbrey R (2000)Methods of

multivariate data analysis applied to the investigation of

fen soils. Anal Chim Acta 420:205–216

Kenkel NC (2006) On selecting an appropriate multivariate

analysis. Can J Plant Sci 86:663–676

Lehto T, Ruuhola T, Dell B (2010) Boron in forest trees and

forest ecosystems. For Ecol Manag 260:2053–2069

Liang KN, Pan YF, Liu WM (2005) Multi-factors fertilization

trial of Tectona grandis in seedling stage. For Res

188(5):535–540

Miller HG (1981) Forest fertilization: some guiding concepts.

Forestry 54(2):157–167

Minchin PR (1987) An evaluation of the relative robustness of

techniques for ecological ordination. Theory Models Veg

Sci 69:89–107

Morari F, Castrignano A, Pagliarin C (2009) Application of

multivariate geostatistics in delineating management zones

within a gravelly vineyard using geo-electrical sensors.

Comput Electron Agric 68:97–107

Nambiar EKS (1995) Sustained productivity of plantations:

science and practice. Bosque 16(1):3–8

Oksanen J (2010) Cluster analysis: tutorial with R. University of

Oulu, OuluOksanen J (2011) Multivariate analysis of ecological commu-

nities in R: vegan tutorial. University of Oulu, Oulu

Ortega RA, Santibanez OA (2007) Determination of manage-

ment zones in corn (Zea mays L.) based on soil fertility.

Comput Electron Agric 58:49–59

Pandey D, Brown C (2000) Teak: a global overview. Unasylva

51(2):1–15

Prasad R, Sah AKY, Bhandari AS (1986) Fertilizer trial in ten

and twenty years old teak plantations in Nadhya Pradesh.

J Trop For 2(1):47–52

R Development Core Team (2011) R: a language and

environment for statistical computing. R Foundation for

Statistical Computing. Vienna, Austria. http://www.r-

project.org

Rennie PJ (1955) The uptake of nutrients by mature forest

growth. Plant Soil 8(1):49–95

Robert PC (2002) Precision agriculture: a challenge for crop

nutrition management. Plant Soil 247:143–149

Srivastava SK, Singh KP, Upadhyay RS (1986) Fine root growth

dynamics in teak (Tectona grandis Linn F.). Can J Plant Sci

16:1360–1364

Theocharopoulos SP, Petrakis PV, Trikatsoula A (1997)

Multivariate analysis of soil grid data as a soil classi-

fication and mapping tool: the case study of a homo-

geneous plain in Vagia, Viotia, Greece. Geoderma

77:63–79

Worrel R, Hampson A (1997) The influence of some forest

operations on the sustainable management of forest soils—

a review. Forestry 70(1):61–85

Yan L, Zhou S, Feng L, Hong-Yi L (2007) Delineation of site-

specific management zones using fuzzy clustering analysis

166 Nutr Cycl Agroecosyst (2014) 98:155–167

123

personal copy

in a coastal saline land. Comput Electron Agric

56:174–186

Yay OD, Alagha O, Tuncel G (2008) Multivariate statistics to

investigate metal contamination in surface soil. J Environ

Manag 86:581–594

Zhou Z, Liang K, Xu D, Zhang Y, Huang G, Ma H (2012)

Effects of calcium, boron and nitrogen fertilization on the

growth of teak (Tectona grandis) seedlings and chemical

property of acidic soil substrate. New For 43:231–243

Nutr Cycl Agroecosyst (2014) 98:155–167 167

123

Author's personal copy