Evaporation losses from bare soils as influenced by ......Evaporation losses from bare soils as...
Transcript of Evaporation losses from bare soils as influenced by ......Evaporation losses from bare soils as...
Evaporation losses from bare soils as influenced by
cultivation techniques in semi-arid regions
H.J. Melloulia, B. van Wesemaelb,*, J. Poesenc, R. Hartmannd
aInstitut National de la Recherche Agronomique de Tunisie, Rue HeÂdi Karray, 2049 Ariana, TunisiabFlood Hazard Research Centre, Middlesex University, Queensway, EN3 4SF Enfield, UK
cFund for Scientific Research Flanders, Laboratory for Experimental Geomorphology, K.U. Leuven,
Redingenstraat 16, 3000 Leuven, BelgiumdFaculty of Agricultural and Applied Biological Sciences, Universiteit Gent,
Coupure Links 653, 9000 Ghent, Belgium
Accepted 25 February 1999
Abstract
The impact of cultivation techniques on the evaporation from bare soils was investigated in the
laboratory. Two soil-types, which are important resources for rainfed cultivation of olives and
almonds in semi-arid regions, were selected: a loamy sand soil and a stony (loam) soil. Evaporation
from the soil surface is an important loss of soil moisture in these farming systems since a large
percentage of the soil is kept bare in order to maximise the water availability for the tree crop. For
the loamy sand soil the impacts of a straw mulch and treatment of the topsoil with olive mill
effluent (OME) were tested. For the stony soil the effects of different rock fragment contents and
distribution within the soil profile were tested. After thoroughly wetting with simulated rainfall and
allowing the soil moisture to redistribute, the columns were subjected to evaporation for 46 days.
Cumulative evaporation depth of soils treated with OME was 28% lower than that of the control
soil. A similar reduction, be it lower (16%) was observed for the soil with a high rock fragment
content by volume (Rv � 0.35 m3 mÿ3). The straw mulch and rock fragment mulch did not have an
impact on the cumulative evaporation depth after 46 days. Furthermore, the time required to reach
half of the total evaporation losses (d0.5) increased from 9 days for the control soil (loamy sand) to
24 days for the soil impregnated with OME and to 15 days for the straw mulch treatment. The same
trend was observed for the stony soils: an increase in d0.5 from 4 days for the control soil
(Rv � 0.19 m3 mÿ3) to 7 days for the soil with Rv � 0.35 m3 mÿ3 and to 8 days for the rock
fragment mulch. These experiments show that the changes in water retention capacity of the topsoil
by treatment with a hydrophobic substance (OME) or an increase in rock fragment content have a
Agricultural Water Management 42 (2000) 355±369
* Corresponding author. Tel.: +44-181-362-6611; fax: +44-181-362-6957E-mail address: [email protected] (B. van Wesemael)
0378-3774/00/$ ± see front matter # 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 3 7 8 - 3 7 7 4 ( 9 9 ) 0 0 0 4 0 - 2
longer lasting effect on the reduction of evaporation losses, and result in a higher and more evenly
distributed soil moisture content. # 2000 Elsevier Science B.V. All rights reserved.
Keywords: Evaporation losses; Cultivation techniques; Topsoils treated with olive mill ef¯uent; Stony soils;
Semi-arid regions
1. Introduction
Evaporation from bare soils results in a considerable loss of moisture and has a direct
impact on crop yield in rainfed agriculture of arid and semi-arid regions. Under annual
field crops the soil surface remains bare for many weeks when the moisture content of the
upper soil layer can be of critical importance during periods of seed germination and
seedling establishment as well as during the subsequent growth of the young crop. In
orchards, a typical component of agriculture in the Mediterranean region, the soil surface
between the trees is kept bare by frequent tillage and is continuously subjected to
evaporation. The very wide spacing between the trees allows a maximum catchment and
soil volume to collect and store the sparse rainfall for individual trees (spacing between
the trees extends to 24 m; Ennabli, 1993).
In the U.S. Midwest up to 50% of the maize evapotranspiration is lost by evaporation
during a normal growing season (Peters, 1960). Research conducted in Tunisia for two
decades demonstrated that evaporation from bare soils accounted for 75% of the annual
precipitation (Riou, 1977). Under similar climatic conditions, this water loss was
evaluated to be 30±61% of the annual precipitation for various rainfed winter wheat-
based production systems (Stewart and Burnett, 1987). By means of an evaporation
simulation model, Floret et al. (1982) estimated that in the south Tunisian steppe region
where the soil surface cover of perennial grasses (Stipa spp. ) is about 30%, half of the
available soil moisture is lost by evaporation from bare soil. In Syria, where soil nutrient
deficiency often severely limits crop growth, as much as 75% of the soil moisture is lost
by evaporation from the soil surface under barley (Cooper et al., 1987). Since these
evaporation losses are very important, their reduction will strongly contribute to soil
moisture conservation for crop production. This is even more urgent in arid and semi-arid
regions, characterised by highly variable and often chronically deficient rainfall. Many of
the countries in these regions will have to take up an important challenge at the beginning
of the next century: increasing food production in order to realise food security for a
growing population while optimising the use of limited water resources (FAO, 1994). In
addition, climate change will affect rainfall depth and its reliability which will increase
the threat of desertification.
Management techniques commonly called `mulching' are mostly based on the
modification of the soil surface conditions. Mulching enhances the formation of thin air-
dry layer on the top bare soil, which hampers capillary rise and slows the evaporation
process. Such a well-developed air-dry surface layer can either be achieved by frequent
tillage, commonly called dry farming (Willis and Bond, 1971; Jalota and Prihar, 1990a;
Mwendera, 1992) or by the application of a soil cover (e.g. straw, sand, gravel or plastic
film). The role of various types of soil cover in reducing evaporation has been widely
356 H.J. Mellouli et al. / Agricultural Water Management 42 (2000) 355±369
studied: crop residue mulches (Bond and Willis, 1969; Unger and Parker, 1976; Smika,
1983; Tanaka, 1985; Jalota and Prihar, 1990b; Gicheru, 1994; Mellouli, 1996), plastic
film (e.g. El Amami and Haffani, 1974), sand mulches (Modaihsh et al., 1985; Bousnina,
1993) and rock fragment mulches (Corey and Kemper, 1968; Groenevelt et al., 1989;
Kemper et al., 1994; van Wesemael et al., 1996). However, these mulches are labour-
intensive to apply and maintain (Gale et al., 1993) and increasingly crop residues have
become a valuable food supply for livestock.
Other techniques to reduce evaporation from bare soils are based on modification of
the hydrodynamics of the topsoil (i.e. to force water to penetrate deep into the soil
profile). Both the aggregate stability of the topsoil and its rock fragment content are
important factors in this respect. The increase in structural stability of the topsoil assists
in the rapid development of a dry mulch layer and can be achieved by mixing the topsoil
with a hydrophobic compound either of synthetic origin (e.g. soil conditioners, Hillel and
Berliner, 1974; Al-Jaloud, 1988; Verplancke et al., 1990) or of natural origin, for
example, Olive Mill Effluents (OME; Mellouli, 1996; Mellouli et al., 1998). OME, better
known in Tunisia as `margines', that is, liquor from olive oil extraction processes, are
characterised by their fertility properties (e.g. Fiestas Ros de Ursinos, 1986), as well as by
their adhesive and hydrophobic behaviour (e.g. Friaaà et al., 1986). Mellouli (1996) has
shown that OME has a beneficial influence on soil aggregation, soil structure stability and
hydrodynamic properties of a sandy soil, for example, liquid±soil contact angle (water
repellency) and saturated hydraulic conductivity were increased. Furthermore, unsatu-
rated hydraulic conductivity and water retention at the pressure potential range
corresponding to the available water capacity can be reduced by the application of
OME.
The main impact of the rock fragment content in stony soils on evaporation is
attributed to the reduction of the water holding capacity and deeper penetration of water
in stony soils compared to non-stony soils (Ravina and Magier, 1984; Kosmas et al.,
1993; Poesen and Lavee, 1994; van Wesemael et al., 1995, 1996). Stony soils are quite
common in semi-arid regions. They cover more than 60% of the land in the
Mediterranean (Poesen and Lavee, 1994). Frequent tillage of, in particular, orchards
leads to the development of a spatial pattern of stoniness on hillslopes (Poesen et al.,
1997). Denudation creates very stony soils on the convexities whereas kinetic sieving and
the downward transport of rock fragments by the tines of the chisel ploughs result in a
rock fragment mulch at the foot of the slopes.
The overall objective of this paper is to investigate the impact of different cultivation
practices with regard to water conservation in two soils which are important resources for
dryland farming where a large proportion of the soil remains bare: (i) a loamy sand soil
typical for a large proportion of the olive plantations in the coastal plains on the southern
fringes of the Mediterranean and (ii) a stony soil typical for many areas of rainfed
agriculture in the uplands which are increasingly used for tree crops such as almonds and
olives (Fig. 1). The specific research questions are:
1. Whether physical modifications of the upper few centimetres of a loamy sand, by
incorporating OME will create a stable and hydrophobic mulch and will have a
favourable effect on the soil water balance by reducing the evaporation.
H.J. Mellouli et al. / Agricultural Water Management 42 (2000) 355±369 357
Fig. 1. Two typical examples of farming systems where a large proportion of the soil is bare throughout the year.
(a) olive groves near Sfax (Tunisia) with a loamy sand soil and an annual precipitation of 190 mm and (b)
almond plantations in the Guadalentin basin (south-east Spain) on stony soils with an annual precipitation of
280 mm.
358 H.J. Mellouli et al. / Agricultural Water Management 42 (2000) 355±369
2. Whether modification of moisture retention capacity in loamy soils by the increase of
rock fragment content (resulting from intensive cultivation on hillslopes) will force
rain water to penetrate deeper in the profile where it should be saved from the
evaporation process.
2. Methods
Laboratory experiments were conducted in order to simulate two conditions that are
common in rainfed agriculture within semi-arid Mediterranean landscapes: loamy sand
soils and stony soils (Figs. 1 and 2).
2.1. Loamy sand soil
The soil used in this experiment is a loamy sand containing 7.9% clay, 16.2% silt and
75.9% sand, with an organic matter content of 1.7%. The OME used in the treatment of
the soil has the following characteristics: pH(H2O) � 5.0; electrical conductivity:
10 mS cmÿ1; density: 1.034 Mg mÿ3; dry matter content: 10.5%; mineral composition:
1%; organic matter: 9.5%. Three soil containers (length � width � height: 48.5 cm �48.5 cm � 50 cm), having a drainage system, were filled for 45 cm from the bottom with
the soil compacted to a bulk density of 1.4 Mg mÿ3. The uppermost 5 cm of these
containers were treated in the following way (Fig. 2):
1. Surface layer of untreated aggregates (0±8 mm) simulating the natural mulch situation
(control);
2. Similar to the `Control' except for a surface cover of straw. Straw was applied at the
optimal rate of 450 g mÿ2 (Bond and Willis, 1969; Smika, 1983);
3. Surface layer of aggregates (0±8 mm) impregnated with OME at a concentration of
1% of active material on dry soil basis as suggested by Mellouli (1996).
The soils were first exposed for 3 h to a rainfall intensity of 20 mm hÿ1 by means
of a rainfall simulator in order to investigate how the different surface conditions
affect water infiltration and redistribution (Kamphorst, 1987; Fig. 2). Then the surfaces
of the containers were covered during 3 days to allow redistribution of soil moisture.
The containers were subsequently subjected to evaporation for 46 days (corresponding
to a cumulative potential evaporation (Ep) of 460 mm) in a controlled environment
simulating the evaporation conditions of Mediterranean regions (Potential evaporation
rate ep � 9.9 � 1.8 mm dayÿ1, temperature fluctuated between 20±308C and relative
humidity between 50±85%). Moisture content was measured by means of Time Domain
Reflectometry (TDR). Two-rod probes were inserted horizontally in the containers
at 2.5 cm depth in the surface layer and at 3 cm interval from 6.5 cm depth. Drainage
was not observed and therefore evaporation was determined by change in soil
moisture storage. Soil moisture storage is defined by the analytical integral, from
the soil surface to a certain depth z, of the smooth moisture content profile, �(z). Since
no analytical integral is available, the storage was obtained by a numerical (trapezoidal)
method.
H.J. Mellouli et al. / Agricultural Water Management 42 (2000) 355±369 359
2.2. Stony soil
The treatments were selected to represent a common scenario induced by tillage of
shallow, stony soils: a stony control soil with a rock fragment content by volume (Rv) of
0.19 m3 mÿ3 that can evolve either to a non-stony soil with a rock fragment `mulch' on
top by kinetic sieving of the rock fragments in case of shallow tillage (Oostwoud
Fig. 2. Soil columns used for the experiments with loamy sand and stony soils.
360 H.J. Mellouli et al. / Agricultural Water Management 42 (2000) 355±369
Wijdenes et al., 1997; Poesen et al., 1997) or to a skeletal soil with an increased rock
fragment content (e.g. Rv � 0.35 m3 mÿ3) by breaking up the bedrock and bringing the
rock fragments to the surface (Fig. 2). Soil columns with a diameter of 19 cm and a
height of 20 cm were filled with an aggregated silt loam soil containing rock fragment
contents by volume (Rv) of 0.19 and 0.35 m3 mÿ3. In addition, a 5 cm thick layer of rock
fragments was applied to a similar silt loam soil (Fig. 2). Each treatment was duplicated.
The fine earth material consisted of well-structured, homogeneous silt loam (13% clay,
78% silt, 9% sand and 0.4% organic matter). Its dry aggregate size distribution was close
to that of a fine seedbed: 14% of the aggregates was smaller than 2 mm, 46% was
between 2 and 11.2 mm, 30% was between 11.2 and 31.5 mm and 10% was larger than
31.5 mm. Well-rounded river gravel (mainly flint and quartzites) were sieved and only the
fraction between 1.7 and 2.7 cm was used. The soil moisture content of the fine earth
equalled 0.18 kg kgÿ1.
Rainfall (30 mm) was applied to the columns with a rainfall simulator (intensity:
35 mm/h; Poesen et al., 1990). The columns were covered with a fine PVC mesh to
reduce degradation of the soil surface. The bottoms of the columns were perforated to
drain excess moisture. The columns were covered with a plastic sheet for 2 days in order
to drain excess moisture and to start the experiment with equilibrated moisture
conditions. Thereafter, the bottom of the columns was closed, the plastic cover was
removed, the columns were weighed and placed on a drying table with two fans blowing
across. After weighing the columns daily, they were replaced in a different position.
Maximum and minimum temperature and relative humidity were monitored on a daily
basis. Three similar columns filled with water were used to measure potential evaporation
(Ep). Potential evaporation varied around a mean of 7.7 mm dayÿ1 (5.6±10.1 mm dayÿ1)
due to variations in mean temperature 20.18C (16±288C) and relative humidity (Rh)
77.8% (58±95%)
3. Results and discussion
3.1. Loamy sand soil
The initial soil moisture distribution with depth was more or less equal in all containers
(Fig. 3(a)). At the start of the evaporation experiment (3 days after infiltration) the
amount of soil moisture in the three columns was equal although a more evenly
distributed moisture profile was obtained in the case of the OME treatment (Table 1;
Fig. 3(b)). This can be attributed to the fact that the aggregates, present at the soil surface,
have acquired hydrophobic and adhesive properties due to their treatment with OME.
Aggregate stability increases due to OME treatment while the aggregates themselves do
not take up moisture which leads to a higher infiltration rate and a deeper penetration of
rain water in the soil through the large cavities between the stable aggregates (Mellouli,
1996; Mellouli et al., 1998). These processes lead to a decrease in water retention in the
upper 20 cm of the soil profile and a more homogeneous distribution of the water content
within the entire profile.
H.J. Mellouli et al. / Agricultural Water Management 42 (2000) 355±369 361
Fig. 3. Soil moisture distribution profiles of the columns used for the experiment with loamy sand soil: (a)
Initial soil moisture content profiles, (b) soil moisture profiles after 60 mm rainfall and redistribution of soil
moisture for 3 days, and (c) soil moisture content profiles after 46 days of evaporation.
After 46 days, reduction of the cumulative evaporation of the OME treatment
compared to the control treatment amounted to 27.9%. The straw mulch did not seem to
be effective since the evaporation reduction was limited to 4.5% of the control (Table 1).
The cumulative evaporation of the different soil columns as a function of the cumulative
potential evaporation (Ep), the environmental evaporative demand, is plotted in Fig. 4(a).
Two phases in the evaporation process can be clearly distinguished: a first phase with a
more or less constant evaporation rate which is energy-limited and a second phase with a
decreasing evaporation rate which is soil moisture content limited (e.g. Black et al.,
1969). The OME topsoil is the most efficient in reducing evaporation losses and affects
both the first and second phase (Fig. 4(a)). Where straw was applied as a mulch, the
whole system reacted as a two-layered mulch. The application of straw is very efficient in
the first stage of evaporation but has no effect in the second phase and after 46 days the
effect is negligible since the length of the first phase is extended. The restriction of the
efficiency of mulches to the constant rate stage was already observed by Corey and
Kemper (1968) and Groenevelt et al. (1989).
The evaporation retarding effect of the two treatments on water conservation can be
observed from the potential evaporation required to lose about one-half of the total
cumulative evaporation of the control treatment (Ep 0.5). The Ep 0.5 equals 74 mm
(reached in 9 days) for the control experiment, 134 mm (reached in 15 days) for the straw
mulch and 232 mm (reached in 24 days) for the OME topsoil (Table 1; Fig. 4(a)).
The presence of a topsoil treated with OME has a beneficial effect on the soil moisture
content distribution after 46 days of evaporation (Fig. 3(c)). A homogeneous moisture
content along the entire soil column at a higher level than that of the control or straw
experiment were observed. Such a higher moisture content within the plough layer is
important for germination and the development of young plants. A homogeneous
distribution of water and its conservation over the upper 50 cm of the profile is also
beneficial for the development of some deep rooting crops.
Table 1Initial and final soil moisture storage and effects of the different treatments on cumulative evaporation at the endof the experiment (E46) and at the time when one-half of the final amount of cumulative evaporation is lost (E0.5
and d0.5)
Loamy sand soils Stony soils
Control Straw OME Control RF mulch Rv � 0.35
Initial moisture (mm) 115.6 117.4 117.8 48.4 45.2 39.7
Final moisture (mm) 57.9 62.3 76.2 5.2 2.8 3.4
E46 (mm) 57.7 55.1 41.6 43.2 42.4 36.3
�E46/E46control (%) 0 4.5 27.9 0 1.9 16.0
Ep 0.5 (mm) 74 134 232 34 62 55
d0.5 (days) 9 15 24 4 8 7
OME, Treatment with olive mill effluent.E46, Cumulative evaporation at the end of the experiment (46 days).�E46/E46control, Relative efficiency in E46 reduction compared to the control experiment.Ep 0.5 Potential evaporation at the time when half of the E46 is lost from the control experiment (30 mm for
loamy sand soils and 22 mm for stony soils).d0.5 Length of period to reach Ep 0.5.
H.J. Mellouli et al. / Agricultural Water Management 42 (2000) 355±369 363
3.2. Stony soil
The soil in the cylinders has compacted somewhat due to the impact of the raindrops
resulting in a soil depth of 15.9 cm for Rv 0.19 m3 mÿ3, 16.9 cm for Rv 0.35 m3 mÿ3 and
17.0 cm for the rock fragment mulch (Fig. 2). The difference in initial soil moisture
Fig. 4. Effect of different treatments on cumulative evaporation (E) as a function of evaporative demand (Ep) for
(a) the loamy sand soil and for (b) the stony soil.
364 H.J. Mellouli et al. / Agricultural Water Management 42 (2000) 355±369
content between the three treatments can be explained by the decrease in water holding
capacity due to an increased volume taken up by rock fragments and a higher macro-
porosity of soils containing rock fragments (Childs and Flint, 1990; van Wesemael et al.,
1995). The volume of the column taken up by rock fragments is similar for control
(Rv � 0.19 m3 mÿ3) and the rock fragment mulch. This is reflected in similar initial soil
moisture contents (Table 1). The column with the highest rock fragment content
(Rv � 0.35 m3 mÿ3) has the lowest water-holding capacity and therefore the lowest
moisture content at the start of the evaporation experiment.
After 46 days, the cumulative evaporation of the column with Rv � 0.35 m3 mÿ3 is
16% less than the control. The rock fragment mulch did not seem to be effective since the
evaporation reduction was limited to 1.9% of the control (Table 1). The column with a
higher rock fragment (Rv � 0.35 m3 mÿ3) shows a shorter constant rate stage evaporation
and a lower evaporation rate thereafter compared to the control experiment
(Rv � 0.19 m3 mÿ3; Fig. 4(b)). This decrease in the evaporation rate can be explained
by the reduced water-holding capacity of stony soils resulting in drainage of excess
moisture. The rock fragment mulch results in a lower evaporation rate during the constant
rate stage, although the constant rate stage lasts longer and the final cumulative
evaporation is similar to that of the control experiment.
Similar to the experiment with the loamy sand soils, the evaporation retarding effect of
the various treatments on water conservation can be observed from the potential
evaporation required to lose about half of the total cumulative evaporation of the control
treatment (Ep 0.5). The Ep 0.5 equals 34 mm (reached in 4 days) for the control experiment,
62 mm (reached in 8 days) for the rock fragment mulch and 55 mm (reached in 7 days)
for the Rv � 0.35 m3 mÿ3 (Table 1).
4. The role of cultivation practices in reducing evaporation losses
Most efforts to prevent moisture loss by evaporation from bare soils have been directed
towards modification of the soil surface conditions. The experiments in this paper have
shown that straw and gravel applied on top of the soil act as two-layered mulches.
However, their efficiency is short-lived, mainly restricted to the constant rate stage
of the evaporation process (e.g. Groenevelt et al., 1989). However, these mulches
tend to lengthen the constant rate stage and therefore on the longer term, cumulative
evaporation is similar to that from the control experiments. These mulches are therefore
best applied in situations where crops are very sensitive to drying during the initial
evaporation phase (e.g. germination of young plants or irrigated crops). The high level
of labour and investments involved in the application and maintenance of these mulches
as well as the scarcity of crop residues in semi-arid regions limits their extensive use
(Gale et al., 1993).
The experiments in this paper show that the modification of the hydrodynamic
properties of the plough layer has a longer lasting effect on water conservation compared
to surface mulching and is therefore more relevant to rainfed agriculture in semi-arid
regions. The impregnation of the loamy sand soil with hydrophobic substances such as
OME at a concentration of 1% of active material on a dry soil basis increases their
H.J. Mellouli et al. / Agricultural Water Management 42 (2000) 355±369 365
aggregate stability and reduces their unsaturated hydraulic conductivity (Mellouli, 1996).
Treating the upper 5 cm of soil columns with OME resulted after 46 days of evaporation
in a homogeneous distribution of water and its conservation over the upper 50 cm. In
particular, tree crop plantations could benefit from such conservation of water over a
prolonged period, since a large proportion of the surface is kept bare between the trees
and the water availability for each tree is stored in a large volume of soil (e.g. Ennabli,
1993). Disposal of OME in the sewage system poses considerable problems because of its
high organic matter and polyphenol content. However, trials in Spain and France with the
spreading of OME at a rate of not more than 100 m3 haÿ1 have proven that the
polyphenols can be broken down rapidly and that the high nutrient content of OME can
partly replace fertilisers (Fiestas Ros de Ursinos, 1986). The magnitude of the
evaporation reduction obtained with OME in the laboratory experiments suggests an
additional advantage of the spreading of OME in olive plantations.
On hillslopes with thin soils, rock fragments tend to move to the surface by tillage with
chisel ploughs (Oostwoud Wijdenes et al., 1997; Poesen et al., 1997). Frequent tillage on
slopes will lead to the creation of two types of soil profiles: (a) a rock fragment mulch, if
the tillage depth does not exceed the soil depth or (b) a profile with a high rock fragment
content if the tines of the chisel plough break up the bedrock (Fig. 2). Poesen et al. (1997)
demonstrated that these two types of soil profiles will generally occur in different
landscape positions dictated by slope characteristics that determine the intensity of tillage
erosion. The thin stony soils will occur on the convexities and the deeper soils with a rock
fragment mulch will prevail at the foot of the slopes. From the results of the laboratory
experiments, it is possible to predict the consequences of this spatial distribution of soil-
types for water conservation. The thin stony soils on the convexities will behave similarly
to the column with Rv � 0.35 m3 mÿ3 and will only retain small amounts of soil
moisture. Water conservation will be efficient even in the longer term by reduced
evaporation and redistribution of soil moisture downslope to the deeper soils on the
footslopes and in valley bottoms. These soils are covered by a rock fragment mulch that
to some extent protects soil moisture from evaporation losses at least during the initial
phase. Intensive cultivation with a tine-like implement (e.g. in south-east Spain two-to-
three times per year, Poesen et al., 1997) will thus cause a clear distinction between
runoff generating areas (shallow stony soils) where water losses are limited due to low
soil moisture retention and water receiving footslopes and valley bottoms with a tillage-
induced rock fragment mulch that protects against evaporation.
The results are in agreement with the calculations of Riou (1977), who indicated that
soil water retention characteristics are crucial in the availability of water for the plants in
semi-arid regions. Field experiments are required to assess the costs and benefits of
cultivation techniques such as the incorporation of OME or the increase of rock
fragments' content by frequent tillage.
5. Conclusions
The water-use efficiency of tree crops such as almonds and olives in semi-arid
environments depends to a large extend upon the evaporation rates from the bare soil in
366 H.J. Mellouli et al. / Agricultural Water Management 42 (2000) 355±369
between the widely spaced trees. The expansion of rainfed tree crops in Mediterranean
environments has stimulated this research into the impacts of the modification of surface
and subsurface characteristics on evaporation losses. The results of laboratory
experiments indicate that the effects of surface treatments (a straw mulch for a loamy
sand soil and a rock fragment mulch for a stony soil) are relatively short-lived, whereas
after 46 days (Ep � 10 mm/day) sub-surface treatments (impregnation with OME for the
loamy soil and increase of rock fragment content of the stony soil) significantly reduce
evaporation losses. These results indicate that cultivation techniques should aim to reduce
the water holding capacity in the topsoil. This can either be achieved by treatment with a
hydrophobic substance (OME), or it will result from frequent tillage of stony soils on
hillslopes whereby the stone content increases by breaking up the bedrock and kinetic
sieving. Field experiments into the impacts of the water-holding capacity on water-use
efficiency are in progress.
Acknowledgements
The research for this paper was financed by the Tunisian±Belgian bilateral co-
operation (ABOS) and the European Union. The first author wishes to thank the ABOS
for enabling him to carry out the research in Belgium. The second and third authors
participated in this research as part of the MEDALUS III (Mediterranean Desertification
and Land-Use) collaborative research project. MEDALUS III was funded by the
European Commission Environment and Climate Research programme (contract number:
ENV4-CT95-0118), and this support is gratefully acknowledged.
References
Al-Jaloud, A.A., 1988. Water use efficiency for crop production in Saudi sandy soils using synthetic polymers,
Ph.D thesis, State University, Ghent, Belgium, 144 pp.
Black, T.A., Gardner, W.R., Thurtell, G.W., 1969. The prediction of evaporation, drainage and soil water storage
for a bare soil. Soil Sci. Soc. Am. Proc. 33, 655±660.
Bond, J.J., Willis, W.O., 1969. Soil water evaporation: Surface residue rate and placement effects. Soil Sci. Soc.
Am. Proc. 33, 445±448.
Bousnina, H., 1993. Bonification du reÂgime hydrique des sols sableux des zones arides par un amendement
argileuxÐEtude dans une oasis du sud tunisien, Ph.D thesis, State University, Ghent, Belgium, 170 pp.
Childs, S.W., Flint, A.L., 1990. Physical properties of forest soils containing rock fragments. In: Gessel, S.P.,
Lacate, D.S., Weetman, G.F., Powers, R.F. (Eds.), Sustained Productivity of Forest Soils. Proc. 7th North
American Forest Soils Conference, University of British Columbia, Faculty of Forestry Publication,
Vancouver, BC, pp. 95±121.
Cooper, P.J.M., Gregory, P.J., Keatinge, J.D.H., Brown, S.C., 1987. Effect of fertilizer, variety and location on
barley production under rainfed conditions in northern Syria. 2. Soil water dynamics and crop water use.
Field Crops Res. 16, 67±84.
Corey, A.T. and Kemper, W.D., 1968. Conservation of soil water by gravel mulches. Hydrology Papers no. 30,
Colorado State University, Fort Collins, CO, 23 pp.
El Amami, S. and Haffani, M., 1974. L'action du paillage sur le bilan hydrique, la croissance des organes
et la maturite du fruit de l'oranger maltaise de Tunisie, Annales de l'INRAT, vol. 46, Fasc. 2, INRAT,
Tunisie, 24 pp.
H.J. Mellouli et al. / Agricultural Water Management 42 (2000) 355±369 367
Ennabli, N., 1993. Les ameÂnagements hydrauliques et hydro-agricoles en Tunisie, Institut National
Agronomique de Tunisie, 255 pp.
FAO, 1994. Water policies and agriculture. In: The State of Food and Agriculture, FAO Agriculture Series 25,
Rome, pp. 228±297.
Fiestas Ros de Ursinos, J.A., 1986. Posibilidades de utilizacion del alpechin como fertilizante, Int. Symp. Olive
by-products Valorization, Seville, Spain, March, 48 pp.
Floret, C., Pontanier, R., Rambal, S., 1982. Measurement and modelling of primary production and water use in
a south Tunisian steppe. J. Arid Environ. 5, 77±90.
FriaaÃ, A., Mensi, R., Kallel, A., 1986. Application of vegetation water in civil engineering ± Soil treatment with
vegetation water, Int. Symp. Olive by-products Valorization, Seville, Spain, March, pp. 263±273.
Hillel, D., Berliner, P., 1974. Waterproofing surface-zone soil aggregates for water conservation. Soil Sci. 118,
131±135.
Gale, W.J., McColl, R.W., Xie, Fang, 1993. Sandy fields traditional farming for water conservation in China. J.
Soil Water Cons. 48, 474±477.
Gicheru, P.T., 1994. Effects of residue mulch and tillage on soil moisture conservation. Soil Technol. 7, 209±220.
Groenevelt, P.H., van Straaten, P., Rasiah, V., Simpson, J., 1989. Modifications in evaporation parameters by
rock mulches. Soil Technol. 2, 279±285.
Jalota, S.K., Prihar, S.S., 1990a. Bare soil evaporation in relation to tillage, Adv. Soil Sci. 12, 187±216.
Jalota, S.K., Prihar, S.S., 1990b. Effect of straw mulch on evaporation reduction in relation to rates of mulching
and evaporativity, J. Indian Soc. Soil Sci. 38, 728±730.
Kamphorst, A., 1987. A small rainfall simulator for the determination of soil erodibility. Netherlands J. Agric.
Sci. 35, 407±415.
Kemper, W.D., Nicks, A.D., Corey, A.T., 1994. Accumulation of water in soils under gravel and sand mulches.
Soil Sci. Soc. Am. J. 58, 56±63.
Kosmas, C.S., Danalatos, N.G., Moustakas, N., Tsatiris, B., Kallianou, Ch, Yassoglou, N., 1993. The impacts of
parent material and landscape position on drought and biomass production of wheat under semi-arid
conditions. Soil Technol. 6, 337±349.
Mellouli, H. J., 1996. Modification physique d'un sable limoneux par les effluents (les margines) des moulins aÁ
huile d'olive. Incidence sur l'eÂvaporation, Ph.D thesis, State University, Ghent, Belgium, 255 pp.
Mellouli, H.J., Hartmann, R., Gabriels, D., Cornelis, W.M., 1998. The use of olive mill effluents (`margines') as
soil conditioner mulch to reduce evaporation losses. Soil Till. Res. 49, 85±91.
Modaihsh, S.A., Horton, R., Kirkham, D., 1985. Soil water evaporation suppression by sand mulches. Soil Sci.
139, 357±361.
Mwendera, E.J., 1992. Analysis of the effect of tillage on soil water conservation, Dissertationes de Agricultura
no. 221. Katholieke Universiteit Leuven, 247 pp.
Oostwoud Wijdenes, D., Poesen, J., Vandekerckhove, L., de Luna, E., 1997. Chiselling effects on the vertical
distribution of rock fragments in the tilled layer of a Mediterranean soil. Soil Till. Res. 44, 55±66.
Peters, D.B., 1960. Relative magnitude of evaporation and transpiration. Agron. J. 52, 536±538.
Poesen, J., Lavee, H., 1994. Rock fragments in topsoils: Significance and processes. Catena 23, 1±28.
Poesen, J., Ingelmo-Sanchez, F., MuÈcher, H., 1990. The hydrological response of soil surfaces to rainfall as
affected by rock fragment cover and position. Catena 19, 451±474.
Poesen, J., van Wesemael, B., Govers, G., Martinez-Fernandez, J., Desmet, P., Vandaele, K., Quine, T., Degraer,
G., 1997. Patterns of rock fragment cover generated by tillage erosion. Geomorphology 18, 183±197.
Ravina, I., Magier, J., 1984. Hydraulic conductivity and water retention of clay soils containing coarse
fragments. Soil Sci. Soc. Am. J. 48, 736±740.
Riou, C., 1977. Evaporation du sol nu et reÂpartition des pluies - Relations eÂtablies en Tunisie aÁ partir des
reÂsultats des cases lysimeÂtriques. Annales de l'INRAT, vol. 50, Fasc.4, INRAT, Tunisie, 24 pp. (Also
published in Cahiers de l'ORSTOM, SeÂr. Hydrol., vol. XIV, no. 3, 1977, pp. 285±295).
Smika, D.E., 1983. Soil water change as related to position of wheat straw mulch on the soil surface. Soil Sci.
Soc. Am. J. 47, 988±991.
Stewart, B.A., and Burnett E., 1987. Water conservation technology and rainfed and dryland agriculture. In:
Jorden, W.R. (Ed.), Water and Water Policy in World Food Supplies, A&M University Press, College
Station, TX, pp. 355±359.
368 H.J. Mellouli et al. / Agricultural Water Management 42 (2000) 355±369
Tanaka, D.L., 1985. Chemical and stubble-mulch fallow influences on seasonal soil water contents. Soil Sci.
Soc. Am. J. 49, 728±733.
Unger, P.W., Parker, J.J., 1976. Evaporation reduction from soil with wheat, sorghum and cotton residues. Soil
Sci. Soc. Am. J. 40, 938±942.
van Wesemael, B., Poesen, J., de Figueiredo, T., 1995. Effects of rock fragments on physical degradation of
cultivated soils by rainfall. Soil Till. Res. 33, 229±250.
van Wesemael, B., Poesen, J., Kosmas, C.S., Danalatos, N.G., Nachtergaele, J., 1996. Evaporation from
cultivated soils containing rock fragments. J. Hydrol. 182, 65±82.
Verplancke, H., Hartmann, R., Al-Jaloud, A.A., 1990. Effect of synthetic polymers on water use efficiency for
crop production on sandy soils in Saudi Arabia. Soil Technol. 3, 45±55.
Willis, W.O., Bond, J.J., 1971. Soil water evaporation: Reduction by simulated tillage. Soil Sci. Soc. Am. Proc.
35, 526±529.
H.J. Mellouli et al. / Agricultural Water Management 42 (2000) 355±369 369