Previous PageTable of ContentsNext Page

Wind and water erosion of non cultivated sandy soils in the Sahel: a case study
in Northern Burkina Faso, Africa

Rajot, J.L.1; O. Ribolzi2; O. Planchon3 and H. Karambiri4

  Keywords: wind erosion, water erosion, crust, grazing area, Sahel


In the Sahel, sandy soils are widespread and support not only most of pearl millet production, the major staple crop in the region, but also forage production for livestock. Parent sediments of these soils have an aeolian origin and hence are prone to wind erosion. However, the clay content, albeit very low, results in the formation of crusts during rainfall, thus leading to runoff and water erosion. Squall lines, major rainfall events of the rainy season, are usually preceded by intense wind. Wind and water erosion is thus closely associated both in time and in space, but they are rarely studied simultaneously. Erosion measurements were carried out during two years (2001, 2002) on a small catchment of grazing land (1.4 ha) at Katchari, Burkina Faso, typical of the Sahel area under 500 mm annual rainfall. Wind erosion occurred at the onset of the rainy season, when soil cover is the lowest, from May to 15th of July, before vegetation growth. Water erosion occurred throughout the rainy season, but some intense events produced most of the total annual erosion.

Wind caused the largest sediment fluxes leading to both erosion (up to 20 Mg ha-1 yr-1) and deposits (up to 30 Mg ha-1 yr-1) according to the area of the catchment. Water erosion is one order of magnitude lower than wind erosion, and is more intense where wind erosion is the highest. Thus the same area is eroded both by wind and water. Conversely, in areas where there are aeolian deposits, water erosion is low and these areas correspond to fertile islands where vegetation grows. At this study scale, there is no land degradation, but intense dynamics leading to a high spatial variability typical of the Sahelian environment. On this uncultivated area, the dynamics were similar to those recorded in other Sahelian cultivated millet fields.

1. Introduction

In the Sahel, sandy soils are dominant. In Niger, for example, they represent more than 80% of the agro-pastoral zone (Gavaud, 1977). They support most of the annual vegetation growth, pearl millet, the main staple crop, and forage for livestock. Consequently, they play a major role for the farmers who represent more than 80% of the Sahelian population (Thiombiano, 2000; Guengant and Banoin, 2003). On these soils, both wind and water erosion occurs, but these two forms of erosion are rarely studied simultaneously. Wind erosion data, scarcer than than that on water erosion, can be found in the literature for cultivated areas (fields and fallows). At the field scale,soil loss due to wind erosion is found to reach very high levels (more than 25 Mg ha-1 yr-1, Bielders et al. 2001). However, very few studies concern uncultivated areas where grazing is the only land use. Soil losses by water erosion in the Sahel seem to be lower (Collinet and Valentin, 1985), but, as measurements were not carried out on the same surface (same surface features and same surface area), they are difficult to compare. However a recent study (Visser, 2004) addressed estimation and comparison of wind and water erosions based on field work and modeling in the same area as the present work. Her results confirm those of previous studies, i.e. sediment and nutrient fluxes associated with wind are generally several orders of magnitude larger than those due to runoff. Her data were obtained at the erosion event scale on plots installed on two cultivated fields and one degraded surface with bare gravelly soil, but no measurement was performed in the grazing area. The objective of our work was to quantify wind and water erosion simultaneously in small grazing catchments with different surface features, over 2 seasonal cycles (2001 and 2002).

2. Materials and methods

2.1 Study area

The study area is located in the North of Burkina Faso (UTM30, WGS84, 809847 m East, 155093 m North), near Dori, 250 km Northeast of Ouagadougou (Figure 1). The climate is of the Sahelian type, with a long dry season and a short rainy season from May to September. Average annual rainfall recorded in Dori from 1925 to 1998 was 512 mm. The grazed areas of the village lands are located on a low longitudinal slope (about 1%). They show two main soil types: i) large areas of bare crusted clayey soil patched with ii) areas of sandy soil that have developed on aeolian sand deposits (microdunes less than 0.7 m thick) where annual vegetation, shrubs and trees grow (Ribolzi et al., 2000). Within this zone a small representative catchment (1.4 ha) was selected composed of five main soil surface types according to the classification of Casenave and Valentin (1992) (Figure 1): (1) bare erosion surfaces (ERO) accounted for 33.6% of the total catchment area, (2) pavement surfaces (G), which were also bare, covered 0.4% of the catchment area, (3) sedimentation surfaces (SED) covering the bottom of ponds and depressions, accounted for 1.2% of the catchment area, (4) runoff type surfaces (RUN) consisting mainly of laminated materials of various textures, represented 4.2% of the catchment area, and (5) the drying surfaces (DRY) which covered the leeward area of sandy microdunes represented 59.9% of the catchment area. Microdune soils accounted for 69% of the total catchment area and constituted almost exclusively the support for annual vegetation, shrubs and trees. The windward sides of microdunes accounted for 14.3% of the total catchment area; they were characterized by a steep fragmented ERO surface (i.e. crumbling of the laminated structure of plasmic and sandy layers called ERO/S) resulting from wind deflation. These observations served ground to the selection of three sub-areas homogeneous in terms of surface feature combinations (Figure 1).

Management of Tropical
Sandy Soils for Sustainable

Figure 1. The study area, map of soil surface features, experimental design and boundaries of the 2 catchments for which water erosion was calculated and of the 3 sub-areas for which wind budget was calculated

Measurements of both wind and water erosion were undertaken from the 1st of June 2001 to the end of September 2002.

2.2 Water erosion measurements

In this study, it was not possible to use classical water erosion plots because artificial boundaries act as windbreaks, causing significant aeolian sand deposits. To avoid such disturbances, water erosion was measured on two natural nested catchments. The upstream catchment (0.3 ha) corresponds to sub-area 3 on Figure 1.

Rainfall was monitored using three simple rain gauges and three rainfall recorders. The stream discharge of each catchment was measured using flow recorders. Suspended matter fluxes at the outlets were estimated from discrete 1-litre water samples collected throughout runoff events with a time interval varying from 2 to 5 minutes. Bedload was collected in sediment traps after each event. The totality of these materials was dried and then weighed.

2.3 Wind erosion measurements

Unlike cultivated land where field/fallow transitions have to be taken into account in wind erosion studies (Bielders et al., 2002), in Sahelian grazing land there is no clear boundary acting on wind erosion. Thus, the limits of the areas under study are determined by those of water erosion (main catchment and upstream catchment [sub-area 3] boundaries) and by surface feature patterns (sub-areas 1 and 2) (Figure 1).

Wind-blown sediment fluxes were obtained using 50 masts equipped with 3 BSNE sand catchers (Fryrear, 1986) of 0.05, 0.15 and 0.3 m in height. The masts were placed 1) on the sub-area boundaries approximately every 20-m and 2) along a transect in the western side of the catchment (Figure 1).

Wind-blown sediments caught in BSNE were collected if possible after each erosion event from June to September and each month from October 2001 to May 2002. The horizontal fluxes were calculated at each mast by integrating the sediment flux density profile between 0 and 0.4 m height. Wind speed and direction were measured using an automatic weather station. An acoustic saltation sensor (Saltiphone, Span and Van den Abeele, 1991) indicated the period during which the fluxes were significant. With this information, it was possible to estimate the mean direction of wind during each storm event, and to determine the upwind and downwind limits of catchments; the incoming and outgoing mass fluxes along the boundaries of the catchments were then calculated by linear interpolation of sediment mass fluxes measured at each mast. The mass budgets within the sub-areas 1, 2 and 3 were calculated by subtracting outgoing from incoming wind-blown sediments.

Along the E-W transect, the BSNE masts were setup at each major surface feature change. When erosive wind direction corresponded to the transect direction (95 ± 15º), which was assumed to correspond to the more intense wind erosion events, it was possible to compute a budget by subtracting downwind from upwind fluxes and dividing the result by the distance between the two measurement locations.

Measurements of wind-blown sediments along the E-W transect were taken during the rainy season 2001.

3. Results and discussion

3.1 Dynamics of wind-blown sediment flux

Sand catchers were collected 57 times during the measurement period. Some data correspond to mixed events occurring at close intervals, or during the monthly period of collection. From meteorological measurements, it is possible to estimate that sixty-eight wind erosion events occurred during the 16 months of measurement. For the common period of measurement (June to September), there were 33 and 21 events in 2001 and 2002, respectively. The flux density at a height of 30 cm, averaged for the 50 BSNE masts (called FD30 below), accumulated over 2001 and 2002 was 25 and 22 g cm-2 , respectively. The inter-annual variability was lower than that measured in Niger from 1996 to 1998 (Rajot, 2001).

Only three events occurred during the dry season from October to March. These latter events represented less than 0.3% of the cumulative FD30 for the whole period. The first event of 2002 occurred on the 6th of April and was linked to the first rainfall of the year. As in a cultivated field in Niger (Rajot, 2001), the Harmattan wind did not produce wind erosion in this grazing area of the Sahel.

For the whole period, eight events produced 53% of the cumulative FD30. Five of them occurred in June, two in July, and one in April. A few events at the onset of the rainy seasons produced most of the wind erosion.

Figure 2 shows the cumulative FD30 according to wind direction classes. 75% of the flux corresponded to wind blowing from the East to the Southeast (between 75º and 165º). This result corresponds closely with the local morphology of the microdune showing a higher slope on the windward side.

Management of Tropical
Sandy Soils for Sustainable

Figure 2. Percentage of flux density at 30 cm height averaged on the 50 BSNE masts and accumulated over the measurement period versus wind direction of events

All these observations correspond closely to other measurements suggesting a similar wind dynamic throughout the Sahel (Bielders et al., 2004).

3.2 Wind-blown sediment budget within sub-areas

From the 68 events producing wind-blown sediment flux during the study period, it was only possible to calculate the mass sediment budget for a subset of forty-two events owing to mixed events or important variations in the wind direction for the other twenty-six. The cumulative FD30 for these forty-two computed events represented 87% of the total FD30 blowing out on the catchments during the entire measurement period. There was no major event, i.e. with high sediment transport, among those events that were not considered. Thus, it will be assumed that the results obtained with these 42 events provide a good image of wind erosion on the catchments.

A Monte Carlo procedure was used for one major erosive event to estimate the standard deviation on budget result taking into account all the uncertainties affecting the computation, namely: the height of catchers, the surface of the opening, the fit of the theoretical equation for the calculation of horizontal flux, the position of the catcher on the area under study and the wind direction. The variation coefficients obtained ranged from 20% to 150% according to the sub-area and the differences between the sub-area budgets always remained highly significant.

Figure 3 shows the cumulative mass balances of aeolian sediment over the period of measurement for the three sub-areas. These areas behaved very differently: the budget is almost systematically positive for the upstream sub-area (#3) whereas it is systematically negative for the downstream one (#1), amounting to about +65 Mg.ha-1 and -35 Mg.ha-1, respectively, over the measurement period. Both erosion and deposition occurred in the centre sub-area (#2), but the budget remained positive over the measurement period (+27 Mg.ha-1). At the catchment scale, these different behaviours of the sub-areas led to an almost balanced budget until the beginning of June 2002 (+3 Mg.ha-1), and a really positive budget at the end of the measurement period (+16 Mg.ha-1).

A high level of wind-blown sediment deposits was also reported by Bielders et al. (2001) for fallow land in Niger which presented similar surface features as sub-area 3 (dry crust with annual and perennial vegetation). The deposits were ascribed to the high surface roughness of these areas. In Niger, the sources of wind-blown sediments were pearl millet fields (Bielders et al., 2001). In this study, net wind erosion occurred on complex natural areas where all the different surface features encountered in the catchment are represented (Figure 1). Thus the surface features from where wind-blown sediments originate are still unclear and need to be assessed.

3.3 Wind-blown sediment budget along the transect

The transect measurements taken across sub-areas 2 and 3 (Figure 1.) enabled a better description of the processes occurring in relation with soil surface features. Only five events required the wind direction to be computed from the transect, but two of them were the more intense events of the 2001 season. General trends appeared and can be summarized from the budget computed from the sum of these five events (Figure 4). First of all, the transect revealed the high spatial variability of wind erosion at the meter scale. There was no systematic behaviour on the 2 main surface features with regards to the budget: erosion may occur on the DRY surface and deposits may occur on the ERO surface. Nevertheless, the larger deposits occurred on the DRY surface whereas the more intense erosion occurred on the ERO surfaces covering windward slope of sandy microdune (ERO/S) or areas where such a surface was present (between 70 and 85 m), as well as on the RUN surface.

ERO/S surfaces are closely associated to DRY surfaces and develop on the same sandy soil. If one considers these 2 surfaces together (between 0 and 9 m and between 25 to 37 m) the sediment budget is negative i.e. the small patches of sandy soil are currently subject to net erosion as suggested by Casenave and Valentin (1989) during drought conditions.

Management of Tropical
Sandy Soils for Sustainable

Figure 3. Wind-blown sediment mass budget (Mg.ha-1) accumulated over the study period for the 3 sub-areas selected because of their surface feature distribution

Management of Tropical
Sandy Soils for Sustainable

Figure 4. Accumulated wind-blown sediment budget for the five events with easterly winds (parallel to the transect orientation) versus distance from the western border of catchment. The various types of surface features (see text for description) are indicated by different shades of gray on horizontal axis

The fact that net deposits may occur on ERO surfaces whereas sand deposits were not observed suggests that these sediments are mobilized by water erosion which often follows wind erosion in the Sahel (Visser et al., 2004). Similarly, the high susceptibility of RUN surfaces to wind erosion shows that water erosion produces sediments that are easily mobilized by wind erosion.

3.4 Dynamics of water erosion

The annual rainfall levels were 325 mm in 2001 and 345 mm in 2002. Both years showed a deficit compared to the mean annual level (512 mm for the reference period of 1925-1998). Rainfall generated 16 floods in 2001 and 13 in 2002. The number of water erosion events was less than half the number of wind erosion events. Although the amount of rain was lower in 2001, more heavy events were observed during this year: rainfall levels exceeded 25 mm for only two events in 2002 compared to four events in 2001 (Figure 5). For the whole catchment, water erosion was twice as high in 2001 as in 2002, but it was the reverse for the upstream catchment (Figure 5). Water erosion, unlike wind erosion, occurred throughout the rainy seasons and did not show a period of clearly higher intensity. As for wind erosion, some events were responsible for a large part of the annual erosion. In 2001, the four most important rainfall events (level >25 mm) were responsible for 60% of water erosion (Figure 5). Karambiri et al. (2003) already showed the similar behaviours in the same catchment for the period 1998-2000.

Management of Tropical
Sandy Soils for Sustainable

Figure 5. Rainfall depth and accumulated soil loss due to water erosion over the study period (rainy seasons 2001 and 2002) for the entire catchment, sub-area 3 (i.e. 100% DRY) and sub-areas 1 and 2 (between 20 and 80% DRY)

3.5 Water erosion within sub-areas

Cumulative soil losses by water for this period were estimated at 6.0, 2.5 and 7.3 Mg ha-1 yr-1 for the entire study area, the upstream zone (sub-area 3) and the downstream part of the catchment (sub-areas 1 and 2), respectively (Figure 5). Soil losses by water were lower upstream than downstream. These results conform closely to observations of Karambiri et al. (2003) in the same area (1998, 1999 and 2000 rainy seasons). They attributed the different behaviours of the two catchments to the soil surface characteristics: Drying surfaces, which are more permeable and supported herbaceous plants, covered the entire sub-area 3 and hence favoured infiltration rather than runoff. In contrast, the downstream zone was patched with impervious-prone erosion crust, which is more susceptible to erosion.

Particle size distribution of sediment exported varies according to the catchment. On the upstream sub-area, the exported sediments were composed mainly of sandy bedloads, while silty-clay suspended particles were dominant downstream (i.e. sub-areas 1 and 2). These results suggest that the clayey erosion crust could be a major source of sediment for water erosion; they correspond closely to the results reported by Karambiri et al. (2003) in the same study area.

3.6 Combined water and wind sediment budget

At the catchment scale, taking into account both wind and water erosion, and considering the whole study period, we estimated a positive cumulative budget of 10.2 Mg/ha. This result masks high spatial and interannual variabilities of the sediment budget (Table 1). There was a high positive budget on sub-area 3 due to wind sediment deposits (65 Mg/ha). Sandy sediments accumulated on sub-area 3 and were partially removed by water erosion (-2.5 Mg/ha) mainly in the form of bedload. Conversely, there was a negative budget on the downstream part of the catchment due to both wind (-0.4 Mg/ha) and water (-7.3 Mg/ha) erosion. For this sub-area, we observed a net wind erosion in 2001 (-6.4 Mg/ha), which was compensated by wind deposits in 2002 (6.0 Mg/ha). Such high spatial variability due to wind erosion in the Sahel was pointed out by Bielders et al. (2001) in a cultivated area. It also appears to be very high in grazing areas.

Table 1. Accumulated soil (Mg/ha) due to water and wind for the entire catchment, sub-area 3 and sub-areas 1 and 2

  water wind sum water wind sum water wind sum
Entire catchment  





Cumulative totals

Sub-area 3 -4.1  1.1 -3.0 -1.9 15.1 13.2 -6.0 16.2 10.2
Sub-areas 1 and 2 -0.9 25.7  24.8 -1.6 39.0 37.4 -2.5 64.7   62.2
  -5.2 -6.4 -11.6 -2.1 6.0 3.9 -7.3 -0.4 -7.7

In these areas, water erosion can locally reach the same order of magnitude as wind erosion and can move higher quantities of sediments (table 1). This result differs from that of Visser (2004) obtained for a cultivated field where wind erosion dominated.

4. Conclusions

For the first time in the Sahel, wind and water erosion was measured from the same surface areas of grazing lands, composed mainly of sandy soils. The main conclusions of this study can be summarized as follows:

  1. Annual wind erosion dynamics for this grazing area are typical of the Sahel and are the same as observed for a cultivated field in Niger.
  2. Wind erosion has a clearly oriented direction and is responsible for the asymmetric morphology of the microdune.
  3. Wind erosion events are more numerous than those of water erosion and at the smallest scale are more intense, moving higher quantities of sediments.
  4. There is a high spatial variability at the local scale with areas of net deposits, where vegetation grows and areas of net erosion where bare soils predominate.
  5. Both wind and water erosion is more intense downstream and appears to be related to the type and size of surface features.

These results revealed the difficulties in estimating land degradation in the Sahel that depends heavily on the study scale. They suggest a strong linkage at the scale of a few meters between sediment source areas where degradation occurs and sediment sink area where vegetation develops in “islands of fertility”. They emphasize the necessity of taking into account both wind and water erosion in order to assess the current land degradation in the Sahel.


The present work was supported by the Département des Ressources Vivantes of the Institute of Research for Development (IRD) and by the French National Research Program on Soil Erosion (PNSE). We acknowledge with gratitude the INERA (Institut National de l’Environnement et de Recherche Agricole, Burkina Faso) for providing access to the site.


Bielders, C.L., Rajot, J.L. and Amadou, M., 2002.Transport of soil and nutrients by wind in bush fallow land and traditionally-managed cultivated fields in the Sahel. Geoderma 109, 19-39, 2002.

Bielders, C.L., Rajot, J.L., Michels, K., 2004. L’érosion éolienne dans le Sahel Nigérien: influence des pratiques culturales actuelles et méthodes de lutte. Sécheresse, 15(1): 19-32.

Bielders, C.L., Vrieling, A., Rajot, J.L. and Skidmore, E., 2001. On-farm evaluation of Field-scale Soil Losses by Wind Erosion under Traditional Management in the Sahel. Proceedings of International Soil Erosion Symposium - ASAE - Honolulu 1-5 janvier 2001.

Casenave, A. & Valentin, C. 1989. Les états de surface de la zone sahélienne - Influence sur l'infiltration. Edition ORSTOM, collection Didactiques, Paris, 1989.

Casenave, A. and Valentin, C. 1992. A runoff capability classification system based on surface features criteria in semi-arid areas of West Africa. Journal of Hydrology, 130, 231-249.

Collinet, J. and Valentin, C, 1985. Evaluation of factors influencing water erosion in West Africa using rainfall simulation. Challenges in African hydrology and water resources IAHS Publ, 144: 451-461.

Fryrear D.W, 1986. A field dust sampler. Journal of Soil and Water Conservation 41: 117-120.

Gavaud, M., 1977 - Les grands traits de la pédogenèse au Niger méridional. Travaux et documents de l’ORSTOM, 76. ORSTOM, Paris, 102 pp.

Guengant, J.P. et Banoin, M. 2003. Dynamique des populations, disponibilités en terres et adaptation des régimes foncier: le Burkina Faso, une étude de cas, Institut National des Sciences des Sociétés (INSS), Institut National de la Statistique et de la Démographie (INSD), Drabo I., Ilboudo F., Tallet B. (coord.), Paris, CICRED, FAO, 2003, 114 p.

Karambiri, H. 2003. Crue et érosion hydrique au Sahel: étude et modélisation des flux d’eau et de matières sur un petit bassin versant pastoral au nord du Burkina Faso. Doctorat Thesis, Université Paris VI, 341 p.

Karambiri, H., Ribolzi, O., Delhoume, J.P., Ducloux, J., Coudrain-Ribstein, A. and Casenave, A., 2003. Importance of soil surface characteristics on water erosion in a small grazed Sahelian catchment. Hydrological Processes, 17(8): 1495-1507.

Rajot, J.L., 2001. Wind blown sediment mass budget of Sahelian village land units in Niger. Bull. Soc. Géol. France, 172(5): 523-531.

Ribolzi O., Auque L., Bariac T., Casenave A., Delhoume J.P, Gathelier R., Pot V, 2000. Ecoulements hypo-dermiques et transferts de solutés dans les placages éoliens du Sahel: Etude par traçage isotopique et chimique sous pluies simulées. C.R. Acad. Sci. Paris, 330: 53-60.

Spaan, W. and Van den Abeele, G.D., 1991. Windborne particle measurements with acoustic sensors. Soil technology, 4: 51-63.

Thiombiano, L. 2000. Etude de l’importance des facteurs édaphiques et pédopaysagiques dans le développement de la désertification en zone sahélienne du Burkina Faso. Thèse d’Etat de l’Université de Cocody, République de Côte d’Ivoire.

Visser, S.M., 2004. Modelling nutrient erosion by wind and water in northern Burkina Faso. Tropical Resource Management Papers No. 53 ISBN 90-6754-785-9.

Visser, S.M., Sterk, G., Ribolzi, O., 2004. Techniques for simultaneous quantification of wind and water erosion in semi-arid regions. Journal of Arid Environments, 59: 699-717.

1 IRD, UR 176 SOLUTIONS, Université Abdou Moumouni (UAM), IRD, BP 11416 Niamey , Niger,
2 IRD, UR 176 SOLUTIONS, IRD-IWMI-NAFRI, BP 06 Vientiane, Laos RDP
3 IRD, UR 176 SOLUTIONS, INAPG, LABO BIOMCO, 78850 Thiverval-Grignon, France
4 EIER, 03 BP 7023 Ouagadougou 03 Burkina Faso

Surface crusts of semi-arid sandy soils: types, functions and management

Valentin, C.1

Keywords: soil crusting, semi-arid region, runoff, wind erosion, water erosion


Soil crusting is increasingly recognized as a major form of soil degradation; it impedes seedling emergence, restricts infiltration and favours rill and gully erosion. Most research on soil crusts has concentrated on the loess belts in Europe and the United States where the soils are both highly productive and very prone to crusting. In comparison, studies on sandy soil crusts remained limited, mainly because these soils are considered as marginal, especially under semi-arid conditions. The objective of this paper is to review studies on surface crusts of semi-arid sandy soils in terms of forming processes, soil-crust types, hydrological and ecological functions, and agricultural management.

Most of these studies have been based on macro- and micro-morphological approaches associated with field rainfall simulation, these methods being more appropriate than laboratory structural stability tests. The analysis of the main soil and climatic factors showed that predicting crusting can rarely be based on a unique factor but requests a combination of factors (e.g. textural properties and organic carbon content).

Soil-crusts types (structural, erosional, depositional, biological) have been related to forming processes and hydrological and erosive properties. Identifying the prevailing role of surface crusts on runoff generation in the semi-arid regions has led to hydrologic models based more on surface conditions than on soil properties. Soil crusts have ambivalent impacts on soil erosion; they protect the soil surface from wind and interrill erosion but favour rill and gully erosion. Surface crusts are inherent to semi-arid sandy ecosystems and favour the concentration of resources, which is pivotal to pastoral and agricultural activities in the semi-arid regions. Due to soil crusts, runoff and runon are important components of the water balance. It is therefore essential for water and land management of the semi-arid sandy regions to account for the spatial and temporal distribution of soil crusts.


Although soil crusting has been mixed up for a long time with its causes (e.g. dispersion) or with its effects (e.g. surface compaction), it is increasingly recognized as a major form of soil degradation (e.g. Auzet et al., 2004). It impedes seedling emergence (e.g. Valenciano et al., 2004; Voortmana et al., 2004), restricts infiltration (e.g. Janeau et al., 2003) and favours rill and gully erosion (e.g., Valentin et al., 2005). Most research on soil crusts has concentrated on the loess belts in Europe (e.g. Bresson and Cadot, 1992) and the United States (e.g. Ruan et al., 2001) where the soils are both highly productive and prone to crusting. In comparison, studies on sandy soil crusts remained limited, because these soils are considered not only as marginal for crop production but also because most scientists have assumed that that are resistant to crusting (e.g. FAO, 1984). By contrast, these last two decades, a significant body of evidence has pointed to the high sensitivity of coarse-textured soils to surface crusting in Northern Senegal (Valentin, 1985), in Northern Niger (Valentin, 1991), in Southern Togo (Bielders et al., 1996), in Southern Niger (e.g. Rockström and Valentin, 1997; Esteves and Lapetite, 2003; Valentin et al., 2004), in Northern Burkina Faso (e.g. Karambiri et al., 2003; Ribolzi et al. 2003, 2005), in Northeastern Thailand (e.g. Hartmann et al., 2002) and in many other parts of the world as Northern China (Duan et al., 2003; Shirato et al., 2005), Zimbabwe (Burt et al., 2001) and Australia (e.g. Chartres, 1992; Isbell, 1995).

The objective of this paper is to review studies on surface crusts of semi-arid sandy soils in terms of factors, processes and soil-crust types, hydrological and ecological functions, and agricultural management.

Materials and methods

Because of the very thin structure of the crusts, generally less than 1 mm thick, many scientists characterized crust types based on micromorphological analysis (e.g. the review by Bresson and Valentin, 1994). A growing number of authors consider that instability tests are unsatisfactory to predict the soil sensibility to crusting (e.g. review by Valentin and Bresson, 1998). Field rainfall simulation has proved an invaluable tool to monitor crust forming processes and their impact on soil infiltrability (e. g. Valentin, 1991; Wace and Hignett, 1991; Casenave and Valentin, 1992; Bielders et al., 1996, Hartmann et al., 2002). Only few authors coupled rainfall micromorphology and rainfall simulation to monitor the various stages of crust forming in sandy soils (e.g., Valentin, 1991; Bielders and Baveye, 1995).

Crust types, processes and properties

The structural crusts are formed in situ while depositional crusts consist of sedimentary microlayers (Chen et al., 1980). Depending on the dominant forming process, two main types of structural crusts have been identified in sandy soils: the sieving crusts (Valentin and Bresson, 1992) and the packing crusts (Janeau et al., 2003). The sieving crusts are made of three well sorted microlayers: a top microlayer of loose coarse sand, a middle microlayer of fine sand with vesicular porosity and a lower dense microlayer of thin particles (Valentin, 1991; Bielders and Baveye, 1995). Packing crusts consist of sand grains or stable micro-aggregates tightly packed at the soil surface with very few macropores. Both types of crusts are influenced by kinetic energy of rainfall (e.g. Valentin, 1986). Sieving crusts develop on very sandy soils with very low organic matter content <1% while packing crusts develop on soils containing more organic matter and fine materials (but with silt <40%; Valentin, 2004).

They consist of a smooth, very dense, hard and thin (of the order of 0.1 mm) microlayer (Valentin and Bresson, 1992).

Structural crusts develop mainly upslope, erosion crusts mid-slope and sedimentation crusts down-slope (e. g., d’Herbès and Valentin, 1997). When these are not removed or destroyed by erosion, tillage or trampling, they tend to be colonized by cyanobacteria, algae, lichens, moses, microfungi, etc. As a result, several authors considered biological crusts as a typical category of surface crusts (e.g. Belnap and Lange, 2001; Eldridge and Leys, 2003) without considering the original physical crusts on which they develop. These underlying crusts greatly determine their hydrological behaviour (e.g., Bresson and Valentin, 1994; Malam Issa et al., 1999; Valentin, 2002) because abiotic (or ‘physical’) soil crusts differ not only in their main morphological characters (Table 1) but also in infiltrability (Table 2). Hence the interest of hydrologists for this classification to predict infiltration and runoff in the semi-arid regions from field observations and thus improve models (e.g., Casenave and Valentin, 1992; Tauer et Humborg, 1993; Bromley et al., 1997; Peugeot et al., 1997; Estevesi and Lapetite, 2003; Ndiaye et al., 2005). Because soil crusting can be identified by significant reflectance changes on the soil’s surface, soil crust-related properties such as water infiltration can be remotely sensed and mapped in semi-arid regions (e.g. d’Herbès and Valentin, 1997; Goldshleger et al., 2004).

Table 1. Main characteristics and properties of soil crusts in sandy soils
Sources: Casenave et Valentin, 1992; Valentin et Bresson, 1992, 2002; Janeau et al., 2003;V alentin, 2004)

Crust type

Thickness (mm)

Other characteristics

Forming process

Main factors

(mm h-1)#



Tightly packed sands or stable micro-aggregates

Compaction under rainfall impact

Silt <40%,
Soil organic matter >1%
Heavy storms

(10-54) n = 14



Vertical textural sorting
with coarse particles at the
top and thin particles at
the bottom. Vesicular porosity

Particle sieving under rainfall impact.
Air trapping
(hence vesicles)

Very sandy soils,
Soil organic matter >1%
Heavy storms

n = 31



Smooth, very dense and
hard microlayer made of
thin particles

Smoothening and erosion of structural crusts

structural crusts
Runoff or/and wind.

n = 20



Vertical textural sorting
with thin particles at the
top and coarse particles
at the bottom. Vesicular

Sedimentation is still water

Pre-existing structural crusts
Accumulation of water

n = 9

# (Range of values), n = number of samples

Soil crusts have ambivalent impacts on soil erosion; they protect the soil surface from wind and interrill erosion but favour rill and gully erosion. Loose sands of the sieving structural crusts are more easily removed by wind (e.g. Goossens, 2004; Hupy, 2004), and water than the more resistant erosion crusts (e.g. Valentin, 1994). The superimposition of a biotic crust tend to make the underlying crust more resistant to erosion (e.g., Malam Issa et al., 2001; Eldridge and Leys, 2003; Valentin et al., 2004; Neff et al., 2005). Runoff produced by soil crusts tends to concentrate and form gullies even in sandy soils (e.g., Peugeot et al., 1997; Leduc et al., 2001; Descloitres et al., 2003; Esteves and Lapetite, 2003). Sandy soils are therefore generally eroded not only by sheet but also by gully erosion, even for very gentle slope gradients (Valentin et al., 2005).

Implications for land and water management

In the semi-arid zones, farmers need to weed several times during the cropping season not only to remove weed covers (e.g., de Rouw and Rajot, 2004) and limit thus competition for nutrients and water resources, but also, and often primarily to destroy the surface crust and increase water intake into the soils (Valentin et al. 2004). However, surface crusts quickly re-establish as a result of the cumulative kinetic energy of the following rainfalls. Table 2 indicates that this critical cumulative rainfall necessary for the crust to form again after tillage tends to increase with mean annual rainfall. Although a part of rainfall is lost through runoff during crust formation, and despite its short-lived positive effect, tillage is therefore essential to increase the amount of water available for crops in semi-arid sandy soils (e.g. Graef and Stahr, 2000). Tillage explains why infiltration is greater in cropped soils than in pasture soils (e.g., Casenave and Valentin, 1992; Burt et al., 2001; de Rouw, 2005).

Because crusts in sandy soils result mainly from the direct impact of raindrops, mulching of crop residues or branches is generally recommended. Since available residues are primarily used for other purposes as livestock feed or roof thatching, mulch is generally restricted to patches covered with erosion crusts. In addition to the effect of sand and seeds accumulation, mulch attracts termites that perforate pre-existing crusts and increases infiltration by a mean factor 2-3 (e.g., Casenave and Valentin, 1992; Mando et al., 1996; Léonard and Rajot, 2001). Manure application and livestock corralling on the most severely crusted patches are also valuable alternatives to restore soil surface properties (e.g. Graef and Stahr, 2000; de Rouw and Rajot, 2004; Schlecht and Buerkert, 2004; de Rouw, 2005).

The proportion of fine particles in the top layer decreases during cultivation (e.g. Ambouta and al., 1996) and increases once the land is returned to fallow (e.g., Ambouta, 1994; Abubakar, 1996). This enrichment in fine particles is primarily due to atmospheric dust deposition (e.g. Orange and Gac, 1990; Valentin et al., 2004). These textural variations influences greatly crusting processes because no erosion crust could develop when clay + silt contents falls below 5% (Ambouta, 1994), which is often the case for cropped sandy soil. By contrast, in the fallow soils, clay + silt content can approach the optimal content of 10% (e.g., Poesen, 1986; Casenave and Valentin, 1989). In the desert regions of China, straw of wheat, rice, reeds, and other plants is half buried and the other half is exposed to fix dunes. This decreases the intensity of sand flux by as much as 99.5%. Where the sand is fixed, fine particles are accumulated and a hard soil crust is formed on the dune surface, improving the stability of the dune surface (Qiu et al., 2004). Once formed, the crusts, which are neither tilled not subjected to trampling, are gradually colonised and consolidated mosses and y green algae (e.g., Li et al., 2002) and protected from further water and wind erosion (Malam Issa et al., 2001; Valentin et al., 2004). As a result, Peugeot et al. (1997) observed in South­western Niger a much higher mean runoff coefficient (MRC) from a fallow (MRC = 23%) than from an adjacent the millet field (MRC = 5%). Most of the runoff concentrates into gullies. Since the bottom of these gullies are highly permeable (Peugeot et al., 1997; Esteves and Lapetite, 2003) a large proportion of the runoff in these gullies contribute to the water table recharge (Leduc et al., 2001).

Table 2. Mean annual rainfall (MAR, mm) and critical cumulative rainfall necessary to form a new crust after tillage (CCR, mm) in sandy soils of the arid and semi-arid zones of West Africa

Location Soil texture MAR (mm) CCR (mm) CCR/MAR (%) Source
Agadez, Northern Niger Sandy 150 25 17 Valentin, 1991
Banizoumbou, Southwester Niger Sandy 560 150 27 Röckstrom and Valentin, 1997
Bidi, Northern Burkina Faso Sandy 620 200 32 Lamachére, 1991
Thysse Khaymor, Central Senegal Sandy loam 660 160 24 Ndiaye et al., 2005

In Northern Senegal, most severe crusting was observed in sandy exclosures where the vegetation but also soil crusts were protected from tillage and trampling (e. g. Valentin, 1985). This process has also been observed during a long-term fencing experiment in a sandy desert of Turkmenistan where crusts extended while bush and herbaceous biomass decreased (Orlovsky et al., 2004). These authors concluded that in this environment, undergrazing, as well as overgrazing, should be considered as a desertification factor. A biological soil crust with high contents of soil organic carbon and fine particles (clay + silt) was also formed within 3 years on sand dunes in an exclosure in a semi-arid, sandy grassland located in Northern China (Shirato et al., 2005).

In semi-arid areas, using mean landscape characteristics leads to a considerable underestimation of infiltration-excess surface runoff (e.g. Giintner and Bronstert, 2004). Re-infiltration and lateral redistribution of surface runoff between adjacent landscape patches need therefore to be accounted of. For instance, the mosaic of runoff generating fallows or pastures and runon-fields can be part of an efficient water-harvesting system (Rockström and de Rouw, 1997; Rockström and Valentin, 1997; Rockström et al., 1999).

Although surface crusts can hamper seedling emergence, they have overall positive effects on plant production in semi-arid sandy soils. Where rainfall input is insufficient for a continuous plant cover, vegetation benefits from the concentration of water. Such concentration is made possible only because crusts generate runoff. Crusts, especially erosion crusts, are thus inherent to the semi-arid ecosystems where they regulate scarce resources (e.g., Valentin and d’Herbès, 1999).


  1. Crusts form on sandy soils with silt + clay content exceeding 5%. Most severe crusting is observed for a silt + clay content of 10%.
  2. Crusts develop on sandy soils where they hamper seedling resistance and often generate runoff despite the pervious nature of the underlying soil.
  3. Accounting soil crust types improve predicting hydrological models.
  4. Crusts and associated concentrated runoff explain gully erosion of sandy soils even on gentle slopes.
  5. In the semi-arid and sandy regions, crusts must not be regarded as detrimental for the ecosystems but rather as an essential component to concentrate the poor water resources.


Abubakar, S.M., 1996. Rehabilitation of degraded lands by means of fallowing in a semi-arid area of Northern Nigeria. Land Degradation and Development, 7(2), 133-144.

Ambouta, J.M.K., 1994. Etude des facteurs de formation d’une croûte d’érosion et de ses relations avec les propriétés internes d’un sol sableux fin au Sahel. PhD., Université Laval, Québec, 97 p.

Ambouta, J.M.K., Valentin, C. and Laverdière, M.R., 1996. Jachères et croûtes d’érosion au Sahel. Sécheresse, 7: 269-275.

Auzet, A.V., Poesen, J., and Valentin, C, 2004. Soil surface characteristics: Dynamics and impacts on soil erosion. Editorial. Earth Surface Processes and Landforms, 29, 1063-1064.

Belnap, J. and. Lange, O.L, ed., 2001. Biological soil crusts: structure, function and management. Springer Verlag, Berlin Heidelberg, 503 pp.

Bielders C.L. and Baveye, P., 1995. Vertical particle segregation in structural crusts: experimental observations and the role of shear strain. Geoderma, 67, 247-261.

Bielders, C.L., Baveye P., Wilding L., Drees R. and Valentin, C, 1996. Tillage-induced spatial distribution of surface crusts on a sandy Paleustult from Togo. Soil Science Society of America Journal, 60(3), 843-855.

Bresson, L.M. and Cadot, L., 1992. Illuviation and structural crust formation on loamy temperate soils. Soil Science Society of America Journal, 56: 1565-1570.

Bresson, L.M. and Valentin, C, 1994. Surface crust formation: contribution of micromorphology. In: Ringrose-Voase, A.J. and Humphreys, G.S., ed. Soil Micromorphology: Studies in Management and Genesis. Development in Soil Science, n°22, Elsevier, Amsterdam, 737-762.

Bromley, J., Brouwer, J., Barker, T., Gaze, S. and Valentin, C, 1997. The role of surface water redistribution in an area of patterned vegetation in Southwest Niger. Journal of Hydrology, 198: 1-29.

Burt, R., Wilson, M.A., Kanyanda, C.W., Spurway, J.K.R. and Metzler, J.D., 2001. Properties and effects of management on selected granitic soils in Zimbabwe. Geoderma 101, 119-141.

Casenave, A. and Valentin, C, 1989. Les états de surface de la zone sahélienne. Influence sur l'infiltration. ORSTOM, Collection “Didactiques”, Paris, France, 230 p.

Casenave, A. and Valentin, C, 1992. A runoff capability classification system based on surface features criteria in the arid and semi-arid areas of West Africa. Journal of Hydrology. 130: 231-249.

Chartres, C.J., 1992. Soil crusting in Australia. In: Sumner, M.E. and Stewart, B.A., ed., Soil crusting: chemical and physical processes. Advances in Soil Science, Lewis Publishers, Boca Rato, Fla., USA, 339-366.

Chen, Y., J., Tarchitzky, J., Brouwer, J., Morin, and Banin, A., 1980. Scanning electron microscope observations on soil crusts and their formation. Soil Science, 130, 49-55.

d’Herbès, J.M. and Valentin, C, 1997. Land surface conditions of the Niamey region (Niger): ecological and hydrological implications. Journal of Hydrology, 188-189, 18-42.

de Rouw, A., 2005. Long-term changes in topsoil and pearl millet production in Sahelian cropping systems. This issue.

de Rouw, A. and Rajot, J.L., 2004. Soil organic matter, surface crusting and erosion in Sahelian farming systems based on manuring or fallowing. Agriculture, Ecosystems, Environment, 104, 263-276.

Descloitres, M., Ribolzi, O. and Troquer, Y. le, 2003. Study of infiltration in a Sahelian gully erosion area using time lapse resistivity mapping. Catena, 53(3): 229-253.

Duan, Z.G., Wang, G., Xiao, H.L. and Dong, Z.B., 2003. Abiotic soil crust formation on dunes in an etremely arid environment: A 43-year sequential study. Arid Land Research and Management, 17(1), 43-54.

Eldridge D.J. and Leys, J.F., 2003. Exploring some relationships between biological soil crusts, soil aggregation and wind erosion. Journal of Arid Environments, 53: 457-466.

Esteves, M. and Lapetite, J.M., 2003. A multi-scale approach of runoff generation in a Sahelian gully catchment, a case study in Niger. Catena, 50, 255-271.

FAO, 1984. Guidelines: land evaluation for rainfed agriculture. FAO Soil Bull. 52. FAO, Rome.

Goldshleger, N, Ben Dor, E., Benyamini, Y and Agassi, M., 2004. Soil reflectance as a tool for assessing physical crust arrangement of four typical soils in Israel. Soil Science, 169(10), 677-687.

Goossens, D., 2004. Effect of soil crusting on the emission and transport of wind-eroded sediment: field measurements on loamy sandy soil, Geomorphology, 58, 145-160.

Graef, F. and Stahr, K., 2000. Incidence of soil surface crust types in semi-arid Niger. Soil and Tillage Research, 55(3-4), 213-218.

Gilntner, A. and Bronstert, A., 2004. Representation of landscape variability and lateral redistribution processes for large-scale hydrological modelling in semi-arid areas; Journal of Hydrology, 297, 136-161.

Hupy, J.P, 2004. Influence of vegetation cover and crust type on wind-blown sediment in a semi-arid climate. Journal of Arid Environments 58, 167-179.

Isbell, R.F., 1995. Sealing, crusting and hardsetting conditions in Australian soils. In: So, H.B., Smith, G.D., Raine, S.R., Schafer, B.M., Loch, R.J., ed., Australian Society of Soil Science, Brisbane, Queensland, Australia, 15-30.

Hartmann C, Poss, R., Janeau J.-L., Bourdon E., Lesturgez G. and Ratanaanupap S., 2002. Use of granular material theories to interpret structural changes in sandy soil. 17th World Congress of Soil Science, Bangkok, Thailand, 14-20-August 2002. Symposium no. 04, Paper no. 2108.

Janeau J.L., Bricquet J.P, Planchon O. and Valentin C, 2003. Soil crusting and infiltration on steep slopes in Northern Thailand. European Journal of Soil Science. 2003, 54(3), 543-554.

Karambiri, H., Ribolzi, O., Delhoume, J.P, Ducloux, J., Coudrain-Ribstein, A. and Casenave, A., 2003. Importance of soil surface characteristics on water erosion in a small grazed Sahelian catchment. Hydrological Processes, 17(8), 1495-1507.

Lamachère, J.M., 1991. Aptitude au ruissellement et à l'infiltration. In: Sivakumar, M.V.K. Wallace, J.S. Renard C. and Giroux C, ed., Soil water balance in the Sudano-Sahelian zone. IAHS Press, Institute of Hydrology, Wallingford, UK, 109-119.

Leduc, C, Favreau, G. and Schroeter, P., 2001. Long-term rise in a Sahelian water-table, the Continental Terminal in Southwest Niger. Journal of Hydrology, 243(1-2), 43-54.

Léonard, J., Rajot, J.L., 2001. Influence of termites on runoff and infiltration: quantification and analysis Geoderma, 104(1), 17-40.

Li, X.R., Wang, X.P., Li, T. and Zhang, J.G., 2002. Microbiotic soil crust and its effect on vegetation and habitat on artificially stabilized desert dunes in Tengger Desert, North China. Biology and Fertility of Soils, 35(3), 147-154.

Malam Issa, O., Le Bissonnais, Y., Défarge, C. and Trichet, J., 2001. Role of a cyanobacterial cover on structural stability of sandy soils in the Sahelian part of western Niger. Geoderma 101, 15-30.

Malam Issa, O., Trichet, J., Défarge, C., Couté, A. and Valentin, C., 1999. Morphology and microstructure of microbiotic soil crusts on a tiger bush sequence (Niger, Sahel). Catena, 37, 175-196.

Mando, A., Stroosnijder, L. and Brussaard, L., 1996. Effects of termites on infiltration into crusted soil. Geoderma, 74, 107-113.

Neff, J.C., Reynolds, R.L., Belnap, J. and Lamothe, P. , 2005. Multi-decadal impacts of grazing on soil physical and biogeochemical properties in Southeast Utah. Ecological Applications, 15(1), 87-95

Ndiaye B., Esteves, M., Vandervaere J.-P., Lapetite J.-M. and Vauclin M., 2005. Effect of rainfall and tillage direction on the evolution of surface crusts, soil hydraulic properties and runoff generation for a sandy loam soil. Journal of Hydrology 307, 294-311.

Orange, D. and Gac, J.Y., 1990. Bilan géochimique des apports atmosphériques en domaines sahélien et soudano-guinéen d’Afrique de l’Ouest (bassins supérieurs du Sénégal et de la Gambie). Géodynamique 5(1), 51-65.

Orlovsky, L., Douriko, M. and Babaev, A., 2004. Temporal dynamics and productivity of biogenic soil crusts in the central Karakum desert, Turkmenistan. Journal of Arid Environments, 56, 579-601.

Peugeot, C., Esteves, M., Galle, S., Rajot, J.-L. and Vandervaere, J.P., 1997. Runoff generation processes: results and analysis of field data collected at the east central supersite of the Hapex-Sahel experiment. Journal of Hydrology, 188-189, 179-202.

Poesen, J., 1986. Surface sealing on loose sediments: the role of texture, slope and position of stones in the top layer. In: Callebaut, F., Gabriels, D. and De Boodt, M., ed., Assessment of Soil Surface Sealing and Crusting. Flanders Centre for Soil Erosion and Soil Conservation, Ghent, Belgium, 354-362.

Qiu, G.Y., Lee, I.B., Shimizu, H., Gao, Y. and Ding, G.D., 2004. Principles of sand dune fixation with straw checkerboard technology and its effects on the environment. Journal of Arid Environments. 56(3), 449-464.

Ribolzi, O., Bariac, T., Casenave, A., Delhoume, J.P., Ducloux, J. and Valles V., 2003. Hydrochemistry of runoff and subsurface flow within Sahelian microdune. European Journal of Soil Science, 54, 1-12.

Ribolzi, O, Hermida, M., Karambiri, H., Delhoume, J.P. and Thiombiano, L., 2005. Wind processes improve water infiltration in Sahelian sandy rangeland. This issue.

Rockstrilm, J., Barron, J., Brouwer, J., Galle, S. and de Rouw, A., 1999. On farm spatial variability of soil and water in pearl millet cultivation. Soil Science Society of America Journal. 63, 1308-1319.

Rockström, J. and Rouw, A. de, 1997. Water, nutrients and slope position in on-farm pearl millet cultivation in the Sahel. Plant and Soil, 195, 311-327.

Rockström, J. and Valentin, C, 1997. Hillslope dynamics of on-farm generation of surface water flows: The case of rainfed cultivation of pearl millet on sandy soil in the Sahel. Agricultural Water Management, 33, 183-210.

Ruan, H.X., Ahuja, L.R., Green, T.R. and Benjamin, J.G., 2001. Residue cover and surface-sealing effects on infiltration: Numerical simulations for field applications. Soil Science Society of America Journal, 65(3): 853-861.

Shirato,Y., Zhang, T.H., Ohkuro, T., Fujiwara, H. and Taniyama, I., 2005. Changes in topographical features and soil properties after exclosure combined with sand-fixing measures in Horqin Sandy Land, Northern China. Soil Science and Plant Nutrition. 51(1), 61-68.

Schlecht E. and Buerkert A., 2004. Organic inputs and farmers’ management strategies in millet fields of Western Niger. Geoderma 121, 271-289.

Tauer, W. and Humborg, G., 1993. Irrigation par ruissellement au Sahel. Télédétection et systèmes d’information géographique pour déterminer des sites potentiels. Centre Technique de Coopération Agricole et Rurale, ACP-CEE, Ede, The Netherlands, 199 p.

Valenciano, J.B., Casquero, P.A. and Boto, J., 2004. Influence of sowing techniques and pesticide application on the emergence and the establishment of bean plants (Phaseolus vulgaris L.). Agronomie, 24(2),113-118.

Valentin, C, 1985. Effects of grazing and trampling on soil deterioration around recently drilled water-holes in the Sahelian zone. In: El Swaïfy, S.A., Moldenhauer, WL. and Lo A., ed., Soil erosion and conservation. Soil Conservation Society of America, Ankeny, 51-65.

Valentin, C. 1986. Effects of soil moisture and kinetic energy on the mechanical resistance of surface crusts. In: C. Callebaut, D., Gabriels, and M. de Boodt, ed., Assessment of Soil Surface Sealing and Crusting. University of Ghent, Belgium, 367-369.

Valentin, C, 1991. Surface crusting in two alluvial soils of Northern Niger. Geoderma. 48, 201-222.

Valentin, C, 1994. Sécheresse et érosion au Sahel. Sécheresse 5:191-198.

Valentin, C, 2002. Book review of J. Belnap and O.L. Lange (Eds), 2002. Biological soil crusts: structure, function and management Ecological Studies n°150, Springer Verlag, Berlin Heidelberg. 2001. 503 p., Geoderma, 1804, 229-301.

Valentin, C, 2004. Encroûtement superficiel, ruissellement et érosion dans les régions tropicales. Habilitation à Diriger des Recherches, Université Pierre et Marie Curie, Paris VI, Paris, France, 59 p.

Valentin, C. and Bresson, L.-M., 1992. Morphology, genesis and classification of soil crusts in loamy and sandy soils. Geoderma, 55, 225-245.

Valentin, C. and Bresson, L.-M., 1998. Soil crusting. in: Lal, R., Blum, W.E.H., Valentin, C, Stewart, B.A., ed. Methodology for assessment of soil degradation. Advances in Soil Science, Lewis Publishers, Boca Rato, Fla., USA, 89-107.

Valentin, C. and Bresson, L.-M., 2002. Predicting structural crust types from soil land climatic factors. European COST meeting, Milncheberg, Germany, 10-12 October 2002, oral communication, one abstract.

Valentin, C. and d’Herbès, J.M., 1999. Water harvesting along a climatic transect across tiger bush patterns in Niger: structure, slope and production. Catena, 37, 231-256.

Valentin, C, Poesen, J. and Yong Li, 2005. Gully erosion: impacts, factors and control. Catena, in press.

Valentin, C, Rajot, J.-L. and Mitja, D., 2004. Responses of soil crusting, runoff and erosion to fallowing in the sub-humid and semi-arid regions of West Africa. Agriculture, Ecosystems and Environment 104, 287-302.

Voortmana, R.L., Brouwer, J., and Albersen, P.J., 2004. Characterization of spatial soil variability and its effect on millet yield on Sudano-Sahelian coversands in S.W Niger. Geoderma 121, 65-82.

Wace, S.A. and Hignett, C.T., 1991. The effect of rainfall energy on tilled soils of different dispersion characteristics. Soil and Tillage Research, 20, 57-67.

1 Institut de Recherche pour le Développement, International Water Management Institute, National Agriculture and Forestry Research Institute, Vientiane, Laos,

An estimation of water retention properties in sandy soils of Southern Brazil

Bortoluzzi, E.C.1, 2; D. Rheinheimer3 and D. Tessier1

Keywords: water retention, water potential, electric charges, organic matter, clay minerals


Soil fabric and soil structure are sensitive to land use and management and their role in water retention properties have to be estimated in order to evaluate soil degradation or remediation. Undisturbed soil blocks were sampled on the campus of the Federal University of Santa Maria in Rio Grande do Sul State – Brazil, on a soil classified as Typic Paleudalf. Soil sampling was carried out on A1, A2, E and B horizons of five soil profiles representing different soil management practices. Their water retention and bulk density were measured from -1 to -107,000 kPa and their cation exchange capacity at soil pH (CECE) and at pH 7 (CEC7) were determined. The contribution of organic carbon and clay minerals to CEC was evaluated according to the model developed by Bortoluzzi et al. (2005). A multiple regression was developed to estimate water retention properties in relation with bulk density and the contribution of clay and organic matter at different water potentials. Bulk density explains part of the water content at high water potentials. In contrast, soil CEC7 provides a satisfactory explanation of the water retention properties at all water potentials studied. At high water potentials, the water retention reported to the organic carbon (OC) charge is considerably higher than that of clay charge, i.e. 60 g cmolc-1 and 30 g cmolc-1 for instance at -1 kPa. The amount of water per charge unit of carbon and of clay was similar at ~ -50 kPa, and the CEC can be used to predict water retention as a single parameter. At higher water potentials, taking into account the clay and organic carbon parameters improves the prediction of water retention properties. These results can be used to evaluate soil structure and therefore the impact of land use and management practices on soil physical properties.


In subtropical areas, the amount of soil water available for plant uptake can be limited and is determined by a number of factors, including climatic conditions but also pore space characteristics. Available water content varies widely depending on soil composition and especially soil texture and organic carbon content as reported by Bauer and Black (1992). Agricultural management can also exert a considerable influence on soil fabric and pore space (Wander and Bollero, 1999), and on soil strength. This evolution can be associated to a decrease in organic carbon content and the effects of various stresses including mechanical compaction (Assouline et al., 1997, Bruand and Tessier, 2000).

Although sand particles are mainly responsible for determining soil pore space, the fine fractions control part of the water retention properties. For instance, in French loamy soils, Bigorre et al. (2000) showed the additive effect of clay and organic carbon content on water retention properties at low water potentials (-1,600 and -107,000 kPa).

The objective of this paper is to determine the role of soil structure and electric charges associated with clays and organic carbon content on water retention properties in a wide range of water potentials for sandy soils in Southern Brazil, in relation to soil management practices.

Materials and methods

Soil location and characteristics

The study site was located on the Federal University Campus of Santa Maria in Rio Grande do Sul State, Brazil (29º42′52′′S and 53º42′10′′W) at an altitude of about 90 m. The parent material is a sandstone of the Santa Maria Formation. The soils are Typic Paleudalfs (USA, 1994). They are well drained and present a strong vertical clay gradient (Bortoluzzi, 2003).

Originally, the natural vegetation was a prairie. Five sites were used in this study: 1) a natural prairie (P); 2) a natural re-vegetation, i.e. a forest (F) cor­responding to little human activity, when the other three sites have different management techniques, i.e. 3) conventional tillage (CT) for the past 30 years; 4) 18 years of conventional tillage followed by 12 years no-tillage (NTCT); and, 5) five years of no-tillage after prairie (NTP).

Soil sampling

At each site a trench 2.5 m long, 1 m wide and 1.7 m deep was opened when the soil was close to its field capacity. Blocks of about 1,700 cm3 and soil cores of 70 cm3 were sampled from the A1a (0-7.5 cm) and a A1b (7.5-15.0 cm) horizon, in the A2 (~15-30 cm), E (~40-55 cm) and Bt (~80-95 cm) horizons.

Analytical techniques

Soil core water content (W) and bulk density measurements were determined at field conditions. From blocks, clods (~3-5 cm3) were prepared in the laboratory for water potentials ranging from -1 to -1,600 kPa. At low potential values, i.e. lower than -1,600 kPa and up to -107,000 kPa, the samples were equilibrated with controlled hygrometry in contact with saturated salt solutions. Water retention and bulk density measurements were carried out using the kerosene method according to Afnor (1996).

The pHH2O was measured in a 1:1 soil/solution proportion with a potentiometer. The Effective Cation Exchange capacity (CECE), measured at soil pH, was determined by exchange with cobalt hexamine trichloride at a concentration of 0.01666 M. The CEC at pH 7.0 (CEC7) was determined by exchange with 1 M NH4+-acetate, buffered at pH 7.0 (Afnor, 1996). The total organic carbon content (OC) was measured by a CN elemental analyser (Fison Carlo Erba).


The main soil characteristics are shown in Table 1. The pHH2O ranged from 3.9 to 6.1. The total organic carbon content (OC) ranged from 0.003 to 0.0137 kg kg-1. CEC values measured at soil pH (CECE) ranged from 14.3 to 74.3 mmolc kg-1. The values of CEC at pH 7.0 were higher and ranged from 27.3 to 105.0 mmolc kg-1.

Table 1. General soil properties and contribution of Clay and OC charges to CEC7




          Contribution of charges to CEC7c




ρa/Mg m-3 OCb/ kg kg-1 × 100 pHH2O/ unit CECE CEC7  Clay OC
    ------ /kg kg-1 × 100------       ------- /mmolc kg 1 --------------
 A1a 16.2 24.8 59.0 1.45 (0.10) 1.37 (0.09) 6.1 (0.09) 65.7 (6.4) 65.7 (5.0) 30.2


  A1b 16.2 24.8


1.58 (0.01) 0.94 (0.05) 5.9 (0.18) 54.0 (4.4) 58.3 (2.1) 32.3


  A2 81.4 22.4 59.0 1.51 (0.00) 0.82 (0.04) 5.0 (0.28) 45.0 (4.0) 52.3 (5.5) 32.5


  AB 76.5 21.1 55.4 1.46 (0.04) 0.69 (0.01) 4.5 (0.07) 43.0 (1.0) 65.7 (3.8) 45.5


  BT 46.8 18.4 34.8 1.33 (0.04) 0.56 (0.02) 4.4 (0.02) 74.3 (1.5) 105.0 (2.6) 89.0


F A 11.8 20.6 67.6 1.11 (0.20) 1.27 (0.07) 4.8 (0.52) 40.7 (5.7) 53.7 (1.5) 21.5


  A 11.8 20.6 67.6 1.51 (0.09) 0.69 (0.06) 4.5 (0.35) 23.0 (1.0) 36.3 (0.6) 20.0


  A 12.8 23.0 64.2 1.53 (0.05) 0.69 (0.02) 4.4 (0.16) 27.7 (3.2) 35.3 (4.0) 20.3


  E 9.5 37.4 53.1 1.61 (0.09) 0.30 (0.01) 4.6 (0.10) 14.3 (1.5) 27.3 (1.5) 18.1


  BT 42.0 17.2 40.8 1.54 (0.05) 0.52 (0.04) 4.6 (0.09) 69.0 (2.6) 91.0 (2.0) 77.9


CT Ap 11.0 24.3 64.7 1.70 (0.06) 0.77 (0.03) 5.3 (0.27) 38.0 (4.0) 41.7 (0.6) 21.0


  Ap 11.0 24.3 64.7 1.72 (0.01) 0.70 (0.04) 5.8 (0.17) 39.3 (5.8) 40.0 (1.0) 21.3


   A2 13.6 26.3 60.1 1.65 (0.01) 0.59 (0.04) 5.4 (0.04) 39.0 (2.0) 38.3 (5.8) 24.2


   E 9.5 37.4 53.1 1.61 (0.09) 0.30 (0.01) 4.6 (0.10) 14.3 (1.5) 27.3 (1.5) 18.1


   BT 42.0 17.2 40.8 1.54 (0.05) 0.52 (0.04) 4.6 (0.09) 69.0 (2.6) 91.0 (2.0) 77.9


NTP   Ap 11.9 24.7 63.4 1.34 (0.06) 1.04 (0.17) 4.9 (0.16) 38.3 (7.5) 47.7 (7.2) 21.1


   Ap 11.9 24.7 63.4 1.59 (0.06) 0.78 (0.09) 4.4 (0.12) 30.3 (2.3) 45.3 (1.2) 23.7


   A2 13.6 26.2 60.2 1.48 (0.05) 0.71 (0.02) 4.4 (0.14) 32.7 (1.5) 42.3 (4.2) 24.6


   E 16.1 24.6 59.3 1.57 (0.03) 0.46 (0.01) 4.3 (0.03) 25.3 (1.5) 36.7 (2.1) 28.1


   BT 46.9 18.4 34.7 1.30 (0.03) 0.53 (0.03) 4.0 (0.08) 72.3 (4.5) 93.3 (0.6) 80.4


NTCT   Ap 13.0 20.0 67.0 1.58 (0.06) 0.79 (0.03) 3.9 (0.02) 25.3 (0.6) 43.3 (1.5) 22.4


   Ap 13.0 20.0 67.0 1.72 (0.00) 0.64 (0.03) 4.1 (0.08) 25.3 (2.1) 41.0 (1.0) 24.4


   A2   14.9 23.6


1.61 (0.08) 0.61 (0.02) 4.4 (0.17) 28.7 (2.3) 37.3 (5.9) 23.9
   E 12.6 14.3 73.1 1.58 (0.04) 0.40 (0.03) 4.4 (0.04) 17.3 (1.5) 34.7 (3.1) 26.4


   BT 30.9 17.9 51.2 1.55 (0.04) 0.48 (0.01) 4.5 (0.03) 48.7 (14.3) 76.0 (12.0) 52.7


a Bulk density, b Organic Carbon. c Clay and OC charge contribution obtained by solving of the equation 1, where the relative contribution (%) from each constituent was applied in the measured CEC7 values. The data in parenthesis are standard deviations of means with three replicates. 

Physical properties

The bulk density (ρ) varied from 1.11 Mg m-3 to 1.72 Mg m-3 (Table 1). The greatest bulk density variations were observed close to the surface at 0-15 cm depth. Below, the soil can be divided into two groups: (i) under prairie or under no-tillage after prairie where the bulk density ranged from 1.30 to 1.33 Mg m-3 and (ii) under forest or under conventional tillage and under no-tillage after conventional tillage where bulk density was close to 1.55 Mg m-3.

The results also showed that the relationship between water content and bulk density at conditions close to field capacity was poorly correlated for cores (R2 = 0.22) (Table 2). On clods, this relationship between water content and bulk density at -1 kPa was improved (R2 = 0.84). Furthermore, the quality of the prediction decreased from high water potentials to low water potentials (0.84 <R2 <0.22).

In order to distinguish the contribution of soil structure and fabric, i.e. the arrangement of soil particles, we used the CEC7 to predict soil water content (Table 2). There is a adequate relationship between water content and CEC7 for water potential varying between -3.2 to -1,600 kPa with determination coefficients higher than R2 = 0.86. The determination coefficients are lower on cores at field conditions (R2 = 0.72) and for clods at -1 kPa (R2 = 0.66) and -107,000 kPa (R2 = 0.69). These relationships were improved when the water was estimated by the CEC reported to the soil volume (cmolc dm-3) and not by mass (cmolc kg-1) (Table 2).

Relevance of soil charges on water retention properties

We used a multiple linear regression equation to evaluate the contribution of clay and organic carbon content to water retention properties (Table 3). For organic carbon, water retention reached 16.34 and 1.58 g g-1 at -1 and -1,600 kPa, respectively. This is in agreement with the water retained per g of carbon reported by Emerson (1995). The clay contribution is considerably lower than organic carbon since the water content at -1 kPa is 0.49 g g-1. This water content is quite stable from -1 kPa to -1,600 kPa (0.49 to 0.35 g g-1). At high water potentials the water content per negative electric charge unit of organic carbon was higher than that of the clay (61.3 and 28.9 at -1 kPa;46.3 and 28.2 at -3.2 kPa; 34.4 and 27.4 at -10 kPa), while at low potential range, it was the contrary (19.5 and 24.8 at -100 kPa; 11.5 and 22.3 at -320 kPa; 4.3 and 21.3 at -1,600 kPa). At -107,000 kPa the organic carbon and clay contribution, per electric charge unit, in water content was lower and similar (3.5 and 2.0, respectively).

Table 2. Linear regression equations between bulk density and cation exchange capacity and water content at different water potentials

Water potential
in kPa

Multiple linear regression equations



Field conditionsa W = 0.451 - 0.185 ρ 0.22** 25
-1 W = 0.829 - 0.394 ρ 0.84*  25
-3.2 W = 0.809 - 0.397 ρ 0.76* 25
-10 W = 0.726 - 0.359 ρ 0.56* 25
-100 W = 0.577 - 0.288 ρ 0.46* 25
-320 W = 0.646 - 0.339 ρ 0.42* 25
-1,600 W = 0.380 - 0.188 ρ 0.24** 25
-107,000 W = 0.049 - 0.022 ρ 0.22** 25

Field conditons

W = 0.043 + 0.024 CEC7 mass

0.72* 25

W = 0.107 + 0.021 CEC7 mass

0.66* 25

W = 0.072 + 0.022 CEC7 mass

0.86* 25

W = 0.044 + 0.023 CEC7 mass

0.87* 25

W = 0.007 + 0.022 CEC7 mass

0.87* 25

W = -0.007 + 0.021 CEC7 mass

0.86* 25

W = -0.035 + 0.022 CEC7 mass

0.89* 25

W = 0.004 + 0.002 CEC7 mass

0.69* 25

Field conditons

W = 0.061 + 0.030 CEC7 vol

0.74* 25

W = 0.115 + 0.030 CEC7 vol

0.79* 25

W = 0.087 + 0.030 CEC7 vol

0.92* 25

W = 0.059 + 0.031 CEC7 vol

0.91* 25

W = 0.023 + 0.030 CEC7 vol

0.90* 25

W = 0.006 + 0.028 CEC7 vol

0.87* 25

W = -0.020 + 0.030 CEC7 vol

0.89* 25

W = 0.004 + 0.003 CEC7 vol

0.77* 25

a In field conditions, samples are soil cores. At given water potential values, the samples are clods (3-5 cm3); b n = number of observations used in regression equations; ρ is bulk density (Mg m-3); CEC7 is cation exchange capacity measured at pH 7 in mass (cmolc kg-1) and in volume (cmolc dm-3); W is water content (g g-1). * Significant at P <0.01, ** Significant at P <0.05 and ns = no significant.

Discussion and Conclusion

CEC value and modelling

The Effective Cation Exchange Capacity (CECE) and the CEC at pH 7.0 (CEC7) were used to estimate the electric charges of the soil constituents according to the following equation (Bortoluzzi et al., 2005):

Management of Tropical
Sandy Soils for Sustainable

Table 3. Multiple linear regression equations between organic carbon and clay and water content at different water potentials

Water potential Multiple linear in kPa

Multiple linear
regression equations




W = 0.49 Clay + 16.34 OC

0.62* 25

W = 0.48 Clay + 12.46 OC

0.77* 25

W = 0.46 Clay + 9.41 OC

0.82* 25

W = 0.42 Clay + 5.45 OC

0.88* 25

W = 0.37 Clay + 3.58 OC

0.83* 25

W = 0.35 Clay + 1.58 OC

0.90* 25

W = 0.003 Clay + 0.93 OC

0.73* 25

W = 28.9 CVP Clay + 61.3 CV carbon

0.63* 25

W = 28.9 CVP Clay + 61.3 CV carbon

0.63* 25

W = 28.2 CVP Clay + 46.3 CV carbon

0.82* 25

W = 27.4 CVP Clay + 34.4 CV carbon

0.87* 25

W = 24.8 CVP Clay + 19.5 CV carbon

0.90* 25

W = 22.3 CVP Clay + 11.5 CV carbon

0.93* 25

W = 21.3 CVP Clay + 4.3 CV carbon

0.96* 25

W = 02.0 CVP Clay + 3.5 CV carbon

0.73* 25

a n = number of observations used in regression equations; W is water content (g g-1); Clay and OC are clay and organic carbon content; CV carbon and CVP clay are the surface charges due to carbon and clay soil constituents. * Significant at P <0.01, ** Significant at P <0.05 and ns = no significant.

where a is the permanent charge density kg-1 clay, mmolc kg-1, b is the pH-dependent charge density per unit of pH and kg-1 clay, mmolc kg-1 pH-1, c is the pH value at ZPC (zero point of charge) of the clay, expressed in units of pH, d is the pH-dependent charge density per unit of pH and kg-1 OC, mmolc kg-1 pH-1, e is the pH value at ZPC (zero point of charge) of the organic carbon, expressed in units of pH and pH is the pHH2O of the soil sample. Clay and OC are expressed in kg kg-1.

The equation 1 was simplified for solving the clay and OC relative contribution to the CEC7.0. Thus, the charges due to clay and OC are shown in the Table 1.

Contribution of structure and fabric to water retention properties

The results show that we can estimate the contribution of the CEC due to organic carbon and clay fraction on water retention properties. In Figure 1, the CEC does not completely explain the water retention at high water potential values, for instance at -10 kPa in sandy horizons, while in Bt horizons the estimation was better. This means that part of the water retention properties is due to grain packing voids in sandy material. On the contrary, at low water potential, namely -1,600 kPa, there is very good agreement between estimated water and measured water content in all horizons.

Management of Tropical
Sandy Soils for Sustainable

Figure 1. Contribution of clay (CPV clay) and organic carbon (CV carbon) charges to water content in two soils subjected to contrasting soil management practices, i.e. conventional tillage, CT and no-tillage after prairie, NTP, where A and B are the water potential at -10 kPa and C and D at -1.600 kPa

Clay and organic matter fractions have very different properties, especially at high water potential values of organic carbon. One of the possible explanations is the specific fabric of organic matter at high water content, whereas its hydration properties at low water potentials appeared to be related to its molecular structure (Feller et al., 1996). In contrast, the clay fabric is mainly an assemblage of kaolinite crystals or domains, mainly face-to-face (Ben Rhaïem et al., 1987). In this case, the clay fabric does not vary to a large extent in the high-water-potential range.

It is also interesting to notice that the water content per negative electrical charge unit of clay and organic carbon was similar at a specific water potential value close to -50 kPa. This value is obtained by the intersection between clay and organic carbon water retention curves. This means that the CEC can be used to predict the soil water content alone on the basis of this specific value. This result confirms other data previously reported by Bigorre et al. (2000). Furthermore, because the water content of the clay was rather constant from -1 kPa to -1,600 kPa, this means that water associated with clay fraction is mainly residual water. This would also suggest that organic carbon is the major component of soil water uptake and release at higher potential.

For our soils, the value of CEC can be used to predict water retention properties due to clay and organic carbon, and water potential. The water retention of clays was very constant within a large range of water potentials and can be considered as residual water content since little water was released or taken up over a large water potential range. This may be due to the specific structure of sandy soils where the fine fractions filling the sand matrix can fully express their water retention properties as a result of the available volume for hydration and swelling.

There is a specific water potential of -50 kPa, where the water content related to the charge unit of the clay and organic carbon. At this water potential in the soils under study, the CEC can be used to predict water retention as a single parameter. For other water potential values, a better prediction was obtained by taking the contribution of clay and organic carbon to the CEC and its contribution to water content.

It is shown that land use and soil management practices orient water retention properties. Close to -10 kPa, the prediction must take the structure of the soil into consideration and is thus associated with changes in packing voids and soil structure due to practices.


Afnor 1996. Qualité des sols: recueil des normes françaises. 3rd edn. Paris la Défense, France.

Assouline, S., Tavares-Filho, J. and Tessier, D. 1997. Effect of compaction on soil physical and hydraulic properties: experimental results and modelling. Soil Science Society of America Journal, 61, 390-398.

Bauer, A. and Black, A.L. 1992. Organic carbon effects on available water capacity of three soil textural groups. Soil Science Society of America Journal, 56, 248-254.

Ben Rhaïem, H., Pons, C.H. and Tessier, D. 1987. Factors affecting the microstructure of smectites. Role of cation and history of applied stresses. In: Special Issue Proc. Int. Clay Conf. Denver, Clays and Clay minerals, The Clay Minerals Society, 292-297.

Bigorre, F., Tessier, D. and Pédro, G. 2000. Contribution des argiles et des matières organiques à la rétention de l’eau dans les sols. Signification et rôle fondamental de la capacité d’échange en cations. Academie de Science de Paris. Earth and Planetary Science, 330, 245-250.

Bortoluzzi, E.C., Tessier, D., Rheinheimer, D.S., and Julien, J.L. 2005. The cation exchange capacity of a sandy soil in Southern Brazil: an estimation of permanent and pH-dependent charges. European Journal of Soil Science. (in press).

Bruand, A. and Tessier, D. 2000. Water retention properties of the clay in soils developed on clayey sediments: significance of parent material and soil history. European Journal of Soil Science. 51, 679-688.

Emerson, W.W. 1995. Water retention, organic carbon and soil texture. Australian Journal Soil Research, 33, 241-251.

Feller, C, Albrecht, A. and Tessier, D. 1996. Aggregation and organic matter storage in kaolinitic and smectitic tropical soils; Structure and organic matter storage in agricultural soils, Advances in Soil Science, CRC Lewis pub. 309-359.

USA. Department of Agriculture. Soil Survey Division. Soil Conservation Service. Soil Survey Staff. Keys to Soil Taxonomy. 6. ed. Washington, 1994. 306 p.

Wander, M.M. and Bollero, G.A. 1999. Soil quality assessment of tillage impacts in Illinois. Soil Science Society of America Journal, 63, 961-971.

1 INRA – Soil Science Unit, 78026 Versailles, France. E-mail:
2 Fundation University of Passo Fundo, 99001-970 Passo Fundo, Brazil. E-mail:
3 Federal University of Santa Maria, Santa Maria, Brazil.

Previous PageTop of PageNext Page