SEARCH

SEARCH BY CITATION

Keywords:

  • Mediterranean soils;
  • rock fragment cover;
  • run-off;
  • simulated rainfall;
  • soil loss

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Rock fragments are a key factor for determining erosion rates, particularly in arid and semiarid environments where vegetation cover is very low. However, the effect of rock fragments in non-cultivated bare soils is still not well understood. Currently, there is a need for quantitative information on the effects of rock fragments on hydrological soil processes, in order to improve soil erosion models. The main objective of the present research was to study the influence of rock fragment cover on run-off and interrill soil erosion under simulated rainfall in Mediterranean bare soils in south-western Spain. Thirty-six rainfall simulation experiments were carried out at an intensity of 26.8 mm h−1 over 60 min under three different classes of rock fragment cover (<50%, 50–60% and >60%). Ponding and run-off flow were delayed in soils with high rock fragment cover. In addition, sediment yield and soil erosion rates were higher in soils with a low rock fragment cover. The relationship between soil loss rate and rock fragment cover was described by an exponential function. After this first set of experiments, rock fragments were removed from sites with the highest cover (>60%) and the rainfall simulation experiments were repeated. The steady-state run-off rate and soil loss increased significantly, showing that run-off and soil erosion were partly conditioned by rock fragment cover. These results have significant implications for erosion modelling and soil conservation practices in areas with the same climate and soil characteristics.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Rainfall-induced erosion in the Mediterranean area is a major problem. Soil surface characteristics such as roughness, structure, vegetation and rock fragments (mineral particles ≥2 mm in diameter) have an important influence on infiltration rates, run-off generation and erosion (Auzet et al. 1995; Papy and Douyer 1991). Quantitative information on the effects of rock fragment cover on hydrological and erosional processes is necessary to model soil erosion and to predict the effects of land-use changes (Cerdà 2001; Poesen and Lavee 1994). Soils containing rock fragments constitute approximately 30% of the soil surface in Western Europe and 60% in the Mediterranean area (Poesen and Lavee 1994), and are widely distributed in Spain (Ingelmo et al. 1994). In recent years, many studies have examined soils containing abundant rock fragments (Abrahams and Parsons 1991, 1994; Bunte and Poesen 1994; Cerdà 2001; Cousin et al. 2003; Martínez-Zavala and Jordán 2008; Poesen and Bunte 1996; Poesen et al. 1998; Valentin 1994; van Wesemael et al. 1996), but more research is needed (Martínez-Zavala and Jordán 2008). In addition, most recent studies deal with the effect of rock fragment cover on the hydrological and erosional response of agricultural soils, but non-cultivated soils have not been taken into account (Cerdà 2001).

The energy of raindrops reaching the soil surface is a critical factor for splash detachment of soil particles (Flanagan and Nearing 1995; Wischmeier and Smith 1978), and detached particles contribute to an increase in sediment yield during the initial stage of run-off generation. When the rainfall intensity exceeds the water infiltration capacity run-off occurs and overland sheet flow is generated (Parsons et al. 1996). Poesen et al. (1994) suggested that rock fragments in an eroding environment have three main effects: protection against raindrop impact and flow detachment, reduction of the physical degradation of the eroding surface and retardation of overland flow velocity. Laboratory experiments have demonstrated that a decrease in rock fragment cover leads to a decrease in infiltration because part of the porosity has disappeared and the soil surface is not prevented from sealing (Brakensiek and Rawls 1994; Grant and Struchtemeyer 1959). However, the effect of rock fragments on the soil surface also depends on their relative position at the surface: rock fragments laying on the soil surface affect erosion by softening the splash effect of raindrops, delaying the time of initiation of run-off and preventing sealing processes (Daba Fufa et al. 2002; Martínez-Zavala and Jordán 2008). In contrast, rock fragments partly embedded in the soil surface contribute to an increase in run-off rates (de Figuereido and Poesen 1998; Poesen and Lavee 1994). However, a number of authors have demonstrated that the effect of rock fragment size on run-off and erosion can be more important than the position of the rock fragments (de Figuereido and Poesen 1998), and that small amounts of fine eroded sediments may be enough to clog macropores (Luce 1997).

Simulated rainfall under laboratory conditions is a widely used method for studying soil erosion on undisturbed soil samples, but few experiments have been carried out under natural conditions (Cerdà 2001). Simulated rainfall experiments allow researchers to control the amount of rainfall, rainfall intensity and time, which is convenient for the study of deformation mechanisms of soil under heavy storms (Meyer 1994), similar to Mediterranean climates. Rainfall simulation methods provide limited information because of the small size of the plots and the design of the simulator. It is difficult to extrapolate data about sediment production to a larger scale, but the results from simulations can be used for comparative purposes (Arnáez et al. 2004). In addition, simulated rainfall on small plots provides detailed information about the vertical processes controlling infiltration, run-off and erosion rates (Lange et al. 2003).

In many non-cultivated Mediterranean soils, shrubland is distributed as a patchy mosaic of bare and vegetation zones, and surface rock fragments may constitute an important protection against rainfall-induced erosion processes and surface sealing (Martínez-Zavala and Jordán 2008). The conditions of soil development in Spain (parent material, geomorphology and climate) have led to a relative abundance of rock fragments in the soil profile and on the soil surface (Ingelmo et al. 1994). The objective of the present study was to examine the effects of rock fragment coverage on the hydrological and erosional responses of Mediterranean bare soils.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Study area

The present study was carried out at Sierra del Bujeo, in Los Alcornocales Natural Park (south-western Spain), at approximately 36°04′45′′N and 5°31′26′′W (Fig. 1). This area is located in the El Aljibe range system. The risk of soil erosion in this area is high owing to heavy storms and the concentration of rainfall between October and April (Jordán and Bellinfante 2000). The parent material is Oligo-Miocene siliceous sandstone, and the soils are acidic, sandy and nutrient poor (mainly Cambisols, Regosols and Leptosols; International Union of Soil Sciences 2006) with rocky outcrops and a high proportion of rock fragments on the soil surface. The climate is Mediterranean, with cool, humid winters and warm, dry summers. Average annual rainfall ranges between 665 mm in the valleys and 1210 mm in the mountains. Mean temperature is mild, 16–18°C, with a monthly maximum of 31°C and a monthly minimum of 5°C. In summer, hillslopes intercept moisture from the prevailing south-eastern winds from the Mediterranean Sea and reduce to some extent the severity of drought.

image

Figure 1.  Map of the study area.

Download figure to PowerPoint

The vegetation is dominated by Quercus suber (cork oak) and locally abundant Quercus canariensis. Open heathland, dominated by Erica australis, among other species, covers the ridges and hilltops. Shrubland cover is very low at the studied plots, and bare soils dominate. A detailed view of the bare soils covered with rock fragments is shown in Fig. 2.

image

Figure 2.  (a) Photo showing the detail of the bare soil surfaces covered with rock fragments and individuals of the pioneer species Drosophyllum lusitanicum. (b) Photo showing the accumulation of coarse rock fragments and sediments at the end of the hillslope.

Download figure to PowerPoint

Soil and sediment analysis

The topsoil (0–5 cm depth) was sampled for physical and chemical analysis. Soil texture was determined using the Bouyoucos method (Ministerio de Agricultura, Pesca y Alimentación 1982). Bulk density was measured using the core method (Blake and Hartge 1986). A small fraction of each sample was air-dried and sieved (0–2 mm) for soil organic carbon analysis using the Walkley and Black method (Walkley and Black 1934). Soil moisture was determined in the first 5 cm gravimetrically before and after the rainfall simulations. The water content at 1500 kPa (wilting point), 33 kPa (field capacity) and 0 kPa (saturation) was determined using a pressure-membrane extraction apparatus (Richards 1941) as described in Ministerio de Agricultura, Pesca y Alimentación (1982).

Rock fragment cover was calculated by considering all mineral particles larger than 2 mm resting on the soil surface. Three classes of rock fragment coverage were considered: RFC1 (<50%), RFC2 (50–60%) and RFC3 (>60%). Most of the rock fragments in the studied plots were laying on the soil surface, whereas the average proportion of rock fragments embedded in the soil was <5%. The size of the rock fragments varied from 2 mm to 10 cm, and they were irregularly shaped.

Rainfall simulation experiments

The rainfall simulations were carried out in May 2007. Twelve soil plots were selected under each class of rock fragment coverage, resulting in 36 soil plots (12 soil plots × 3 classes = 36 experiments). A portable rainfall simulator similar to that described by Navas et al. (1990) was used. The structure is supported by four legs in the shape of a truncated pyramid and is covered with a wind protector. The length of the legs can be adjusted so that the simulator can be levelled when placed on a sloped surface. The nozzle at the top of the structure is 3.5 m high and it is connected through a rubber pipe to a mobile pump. Water falls from the nozzle onto an area of 1256.6 cm2, limited by a steel ring (40 cm in diameter). The ring was carefully tapped into the soil to avoid leakage and to direct run-off flow into the outlet of the plot. The rainfall intensity was measured by five rain gauges distributed uniformly over the plot. The rainfall simulator was calibrated at a constant rate of 26.8 mm h−1 and the simulation time was 60 min. Distilled water (pH 5.6) was used for the simulations because the soil response may be influenced by the chemical composition of the water (Agassi et al. 1994). A gutter installed on the downstream side of the plots conducted the run-off to a sample collection box.

Following Cerdà (2001), three variables were measured: time to ponding from the start of the application (TP), time to surface run-off (TR) and time to run-off outlet (TO). Time to ponding was measured when 40% of the surface showed ponds on flat or concave microsurfaces. Run-off occurred without previous ponding on the steeper microsurfaces, although it could be detected as shining on such areas before run-off began. According to Cerdà (2001), such visual determinations identify the areas where the top few millimetres of the soil are saturated. The same researcher made these assessments at all plots. The volume of run-off was determined every 5 min. Run-off water was also sampled every 5 min for determination of sediment load in the water. The samples were dried (110°C) until a constant weight and sediment yield was quantified by weighing. Soil loss was calculated by multiplying the sediment concentration by the run-off volume for each measurement interval. The infiltration rate was calculated as the difference between the rain intensity and the run-off rate. A constant run-off rate was observed after 40 min in all plots, and the steady-state run-off rate could be calculated as the average value of the last five measurements.

Fifteen minutes after the tests were finished at the RFC3 plots, the rock fragments on the soil surface were removed and the rainfall simulations were repeated (Fig. 3). The steady-state infiltration rate was not affected by the initial soil moisture. A comparison between the steady-state infiltration rate after the first and second round of tests shows the influence of the rock fragments on infiltration.

image

Figure 3.  Detail of one plot under rock fragment coverage >60% before (a) and after (b) removal of the rock fragments.

Download figure to PowerPoint

Data analysis

The relationship between run-off and time of simulation was described by a logistic function closely fitted by least squares approximation:

  • image

where a1 and a2 are the lower and upper limits (asymptotes) of the function, x0 is the value of x at which y is the mid-point between a1 and a2, and p is an exponent controlling the shape of the curve.

The relationship between soil loss rate and rock fragment cover can be expressed by an exponential equation:

  • image

where slr is the soil loss rate, r is the percentage cover of rock fragments and b and c are empirical coefficients. Similar equations have also been reported in Canada by Chow and Rees (1995), in the semiarid tropics in India by Mandal et al. (2005) and in sparse Quercus forests in Spain by Martínez-Zavala and Jordán (2008). All computations and graphical displays were made using Statistica version 6 (StatSoft 2001).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Soil analysis

Soil characterization is shown in Table 1. The soil pH was acidic (3.2–6.9). Most of the soil plots were sandy loam or sandy clay loam. The average gravel content in the first 5 cm was 18%, with values ranging from 0.2 to 70.5%. Bulk density ranged from 0.84 to 1.47 g cm−3. The organic matter content ranged from 1.1 to 5.6% and the initial soil moisture at the soil surface ranged from 2.2 to 4.8%.

Table 1.   Soil properties of the studied plots under the different classes of rock fragment cover classes
PlotRock fragment cover (%)pHOrganic matter (%)Sand (%)Clay (%)Coarse fragments (%)Soil moisture (%)Saturation water (%)Field capacity (%)Wilting point (%)Available water (%)Bulk density (g cm−3)
  1. RFC, rock fragment cover.

RFC1
 1–137.45.31.867.410.71.12.641.819.59.410.081.47
 1–239.75.01.715.955.54.63.754.346.532.314.211.37
 1–340.45.81.943.12.80.24.634.625.27.917.221.39
 1–440.44.41.749.626.338.73.147.726.715.511.211.33
 1–540.65.62.777.512.81.22.441.618.49.88.611.40
 1–643.96.52.011.360.75.23.955.149.835.714.061.20
 1–745.84.61.153.531.87.82.548.427.917.910.061.37
 1–847.33.82.169.217.012.32.744.121.212.09.181.32
 1–947.36.92.253.621.769.13.046.324.513.411.121.39
 1–1048.35.61.653.218.642.62.945.523.612.211.441.30
 1–1148.84.31.383.75.42.32.636.714.76.48.291.41
 1–1249.26.23.264.716.136.82.844.021.411.59.871.23
 Mean44.15.31.953.623.318.53.145.026.615.311.281.35
 Standard deviation4.21.00.622.118.222.60.76.110.79.32.650.08
RFC2
 2–151.36.53.252.717.32.03.345.123.311.511.791.22
 2–251.36.43.024.627.29.04.749.731.215.315.851.12
 2–352.56.52.159.216.70.93.144.522.311.410.871.27
 2–453.05.23.973.714.00.82.742.619.510.68.901.23
 2–553.95.02.170.822.95.22.545.322.514.28.281.34
 2–655.33.92.969.818.270.52.644.321.312.39.021.30
 2–755.66.03.461.110.637.43.541.920.49.211.191.22
 2–855.85.12.842.119.227.23.946.425.512.113.391.18
 2–956.86.32.56.256.50.13.855.148.633.615.011.07
 2–1057.64.32.943.228.726.43.448.628.416.112.351.17
 2–1158.55.13.737.929.326.73.649.229.916.813.091.08
 2–1260.04.13.561.820.34.52.745.322.913.09.900.95
 Mean55.15.43.050.323.417.63.346.526.314.711.641.18
 Standard deviation2.81.00.620.411.921.20.73.77.96.42.420.11
RFC3
 3–163.24.65.477.15.530.03.037.216.06.79.301.07
 3–265.24.94.662.38.743.53.240.819.98.511.330.94
 3–365.46.44.653.919.17.33.345.723.812.611.140.86
 3–465.55.44.364.43.35.23.335.918.46.611.760.94
 3–565.94.54.381.32.342.02.232.313.54.98.641.20
 3–666.95.13.456.28.11.43.940.220.88.212.530.97
 3–767.95.73.360.013.89.42.943.421.310.410.831.02
 3–867.94.84.954.123.914.92.946.925.314.510.830.84
 3–968.04.04.158.222.246.12.846.224.113.910.130.96
 3–1068.34.94.863.626.516.22.946.724.615.69.050.87
 3–1175.33.25.274.66.84.03.038.517.07.49.630.85
 3–1276.85.15.642.57.01.54.841.324.38.815.510.85
 Mean68.04.94.562.412.318.53.241.320.79.910.890.95
 Standard deviation4.10.80.711.08.517.20.64.73.93.51.860.11

The soil water properties were not related to the rock fragment cover. The saturation water content ranged from 32.3 to 55.1%. The field capacity (33 kPa) was 13.5–48.6% and the wilting point (1500 kPa) was 4.9–35.7%.

Time to ponding and time to run-off

Table 2 shows the response of the studied plots to the simulated rainfall. As expected, when the rock fragment cover increases run-off is retarded. The first change observed on the soil surface after rainfall starts is ponding. Ponding was observed first in areas where the soil surface was crusted. The rainfall infiltrated faster in areas near the rock fragments because of the greater porosity and aggregation of soil particles.

Table 2.   Initial soil moisture of the studied plots, time to ponding, time to run-off and time to run-off to the outlets under different rock fragment cover classes
PlotSoil moisture (%)Time to ponding (s)Time to run-off (s)Time to run-off to the outlet (s)
  1. RFC, rock fragment cover.

RFC1
 1–12.686225414
 1–23.7108270427
 1–34.6124293421
 1–43.1130273284
 1–52.4118276389
 1–63.9152323364
 1–72.5246410547
 1–82.7209347555
 1–93.0209360450
 1–102.9156292375
 1–112.6155334437
 1–122.8252409575
 Mean3.1162.1317.7436.5
 Standard deviation0.754.756.985.8
RFC2
 2–13.3207377503
 2–24.7229353465
 2–33.1270439547
 2–42.7222347418
 2–52.5221399512
 2–62.6257419560
 2–73.5282431684
 2–83.9206386596
 2–93.8254411551
 2–103.4238369529
 2–113.6270395526
 2–122.7318487589
 Mean3.3247.8401.1540.0
 Standard deviation0.733.639.667.3
RFC3
 3–13.0345522778
 3–23.2435602829
 3–33.3339480752
 3–43.3286456714
 3–52.2393568755
 3–63.9344519724
 3–72.9359479588
 3–82.9389514683
 3–92.8349519749
 3–102.9419588763
 3–113.0494627818
 3–124.8468594756
 Mean3.2385.0539.0742.4
 Standard deviation0.660.155.363.2
All classes
 Mean3.2264.97419.25572.97
 Standard deviation0.6105.35105.08146.95

Time to ponding was fastest at the soil plots with a low rock fragment cover. The average time to ponding was 162.1 s at the RFC1 plots, 247.8 s at RFC2 and 385.0 s at RFC3. Surface run-off and run-off in the outlet started later as the rock fragment cover increased. Times to run-off were 317.7 (RFC1), 401.1 (RFC2) and 539.0 s (RFC3). Finally, the average times to run-off in the outlet were 436.5 (RFC1), 540.0 (RFC2) and 742.4 s (RFC3).

Run-off and infiltration rates

The run-off rate decreased when the rock fragment cover increased (Fig. 4). The average run-off coefficients were 21.8% (RFC1), 13.1% (RFC2) and 5.2% (RFC3) (Table 3). Steady-state run-off rates were reached after 40 min at all rock fragment cover intervals. The final run-off rates were very low for RFC3, but were higher for RFC1. The steady-state infiltration rates increased with rock fragment cover. The average values were 0.6, 1.1 and 1.4 mm h−1 for RFC1, RFC2 and RFC3, respectively.

image

Figure 4.  Relationship between run-off rate and time for the different rock fragment cover classes (RFC1 [<50%], RFC2 [50–60%] and RFC3 [>60%]). Error bars are ± 0.95 standard error.

Download figure to PowerPoint

Table 3.   Run-off coefficients, maximum run-off rates, steady-state infiltration rates and average run-off rates for the three different rock fragment cover classes
 RFC1RFC2RFC3
  1. RFC, rock fragment cover.

Runoff coefficients (%)21.813.15.2
Maximum run-off rate (mm h−1)12.66.82.7
Steady-state infiltration rate (mm h−1)0.61.11.4
Average run-off rate (mm h−1)8.04.81.9
Standard deviation4.52.20.9

Sediment yield and soil loss

Sediment yield increased linearly at the beginning because it takes time for the soil to get wet and for the soil particles to become detached from the surface. The sediment concentration increased during the first minutes of rainfall. After 20 min from the beginning of the experiments, a peak in sediment yield was reached for all rock fragment cover classes. After this peak in the sediment yield there was a steady decrease in the sediment concentration (Fig. 5).

image

Figure 5.  Relationship between sediment yield and time for the different rock fragment cover classes (RFC1 [<50%], RFC2 [50–60%] and RFC3 [>60%]). Error bars are ± 0.95 standard error.

Download figure to PowerPoint

Sediment yield decreased when rock fragment cover increased. The average sediment yield decreased from 0.54 (RFC1) to 0.34 (RFC2) and 0.21 g L−1 (RFC3). The average soil loss rate decreased from 1.90 (RFC1) to 0.60 (RFC2) and 0.15 kg m−2 h−1 (RFC3). Figure 6 shows the exponential relationship between the soil loss rate and the rock fragment cover.

image

Figure 6.  Relationship between soil loss rate and rock fragment cover (RFC1 [<50%], RFC2 [50–60%] and RFC3 [>60%]). Both parameters are related by the equation: y = 138.6 e−0.1002x; R2 = 0.836.

Download figure to PowerPoint

Effects of rock fragment removal

After the rock fragments were removed, the average rock fragment cover decreased from 68.02 to 3.04% (Table 4). The latter value was not zero, in part, because of the embedded mineral particles. The initial average soil moisture content was 3.2%, but this value increased up to 20.6% after the second set of rainfall simulations. The average time to ponding decreased to 38% with a lower rock fragment cover. As a consequence, time to run-off decreased from 539.00 to 258.25 s (48% of the initial value). The run-off coefficient varied from 4.83% in the first set of simulations to 11.07% in the second set, whereas the steady-state run-off rate increased from 2.60 to 4.58 mm h−1.

Table 4.   Rock fragment cover, initial moisture, time to ponding (TP), time to run-off (TR), run-off coefficient (RC), steady-state run-off rate, sediment yield and soil loss rate at the rock fragment cover >60% plots in the first and second tests
PlotRock fragment cover (%)Moisture (%)TP (s)TR (s)TR–TP (s)RC (%)Steady-state run-off rate (mm h−1)Sediment yield (g L−1)Soil loss rate (kg m−2 h−1)
First test
 Mean68.03.23855391544.82.618.80.2
 Standard deviation4.10.66055231.70.91.30.1
Second test
 Mean3.020.615025810911.14.6190.63.6
 Standard deviation1.33.43537273.61.493.31.7

The interval of time between ponding and run-off initiation (TR–TP) was shorter when the rock fragment cover was low. The average TR–TP times were different for each run (P = 0.0002), although no significant correlation was observed between rock fragment cover and TR–TP.

Higher run-off coefficients and steady run-off rates led to an increment in sediment yield and soil loss. The average sediment yield increased from 18.84 g L−1 in the first test to 190.61 g L−1 in the second test. In addition, the average soil loss rate increased from 0.15 to 3.62 kg m−2 h−1.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

According to Poesen et al. (1994), the influence of surface rock fragments on sediment yield from bare interrill areas depends on the effect of the rock fragments on run-off generation and flow concentration. Rock fragments on the soil surface increase roughness and interception of raindrops, delaying ponding and run-off. This enhanced the infiltration rate during simulated storms, as shown by Poesen et al. (1990), Abrahams and Parsons (1994), Cerdà (2001) and Mandal et al. (2005). This effect can be explained by the irregularity of the surface in stony soils. In the exposed soil surface between rock fragments, the pressure of the water column is greater and water infiltrates more quickly and more deeply. Rock fragments channel the water flow between them, generating deeper, more hydraulically efficient flow, as shown by Wainwright (1996) and Martínez-Zavala and Jordán (2008). In agreement with Cerdà (2001) and Mandal et al. (2005), we found that water infiltrated more easily in the area under the rock fragments.

The run-off rate varied as a logistic function of time. The run-off rate increases at the beginning of the simulated rainfall experiments, but becomes stable after a period of time. It is likely that soil sealing after some minutes of simulated rainfall contributes to the stable run-off and infiltration rates. A steady final run-off rate is reached faster when the rock fragment cover is high. Long or frequent storms might be more dangerous when the rock fragment cover is low and the soil is exposed because rock fragments on the soil surface reduced the speed of the run-off flow and the run-off rate. Slow run-off flow supports low quantities of sediment, reducing soil sealing and increasing infiltration.

Corey and Kemper (1968) and Ingelmo-Sánchez et al. (1980) demonstrated that rock fragment cover enhances infiltration and reduces evaporation after heavy storm events. Mediterranean soils with an appreciable rock fragment cover can store more water than soils with a low rock fragment cover (Danalatos et al. 1995). Thus, rock fragments on the soil surface might be important for bare soils under Mediterranean conditions, in order to control the soil water balance.

The rock fragment cover contributed to reducing soil erosion. Laboratory experiments have shown that the removal of rock fragments leads to a decrease in infiltration because part of the porosity has disappeared and the soil surface is not prevented from sealing by rock fragments (Grant and Struchtemeyer 1959). Observed run-off coefficients were much higher when the rock fragments were removed from the soil surface. These differences cannot be attributed only to rock fragment cover without taking into account other characteristics of the rock fragments (as shown by Valentin 1994). A number of factors, such as relative position, size and porosity, influence infiltration, run-off, erosion and water retention (Childs and Flint 1990; Poesen and Lavee 1994). The position of rock fragments, laying on the surface or embedded in the soil, conditions the hydrological response of the soils because the evaporation rate and water flow are modified (Jury and Bellantuoni 1976; Pérez 1998; Poesen and Lavee 1994).

The removal of rock fragments affected the hydrological response and the soil erodibility. The steady-state run-off rate was increased after the rock fragments were removed. This variable was not affected by the initial soil moisture differences between the experiments because both run-off and infiltration were stable after 20–30 min of simulated rainfall. It is necessary to examine how successive rainfall events influence soil physical properties. Changes in macroporosity have been studied by other authors, who have not detected an influence on soil sealing (Schiettecatte et al. 2005).

Removing the rock fragments increased the soil loss rate. Cerdà (2001) pointed out that this effect results, in part, from the faster run-off, but the main cause is the increased soil erodibility. In agreement with other microscale experiments (Poesen et al. 1994), we observed that rock fragments at the soil surface had a negative effect on sediment yield and can be considered to be natural soil surface stabilizers.

Conclusions

A high rock fragment cover on the soil surface contributes to a delay in ponding. Ponding is detected later on areas near rock fragments, where rainfall infiltrated faster because of the greater porosity and aggregation of soil particles. Rock fragment cover contributed to delayed run-off flow and increased infiltration rates, diminishing the soil loss rates. The relationship between soil loss and rock fragment cover at the studied area was expressed by an exponential function.

After the rock fragments were removed, the time to ponding and the time to run-off decreased considerably. Run-off coefficients, final run-off rates and soil erosion increased after the rock fragments were removed.

These results have implications for erosion modelling and soil conservation in areas with the same climate and soil characteristics, but further research and standard soil loss estimations are still needed.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The authors are grateful to José Luis Lozano (University of Sevilla) for preparing the soil samples and helping with some of the laboratory analysis, and to Dr Artemi Cerdà (University of Valencia) who reviewed a previous version of the manuscript. Dr R. Green checked the style. Dr Masanori Saito (Tohoku University) and Dr Morihiro Maeda (Okayama University) helped to improve the original manuscript. We are also grateful to the staff of the Natural Park for their attention and assistance during part of the field work.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • Abrahams AD, Parsons AJ 1991: Relation between infiltration and stone cover on a semiarid hillslope, southern Arizona. J. Hydrol., 122, 4959.
  • Abrahams AD, Parsons AJ 1994: Hydraulics of interrill overland flow on stone-covered desert surfaces. Catena, 23, 111140.
  • Agassi M, Shainberg I, Van Der Merwe D 1994: Effect of water salinity on inter-rill erosion and infiltration: laboratory study. Aust. J. Soil Res., 32, 595601.
  • Arnáez J, Larrea V, Ortigosa L 2004: Surface runoff and soil erosion on unpaved forest roads from rainfall simulation tests in northeastern Spain. Catena, 57, 114.
  • Auzet AV, Boiffin J, Ludwig B 1995: Concentrated flow erosion in cultivated catchments: influence of soil surface state. Earth Surf. Proc. Land., 20, 514.
  • Blake GR, Hartge KH 1986: Bulk density. In Met, 2nd edn., Ed. AKlute., pp. 363375, American Society of Agronomy, Madison.
  • Brakensiek DL, Rawls WJ 1994: Soil containing rock fragments: effects on infiltration. Catena, 23, 99110.
  • Bunte K, Poesen J 1994: Effects of rock fragment size and cover on overland flow hydraulics, local turbulence and sediment yield on an erodible soil surface. Earth Surf. Proc. Land., 19, 115135.
  • Cerdà A 2001: Effects of rock fragment cover on soil infiltration, interrill runoff and erosion. European Journal of Soil Science, 52, 5968.
  • Childs SW, Flint AL 1990: Physical properties of forest soils containing rock fragments. In Sustained Productivity of Forest Soils Ed. SPGessel, DSLacate, GFWeetman and RFPowers., pp. 95121, University of British Columbia, Faculty of Forestry, Vancouver.
  • Chow TL, Rees HW 1995: Effects of coarse-fragment content and size on soil erosion under simulated rainfall. Can. J. Soil. Sci., 75, 227232.
  • Corey AT, Kemper WD 1968. Conservation of Soil Water by Gravel Mulches. Colorado State University Hydrology Paper No 30, Fort Collins.
  • Cousin I, Nicoullaud B, Coutadeur C 2003: Influence of rock fragments on the water retention and water percolation in a calcareous soil. Catena, 53, 97114.
  • Daba Fufa S, Strauss P, Schneider W 2002: Comparison of erodibility of some hararghe soils using rainfall simulation. Communications in Soil Science and Plant Analysis, 33, 333348.
  • Danalatos NG, Kosmas CS, Moustakas NC, Yassoglou N 1995: Rock fragments: II. Their impact on soil physical properties and biomass production under Mediterranean conditions. Soil Use and Management, 11, 121126.
  • De Figuereido T, Poesen J 1998: Effects of rock fragment characteristics on interrill runoff and erosion of a silty loam soil. Soil & Tillage Research, 46, 8195.
  • Flanagan DC, Nearing MA (Eds.). 1995. USDA-Water Erosion Prediction Project: technical Documentation, NSERL Rep, vol. 10. Natl. Soil Erosion Res. Lab, West Lafayette.
  • Grant WJ, Struchtemeyer RA 1959: Influence of the coarse fraction in two Maine potato soils on infiltration, runoff and erosion. Soil Science Society of America Proceedings, 23, 391396.
  • Ingelmo F, Cuadrado S, Ibañez A, Hernandez J 1994: Hydric properties of some Spanish soils in relation to their rock fragment content: implications for runoff and vegetation. Catena, 23, 7385.
  • Ingelmo-Sánchez F, Cuadrado-Sánchez S, Blanco de Pablos A 1980: Evaporación de agua en suelos de distinta textura. Anuario del Centro de Edafología y Biología Aplicada de Salamanca, 6, 255280.
  • International Union of Soil Sciences 2006. World Reference Base for Soil Resources. 2nd edn. World Soil Resources Reports, 103. FAO, Rome.
  • Jordán A, Bellinfante N 2000: Cartografía de la erosividad de la lluvia estimada a partir de datos pluviométricos mensuales en el Campo de Gibraltar (Cádiz). Edafología, 7-3, 8392.
  • Jury WA, Bellantuoni B 1976: Heat and water movement under surface rocks in a field soil, II. Moisture effects. Soil Science Society of America Journal, 40, 509513.
  • Lange J, Greenbaum N, Husary S, Ghanem M, Leibundgut C, Schick AP 2003: Runoff generation from succesive simulated rainfalls on a rocky, semi-arid, Mediterranean hillslope. Hydrol. Process., 17, 279296.
  • Luce CH 1997: Effectiveness of road ripping in restoring infiltration capacity of forest roads. Restoration Ecology, 5, 265270.
  • Mandal UK, Rao KV, Mishra PK et al. 2005: Soil infiltration runoff and sediment yield from a shallow soil with varied stone cover and intensity of rain. European Journal of Soil Science, 56, 435443.
  • Ministerio de Agricultura, Pesca y Alimentación, 1982. Métodos oficiales de análisis de suelos. Ministerio de Agricultura, Pesca y Alimentación, Madrid.
  • Martínez-Zavala L, Jordán A 2008: Effect of rock fragment cover in interrill soil erosion from bare soils in Western Andalusia, Spain. Soil Use and Management, 24, 108117.
  • Meyer LD 1994. Rainfall simulators for soil erosion research. In Soil Erosion Research Methods, 2nd edn., Ed. RLal., pp. 7595, Soil and Water Conservation Society (Ankeny) and St Lucie Press, Delray Beach.
  • Navas A, Alberto F, Machín J, Galán A 1990: Design and operation of a rainfall simulator for field studies on runoff and soil erosion. Soil Technology, 3, 385397.
  • Papy F, Douyer C 1991: Influence des états de surface du territoire agricole sur le déclenchement des inondations catastrophiques. Agronomie, 11, 201215.
  • Parsons AJ, Wainwright J, Abrahams AD 1996: Runoff and erosion on semi-arid hillslopes. In Advances in Hillslope Processes Ed. MGAnderson., pp. 10611078, Wiley, Chichester.
  • Pérez FL 1998: Conservation of soil moisture by different stone covers on alpine talus slopes (Lassen, California). Catena, 33, 155177.
  • Poesen J, Bunte K 1996: The effects of rock fragments on desertification processes in Mediterranean environments. In Mediterranean Desertification and Land Use Ed. CJBrandt and JBThornes., pp. 247269, John Wiley & Sons, Chichester.
  • Poesen J, Ingelmo-Sánchez F, Mucher H 1990: The hydrological response of soil surfaces to rainfall as affected by cover and position of rock fragments in the top layer. Earth Surf. Proc. Land., 15, 653671.
  • Poesen J, Lavee H 1994: Rock fragments in top soils: significance and processes. Catena, 23, 128.
  • Poesen JW, Torri D, Bunte K 1994: Effects of rock fragments on soil erosion by water at different spatial scales: a review. Catena, 23, 141166.
  • Poesen J, Van Wesemael B, Bunte K, Solé A 1998: Variation of rock fragment cover and size along semiarid hillslopes: a case-study from southeast Spain. Geomorphology, 23, 323335.
  • Richards LA 1941: A pressure-membrane extraction apparatus for soil solution. Soil Sci., 51(5), 377386.
  • Schiettecatte W, Jin K, Yao Y et al. 2005: Influence of simulated rainfall on physical properties of a conventionally tilled loess soil. Catena, 64, 209221.
  • Valentin C 1994: Surface sealing as affected by various rock fragment covers in West Africa. Catena, 23, 8797.
  • Wainwright J 1996: Infiltration, runoff and erosion characteristics of agricultural land in extreme storm events, SE France. Catena, 26, 2747.
  • Walkley A, Black IA 1934: An examination of the Degtjareff method for determining soil O.M. and a proposed modification of the chromic acid titration method. Soil Sci., 37, 2938.
  • Van Wesemael B, Poesen J, Kosmas CS, Danalatos NG, Nachtergaele J 1996: Evaporation from cultivated soils containing rock fragments. J. Hydrol., 182, 6582.
  • Wischmeier WH, Smith DD 1978: Predicting rainfall erosion losses. A guide to conservation planning. USDA Agricultural Handbook 537, Washington.