Surface erosion and underground leakage of yellow soil on slopes in karst regions of southwest China
Abstract
Surface erosion and underground leakage loss simultaneously exist on slopes with a double-layered structure (surface and underground) and have caused rocky desertification in karst regions. Because of the great difficulty of directly determining the underground leakage loss of soil and water, soil erosion processes in this region remain unclear, especially underground leakage loss. The aim of this study was to reveal the plot-scale characteristics of surface soil erosion and underground leakage loss of yellow soils on karst slopes. Simulated rainfall tests and field runoff plot monitoring were conducted to achieve this aim. We found that surface erosion formed on steep slopes (30°) or at greater rainfall intensities (over 50 mm hr−1) and that surface runoff and surface soil loss dominated the total runoff and soil loss on karst slopes at greater rainfall intensities. However, underground leakage loss always occurred at different slopes and rainfall intensities. For small rainfall intensity cases (lower than 80 mm hr−1), underground runoff and underground soil leakage loss dominated the total runoff and soil loss and showed an underground runoff ratio of over 90% and an underground soil leakage ratio of over 44%. Rainfall intensity had significant effects on surface runoff depth and surface soil loss rates but an insignificant influence on underground leakage loss. Slope gradient and slope length influenced the runoff yield, and slope degree had a greater effect on the surface soil loss than had slope length. This study improves our understanding of surface erosion and underground leakage loss in karst rocky desertification regions.
1 INTRODUCTION
The southwest karst region in China, centered in Guizhou Province, is the largest continuous karst region in the world (Sweeting, 1993). Guizhou Province is known as China's karst province. The karst rocky desertified land of Guizhou Province covers 35,000 km2, approximately 20% of the area of the province (Wang et al., 2004). This region is characterized by fragile ecological environments and serious soil erosion that has caused extreme rocky desertification, soil structural deterioration, and soil impoverishment (Wang, Liu, & Zhang, 2004). Karst rocky desertification, a consequence of soil erosion and bedrock exposure, is one of the world's most serious social and environmental concerns (Bai, Zhang, Chen, & He, 2010; García-Ruiz, Beguería, Lana-Renault, Nadal-Romero, & Cerdà, 2017; Liu, Wang, & Deng, 2008; Xiong et al., 2009). It is well known that soil erosion can directly accelerate rocky desertification (Febles-González, Vega-Carreño, Amaral-Sobrinho, Tolón-Becerra, & Lastra-Bravo, 2014; Wei, Yan, Xie, Ni, & Loa'iciga, 2016); however, the mechanism of soil erosion in this region is not clear. Therefore, a complete understanding of the soil erosion mechanism and characteristics has become more urgent.
Because of their unique climatic background and geological structure, karst slopes have a double-layered hydrogeological structure (surface and underground) in epikarst zones (Auler & Smart, 2003; Dai, Liu, Shao, & Yang, 2015; Dai, Peng, Yang, & Zhao, 2017). The surface exhibits a landscape of discontinuous regolith and large bare rock areas (Chen, Liu, Wang, & Zhang, 2011). The underground hides many sinkholes and fissures formed from the dissolution of soluble rocks (Dobecki, 2006; Farrant & Cooper, 2008). Fissures are any discontinuity within the rock mass that is either initially open or capable of being opened by dissolution to provide a route for water movement (Dai, Peng, Zhao, Shao, & Yang, 2017). Fissures act as the foci for centripetal flow paths that underdrain the fissured epikarst (Hartmann, Lange, Weiler, Arbel, & Greenbaum, 2012; Williams, 2008). Most rainwater and soil are transported to underground rivers through the fissures and form soil underground leakage (Peng, Dai, Li, Yuan, & Zhao, 2017; Peng, Dai, Yang, & Zhao, 2016). Little runoff and soil loss are observed on the surfaces of karst slopes because of their special soil loss type (Peng & Wang, 2012), which is obviously different from soil loss in nonkarst regions (Bewket & Teferi, 2009; Cerdà, Flanagan, & Le Bissonnais, 2009; Cerdà, Lavee, Romero-Díaz, Hooke, & Montanarella, 2010; Hancock, Wells, Martinez, & Dever, 2015; Meyer, Poesen, Isabirye, Deckers, & Raes, 2011; Peng, Shi, Jiang, Wang, & Li, 2014).
Surface erosion and underground leakage loss simultaneously exist on karst slopes and have caused rocky desertification in karst regions (Dai et al., 2015; Zhang, Bai, & He, 2011). Some researchers have studied soil erosion on karst slopes. For example, Peng and Wang (2012) used the large runoff field method to measure surface runoff and soil loss, Luo, Zhou, and Wang (2011) adopted exposed root anatomical features to imply karst soil erosion, Feng et al. (2016) used the revised Universal Soil Loss Equation (RUSLE) to estimate the annual soil erosion rates, and Xu, Shao, Kong, Peng, and Cai (2008) integrated the RUSLE with geographic information system to estimate soil loss and identify risk erosion areas. However, those studies focused only on surface erosion. Some studies have also described the underground leakage loss. Wang et al. (2014) proposed the erosion–creep–collapse mechanism of underground soil loss for karst rocky desertification. Zhou, Tang, Yang, and Zhang (2012) revealed the detailed inference of the creep mechanism in the underground soil loss using a conceptual creep model. Wei et al. (2016) described the soil leakage phenomena qualitatively and quantitatively at small, medium, and large spatial scales using the isotope 137Cs tracing method. Because of the great difficulty of directly determining the underground leakage loss of soil and water, soil erosion processes in this region remain unclear, especially the underground leakage loss and the rate and relative contributions of soil underground leakage and surface erosion. Accurately understanding karst soil erosion would face the great obstacle of underground leakage loss measurement because of the dualistic overground and underground systems (Luo et al., 2011).
This paper represents an initial effort to study surface erosion and underground leakage loss from an integrated soil–bedrock–fissure system perspective using karst bare slopes as an example. A combination of simulated rainfall tests and runoff plot monitoring was conducted to measure plot-scale surface loss and underground leakage loss of soil and water. The study specifically aims to (a) reveal the formation of surface erosion and underground leakage on karst slopes, (b) determine the rates and contributions of surface erosion and soil underground leakage, and (c) analyze the influences of rainfall intensity, slope degree, and slope length on surface erosion and underground leakage. Results from this study could provide a better understanding of soil erosion processes in karst rocky desertification regions.
2 MATERIALS AND METHODS
2.1 Experimental materials
2.1.1 Test soils
Test soils were collected from a typical karst slope area (Figure 1) in Dafang County, Bijie City, of Guizhou Province, China. The rock stratum in this area is mostly limestone, sand, and shale, and outcrops are mostly carbonate limestone. Regional soils are dominated by yellow soil and paddy soil. The test soil was yellow soil classified as Haplic Acrisol in the Food and Agriculture Organization Taxonomy; it is the largest soil type in Guizhou Province, with a total area of 738 million hm2. The test soil was characterized by acidic, strong clay, and poor nutrient content. Yellow soil can be developed in a variety of parent materials, dominated by granite, sandstone, shale, and the quaternary red clay and limestone weathering, which is represented by the Guizhou Plateau. The test soil was not screened, but a large soil mass was dispersed for simulated rainfall and runoff field tests. The soil particle size distribution was as follows: 23.48% clay (<0.001 mm), 22.82% fine silt (0.001–0.005 mm), 16.34% medium silt (0.005–0.01 mm), 21.80% coarse silt (0.01–0.05 mm), and 15.57% sand (>0.01 mm).
2.2 Test equipment
Experiments were conducted at Guizhou University, Guiyang City, China. The experimental equipment consisted of a rainfall simulator and a steel tank. The rainfall simulator (QYJY-502) was a portable fully automatic simulator with four stainless steel down-sprayers, similar to that described by Cerdà (1998). The sprayers were 6 m in height, 6.5 m × 6.5 m in effective rainfall area, and more than 85% uniform in the distribution of raindrops. Rainfall intensity ranged from 6 to 180 mm hr−1 and could be remotely controlled or manually adjusted with less than 30 s of adjustment time and 6 mm hr−1 of adjustment precision. Raindrop velocity could meet natural rainfall characteristics. Two rain gauges were installed at both sides of the steel tank to determine actual rainfall intensity.
The size of the steel tank with drainage holes was 4.0 × 1.5 × 0.35 (Length × Width × Depth, m; Figure 2). The drainage holes had a diameter of 5 cm and were uniformly formed for free drainage of infiltrating water. The density of drainage holes could be adjusted by changing the contact area of the holes between the movable tank floor and the fixed tank floor. Surface and underground runoff sediment could be collected at the surface and the lower end of the steel tank, respectively.
2.3 Experimental design and methods
2.3.1 Simulated rainfall test
- Bedrock bareness rate is the ratio of the outcrop area of bedrocks to the horizontal projected area of the slope (Dai et al., 2015). The experimental bedrock bareness rate was designed to be 20% to represent the wide distribution of cultivated land in the karst areas of Guizhou Province; it was simulated in the test steel tank by randomly arranged limestone rocks with a diameter of greater than 35 cm. Soils with a total depth of 30 cm were homogenously backfilled layer by layer in rock spaces. Surface soil was leveled, and the area of contact between soils and rocks was compacted by hand to eliminate the influence of edge effects.
- Drainage holes at the bottom of the steel tank were used to simulate pore fissures in karst areas and to measure underground runoff and sediment. Underground pore fissure degree is the ratio of drainage holes to the floor area of the steel tank. The experimental underground pore fissure degree was 5%.
- Soil layers were divided into three sublayers with a total depth of 30 cm. Soil compactness was 1070, 760, and 410 kPa for the lower sublayer, median sublayer, and upper sublayer, respectively, according to soil compaction field measurements.
- Experimental slopes were set to 5°, 10°, 15°, 20°, and 30°.
- Experimental rainfall intensities were designed to represent light rainfall intensity (30 and 50 mm hr−1), moderate rainfall intensity (80 and 120 mm hr−1), and heavy rainfall intensity (150 mm hr−1), on the basis of the characteristics of erosive rainfall in karst areas of Guizhou Province.
A light rain was simulated before each rainfall experiment to accelerate soil sedimentation and reach the saturated soil water content. Timing was started when runoff was produced from either the surface or underground. The surface and underground runoff sediment samples were collected according to a method using 10-min intervals. Runoff volume was measured using large plastic metering buckets. Sediment yield was determined by the oven-drying method (sediment samples were dried in an air-forced oven to a constant weight at 105 °C and weighed). The rainfall duration of each rainfall event was 90 min. Each rainfall event was repeated three times. The soil in the steel tank was replaced by fresh soil after each rainfall event.
2.4 Natural rainfall monitoring on runoff plots
Eight runoff plots (Figure 3) were built in 2011 at Guizhou University, and their soils were filled according to the method of the simulated rainfall test. Details of the runoff plots are presented in Table 1. Rain events were observed from January to December in 2014.
| Plots | Slope (°) | Length (m) | Width (m) | Soil type | Treatment |
|---|---|---|---|---|---|
| I | 5 | 4.7 | 2 | Yellow soil | Soil particle distribution is the same as in the experimental soil used for the rainfall simulation tests. |
| II | 15 | 5.1 | 2 | ||
| III | 25 | 5.2 | 2 | ||
| IV | 45 | 6.5 | 2 | ||
| V | 15 | 10 | 2 | ||
| VI | 8.8 | 15 | 2 | ||
| VII | 13 | 20 | 2 | ||
| VIII | 3 | 20 | 5 |
For each rainfall event, the runoff depth in the square pools was measured by a ruler, and 0.5 L of sediment-laden water was sampled to measure sediment production after full stirring and mixing of water and sediment in the square pools. The sediment concentration was determined by the oven-drying method. Finally, the square pools were cleaned for the next rainfall event.
3 RESULTS
3.1 Loss leakage of runoff for simulated rainfall events
Water loss leakage in karst areas has become a serious environmental and social issue. Table 2 shows the runoff depths yielded on slopes of yellow soil under different rainfall intensities and slope degrees. Rainfall intensity significantly (p < .05) influenced runoff yields on surfaces but insignificantly affected runoff yields in underground pore fissures of the karst slopes of the yellow soil. Surface runoff depth significantly increased with an increase of rainfall intensity, whereas underground runoff depth presented no obvious relation. Rainfall intensity obviously determined surface runoff depth. Surface runoff depth showed significant differences between different rainfall intensities, except for small rainfall intensities (30 and 50 mm hr−1). Small rainfall intensity also produced significant differences in surface runoff depth, compared with moderate rainfall intensity (80 and 120 mm hr−1) and heavy rainfall intensity (150 mm hr−1). However, there were no obvious changes in underground runoff depth.
| Slope (°) | Surface runoff depth under different rainfall intensities (mm) | Underground runoff depth under different rainfall intensities (mm) | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| 30 mm hr−1 | 50 mm hr−1 | 80 mm hr−1 | 120 mm hr−1 | 150 mm hr−1 | 30 mm hr−1 | 50 mm hr−1 | 80 mm hr−1 | 120 mm hr−1 | 150 mm hr−1 | |
| 5 | 0.00 Db | 0.00 Db | 64.58 Cc | 114.10 Bd | 143.71 Ae | 39.90 Ecd | 68.68 Bb | 52.37 Da | 63.07 Ca | 79.64 Aa |
| 10 | 0.00 Db | 0.00 Db | 66.34 Cc | 119.14 Bc | 160.44 Ad | 42.99 Dab | 62.96 Ac | 51.96 Ca | 58.39 Bb | 62.62 Ab |
| 15 | 0.00 Db | 0.00 Db | 73.07 Cb | 134.76 Ba | 173.41 Ab | 41.41 Dbc | 64.36 Ac | 36.58 Ed | 44.34 Cd | 48.83 Bd |
| 20 | 0.00 Db | 0.00 Db | 78.39 Ca | 134.62 Ba | 180.20 Aa | 44.52 Ba | 74.32 Aa | 39.02 Dc | 41.15 Ce | 41.86 Ce |
| 30 | 2.89 Da | 4.04 Da | 72.36 Cb | 128.36 Bb | 169.74 Ac | 38.49 Ed | 68.13 Ab | 44.26 Db | 47.15 Cc | 50.81 Bc |
- Note. The same column marked with different lowercase letters indicates significant differences between groups (p < .05), the same row marked with different capital letters indicates significant difference between groups (p < .05), and the same letter indicates that the differences between groups were not significant (p > .05).
Slope conditions affected the occurrence of surface runoff but had little influence on underground runoff. Surface runoff was not generated on certain slopes (5–20°) at small rainfall intensity. A critical rainfall intensity for surface runoff yield existed and changed between 50 and 80 mm hr−1 on karst slopes of the yellow soil. However, runoff occurred on surfaces of karst slopes when the slope was increased to 30°. Moreover, for surface runoff yield cases (80–150 mm hr−1), surface runoff depth first increased and then decreased with the increase of the slope degree, which always reached the maximum at a slope of 20°. The critical slope of maximal surface runoff depth was 20°. However, slope had no obvious effect on underground runoff.
Figure 4 shows the ratio of surface runoff and underground pore fissure runoff to total runoff depth. Underground runoff dominated the total runoff on karst slopes for small rainfall intensity cases. The underground runoff yield ratio was more than 90%. For greater rainfall intensity cases, surface runoff and underground runoff shared the total runoff on karst slopes. The surface runoff yield ratio was as high as 55%, and the underground runoff yield ratio was lower than 45%.
3.2 Loss leakage of soil for simulated rainfall events
The loss leakage of soil on karst slopes is the main reason for the occurrence and development of rocky desertification (Peng et al., 2016). Table 3 shows the soil loss rate on the surface and the soil leakage rate underground under different rainfall intensities and slope degrees. As with the runoff loss leakage, rainfall intensity had a significant (p < .05) effect on the surface loss rate of soil but an insignificant influence on the underground leakage rate of soil on karst slopes. The surface loss rate of soil for small rainfall intensity cases was significantly different than for the moderate rainfall intensity and heavy rainfall intensity cases. Beyond the small rainfall intensity cases, the surface loss rate of soil increased with the increase of rainfall intensity and was higher than the underground leakage rate of soil. However, the underground leakage rate of soil had no obvious changes: Changes were between 0.11 and 0.33 g·m−2·min−1.
| Slope (°) | Surface soil loss rates under different rainfall intensities (g·m−2·min−1) | Soil underground leakage rates under different rainfall intensities (g·m−2·min−1) | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| 30 mm hr−1 | 50 mm hr−1 | 80 mm hr−1 | 120 mm hr−1 | 150 mm hr−1 | 30 mm hr−1 | 50 mm hr−1 | 80 mm hr−1 | 120 mm hr−1 | 150 mm hr−1 | |
| 5 | 0.00 Db | 0.00 Db | 0.84 Bb | 1.00 Bd | 1.50 Ab | 0.15 Dab | 0.24 Ba | 0.29 Aab | 0.20 Cc | 0.21 BCc |
| 10 | 0.00 Db | 0.00 Db | 0.92 Cb | 1.25 Bc | 1.70 Aab | 0.13 Cbc | 0.19 Bb | 0.31 Aa | 0.31 Aa | 0.30 Aa |
| 15 | 0.00 Cb | 0.00 Cb | 1.19 Ba | 1.76 Aa | 1.75 Aab | 0.14 Cabc | 0.14 Cc | 0.22 Bc | 0.22 Bc | 0.29 Aab |
| 20 | 0.00 Cb | 0.00 Cb | 1.29 Ba | 1.50 Bb | 1.91 Aa | 0.17 Ba | 0.24 Aa | 0.26 Ab | 0.26 Ab | 0.26 Ab |
| 30 | 0.09 Ca | 0.16 Ca | 1.27 Ba | 1.72 Aab | 1.61 Ab | 0.11 Cc | 0.13 Cc | 0.27 Bb | 0.33 Aa | 0.28 Bab |
- Note. The same column marked with different lowercase letters indicates significant differences between groups (p < .05), the same row marked with different capital letters indicates significant difference between groups (p < .05), and the same letter indicates that the differences between groups were not significant (p > .05).
Slope had some effects on the surface loss rate of soil but little influence on the underground leakage rate of soil on karst slopes. Both the surface loss rate of soil and the underground leakage rate of soil presented no significant differences for different slopes. As with the runoff loss leakage, the slope affected the occurrence of the surface loss of soil. The surface soil loss did not occur at small rainfall intensity when slopes changed between 5° and 20° but did occur at the 30° slope condition. For the surface soil loss cases, the surface loss rate of soil first increased and then decreased with an increase of slope to a maximum at slopes of 15° or 20°. Accordingly, a critical slope condition of maximal surface soil loss also existed.
As shown in Figure 5, the surface loss ratio and the underground pore fissure leakage ratio of soil were similar to variations in the runoff leakage loss. The underground leakage loss of soil dominated the total soil loss under small rainfall intensity conditions on karst slopes. The underground leakage ratio of soil accounted for more than 44% of the total soil loss on karst slopes. Similarly, the surface loss ratio of soil dominated the total loss under greater rainfall intensity conditions on karst slopes. The surface loss ratio of soil held more than 74% of the total soil loss on karst slopes. However, the underground leakage ratio of soil was less than 26%.
3.3 Soil and water loss for natural rainfall events
3.3.1 Runoff loss
A total of seven rain events, with a rainfall range of 9.3–104.3 mm, were observed during the runoff plot monitoring period. Figure 6 shows the results of surface runoff yield and soil loss recorded in each runoff field. With an increase of rainfall, surface runoff yields generally increased (Figures 6 and 7). When rainfall increased from 9.5 mm (Sep 9) to 104.3 mm (Jun 5), the surface runoff yield in those runoff fields increased by 3.7–7.5 times. In individual rainfall events, Runoff Field III (25° slope and 5.2-m slope length) had the greatest surface runoff yield increase, from 2.1 to 15.1 L m−2; however, Runoff Field VIII (3° slope and 20-m slope length) had the lowest increase. Although rainfall to a certain extent determines the runoff yield on slopes, the gradient and length of the slope influence runoff yield. As a result, the greatest runoff did not occur in the slope with larger slope degree or slope length under the same rainfall conditions, and vice versa.
The surface runoff coefficients in each runoff field at different slope conditions are shown in Figure 7. Variation trends of surface runoff coefficients were similar under different rainfall conditions. With an increase in slope degree (3–45°), the surface runoff coefficient exhibited two peaks at slope degrees of 5° and 25°, with ranges of 0.12–0.63 and 0.13–0.88, respectively. The surface runoff coefficient at 25° slope degree showed the greatest values. The critical slope degree condition of maximal surface runoff coefficient was 25°. Additionally, the rainstorm event with a rainfall of 104.3 mm had the greatest surface runoff coefficient, with a range of 0.34–0.88. Except for the greater rainfall, the surface runoff coefficient under different slope degree conditions had little fluctuation.
3.3.2 Soil loss
As shown in Figure 6, surface soil loss exhibited different variations than did surface runoff yield. There was no obvious relationship between the surface soil loss and the rainfall. For individual rainfall events, Field IV (45° slope and 6.5-m slope length) had the greater surface soil loss, with a range of 0.25–0.89 g m−2, and Field I (5° slope and 4.7-m slope length) had the smaller range (0.02–0.27 g m−2). Slope degree had a greater effect on the surface soil loss than had slope length. Compared with runoff loss, the surface soil loss in each runoff field was not determined by rainfall but by slope degree.
Figure 8 shows the surface soil loss rate in each runoff field at different slope conditions. There was a critical slope value for the influence of the slope on the surface soil loss rate on yellow soil slopes. With the increase of the slope degree, the surface soil loss rate had an overall growth trend. When the slope degree changed between 3° and 15°, the surface soil loss rate exhibited an increasing trend, with maxima at 15° of the slope degree, with a range of 0.21–0.69 g·m−2·hr−1. When slope degree increased from 15° to 25°, the surface soil loss rate rapidly dropped. There was a critical slope degree (15°) for the maximal surface soil loss rate.
4 DISCUSSIONS
4.1 Surface erosion on karst slopes
The surface soil erosion was obviously different on karst slopes compared with the nonkarst areas. Previous studies have indicated that soil surface components (vegetation, rock outcrops, fractures, and soil crusts), topographic position, antecedent soil moisture, and rainfall characteristics affect the surface erosion process (Cerdà, 2001; Chen et al., 2011; Li et al., 2011; Poesen, Torri, & Bunte, 1994). Surfaces of karst slopes have been characterized by discontinuous regolith, thin soils, patchy vegetation, and large areas of bare rock (Li et al., 2017). The degree of bare bedrock typically determines the surface soil erosion on karst slopes (Dai et al., 2015), which is similar to the surface roughness, which reflects the morphological resistance characteristics of surface topography in nonkarst areas (Zhao, Wang, Liang, Wang, & Wu, 2013). This also determines the difference in surface erosion between karst slopes and nonkarst slopes.
In our study, critical slope conditions (15–25°) existed for the maximal surface runoff depth and the surface soil loss rate because the runoff and sediment yields increased with the increase of slope degree and then decreased and reached the maximum at slopes of 15–25°. The existence of a critical slope is of course not because of the decrease of runoff but is closely related to the overland flow velocity and the movement mechanisms of sediments in runoff. Moreover, the critical slope of hillslope erosion mainly depends on the flow rate, water depth, and particle sizes in the overland flow (Hu & Jin, 1999). Additionally, surface erosion on karst slopes only formed at steep slopes (30°) or greater rainfall intensities (over 50 mm hr−1). Relevant research has also reported that surface runoff and soil loss are mainly created by heavy rainfall storms with a rainfall intensity of over 30 mm hr−1 (Peng & Wang, 2012). Compared with nonkarst areas (Maetens et al., 2012), the results of investigations of large runoff fields (Chen, Yang, Fu, He, & Wang, 2012) have shown that the amounts of surface runoff and soil loss on karst slopes were very small, that event surface runoff coefficients were less than 5% on hillslopes with different land uses and that the soil erosion moduli were mostly weak (lower than 30 t·km−2·a−1). This was because bare bed rocks with many joints, fissures, pores, or holes could percolate and absorb a large amount of rainwater and could also absorb some rainwater after long-term weathering (Xiong, Li, & Long, 2012); moreover, the intercepting and gathering effects of bedrock reduced the surface runoff velocity (Bou Kheir, Abdallah, & Khawlie, 2008).
Some studies have also suggested that rainfall storms with large antecedent precipitation could also produce large runoff and soil loss (Peng & Wang, 2012). However, the results of our study indicated that both the surface runoff and soil loss dominated the total water and soil loss on karst slopes for rainfall storms with rainfall intensities over 50 mm hr−1, with a surface runoff ratio of over 55% and a surface loss ratio of over 74% (Figures 3 and 5). Relevant studies have also proven that the presence of rock fractures promotes infiltration in limestone karst landscapes, whereas bare patches and rock outcrops act as sources for runoff (Lange et al., 2003; Li et al., 2011). As a result, greater soil and water are eroded in the form of surface runoff under rainstorm conditions. The soft and hard interfaces of rock and soil may be the main channel for moving water and soil particles (Feng, Chen, Zhang, Nie, & Wang, 2011), and soil eroded in karst caves was reported to come from the interface of rock and soil (Wei et al., 2016). The movement of soil along the interfaces of soil and rock may thus be one of the main reasons for rocky desertification; this requires further study.
4.2 Underground leakage loss on karst slopes
Soil underground leakage loss is a unique type of soil and water loss in karst areas (Peng et al., 2016; Wei et al., 2016). Underground systems of karst slopes are characterized by the existence of irregular solution networks of pores, fissures, fractures, and conduits of various sizes and forms karstified from limestone. Underground leakage loss can occur through subsurface pores, fissures, holes, and underground leakage phenomena, which underestimate the amount of soil erosion in karst areas (Li et al., 2016; Yang, Tang, & Zhou, 2011). Soil underground leakage loss is mainly enhanced, directly or indirectly, by those karst features in karst areas (Mohammadi, Raeisi, & Bakalowicz, 2007). In this study, the underground leakage loss of water and soil always occurred at different slope and rainfall intensity conditions in karst regions, which indicates that underground leakage loss is more prone to occur than surface erosion is.
Results of field monitoring showed that the loss leakage of soil through underground channels is the major form of soil and water loss in the rocky karst desertification area in Chenqi village and is more serious than the surface loss of soil (Yang et al., 2011). This is attributed to the fact that many surface tundishes, sinkholes, and karrens have been developed in the karst areas of carbonate rock. However, the results of our study indicated that underground runoff and underground leakage loss of soil dominated the total runoff and soil loss on karst slopes at small rainfall intensities, showing an underground runoff ratio of over 90% and an underground soil leakage ratio of over 44% (Figure 4). However, at greater rainfall intensity, underground runoff and underground leakage loss of soil were greatly lower than at the surface (underground runoff ratio lower than 45% and soil leakage ratio lower than 26%; Figure 5). These results agree with results of isotope 137Cs tracing on slopes in karst valley areas (Wei et al., 2016); the soil surface loss (75%) was higher than the underground leakage (25%). Additionally, soil particles in karst slopes had a tendency for underground loss with rain water along negative terrain, although the amount of soil erosion was slight, which indicated that soil erosion was dominated by surface erosion (Feng et al., 2011).
Indeed, some studies have proven that soil particles are carried by rainfall runoff leakage into well-developed karst cracks. These particles interact with each other: They are sifted, rearranged, and eventually piled up (Tang et al., 2016). The different effects of fissures filled with soils and open fissures and the mechanisms of water and soil transport in underground pore fissures require further study.
5 CONCLUSIONS
Surface erosion and underground leakage simultaneously exist on slopes in karst regions. The plot-scale characteristics of surface erosion and underground leakage loss of yellow soil on karst slopes were studied through simulated rainfall tests and field runoff plot monitoring. Surface erosion only formed at steep slopes (30°) or greater rainfall intensities (over 50 mm hr−1). The underground leakage loss of water and soil always occurred at different slope and rainfall intensity conditions in karst regions. For small rainfall intensity cases (lower than 80 mm hr−1), underground runoff and underground leakage loss of soil dominated the total runoff and soil loss on karst slopes, with underground runoff yield ratios of over 90% and underground soil leakage ratios of over 44%. For greater rainfall intensity cases, surface runoff and underground runoff shared the total runoff on karst slopes, with ratios of over 55% and lower than 45%, respectively; however, surface soil loss dominated the total soil loss, with a surface loss ratio of over 74%.
Rainfall intensity had significant effects on surface runoff depths and surface soil loss rates but insignificant effects on underground leakage loss. The surface runoff depth and surface soil loss rate increased with the increase of rainfall intensity; however, that of underground leakage had no obvious changes. Although rainfall determined the runoff yields on slopes to a certain extent, slope degree and slope length influenced the runoff yield and slope degree had a greater effect on the surface soil loss than slope length had. Slope degree had certain effects on surface erosion but little influence on underground leakage on karst slopes. Critical slope conditions (15–25°) existed for the maximal surface runoff depth and the surface soil loss rate. As seen, controlling the formation of underground leakage is the key to prevent soil water loss in karst areas. Otherwise, the development of rocky desertification will be intensified. However, the effect of rainfall intensity and slope degree on surface erosion or underground leakage loss lies on the interfaces of soil and rock and the connectivity of fissures.
ACKNOWLEDGMENTS
This work was supported by the National Key Research and Development Program of China (2016YFC0502604), the National Natural Science Foundation of China (41671275, 41461057), Shuangchuang Talent Plan of Jiangsu Province, the Major Project of Guizhou Province (Qian Ke He Major Project [2016]3022), and the First-class Discipline Construction Project in Guizhou Province (GNYL[2017]007).
CONFLICT OF INTEREST
There is no conflict of interest regarding this paper.
