Identification of priority areas for soil erosion control based on minimum administrative units and karst landforms in karst areas of Guizhou

Abstract

Soil erosion is one of the most serious ecological threats in karst areas of Southwest China. The identification of priority areas for remediation and its driving factors is essential to improving the efficiency of prevention and control. The present study systematically considered natural and socio-economic factors not involved in the revised universal soil loss equation (RUSLE) model, and determined priority areas for soil erosion management based on minimum administrative units and karst landforms. Then, the driving factors were identified by using geographic detector. The results showed that the priority areas were mainly concentrated in the southwest, southeast and northeast, overlapping with the severely eroded areas (Erosion rate=45.79 t·ha⁻¹·a⁻¹). Gradient risk zones had geomorphological differences, but the most eroded zones were all controlled by bedrock exposure rates, elevation, or slope position. The spatial correlation and high erosion rate of priority areas provided opportunities to optimize the efficiency and cost of control. Driving factors were affected by karst landforms. The explanation power of slope position on soil erosion was higher in the peak cluster depressions and karst basins with small undulations ( $\bar{q} = 11.85 %$ ), while the karst gorges, trough valleys and plateaus with large undulations gradually decreased ( $\bar{q} = 6.01 %$ ). The interaction of driving factors will enhance the explanatory power for soil erosion. Among them, the repetition rate of elevation was 60%, and the repetition rate of lithology and development index was 40%. This study provides useful information for identifying and managing priority areas for soil erosion control, and enriches the theory of soil and water conservation in karst areas.

Keywords

Minimum administrative units karst landforms priority areas driving factors soil erosion

I Introduction

Southwest China has one of the three major karst concentrated distribution areas in the world, with an area of more than 300 km², accounting for about 1/3 of China’s total land area (Peng and Wang, 2012; Li et al., 2016). In recent decades, due to uncontrolled human activities and specific geochemical processes, soil erosion and its derived exposed bedrock have become a serious environmental problem in the karst area of southwest China, threatening the ecological security of the Yangtze River and Pearl River basins (Cao et al., 2021; Jing et al., 2023). The soil and water resources in karst areas are highly precious. Under similar climate background, the soil formation rate may be dozens of times higher in non-karst areas than it is in karst areas (Jiang et al., 2014; Feng et al., 2016). Therefore, once the soil layer is lost, it will be extremely difficult to recover (Cheng et al., 2017; Tang et al., 2019). The identification of priority areas for soil erosion control and its remediation measures plays an important role in optimizing efficiency and cost (Betelhem Fetene et al., 2022; Zhang et al., 2010). However, compared with those on the Loess Plateau and the black soil area, studies on priority areas for soil erosion control in karst areas have been limited. Systematic research on the driving mechanism underlying soil erosion in priority areas is also scant.

Towns are the lowest level units in China’s administrative divisions. It is not only the carrier and coordinate for formulating policies and planning projects, but also the implementation unit of soil and water loss control in small watersheds (Zhao, 2017; Pan and Wen, 2014; Wen and Deng, 2020). However, at present, priority areas of soil erosion control are mainly identified by erosion rate and its derived indicators or related control factors (Tamene et al., 2017; Asmamaw and Mohammed, 2019; Rafiei et al., 2020; Panagos et al., 2021). Because of the lack of consideration of administrative boundaries and spatial relevance, the feasibility of these methods is low in actual management and planning. Recently, spatial clustering has been increasingly used to identify priority areas of soil erosion in order to provide efficient and low-cost management solutions (Vijith et al., 2018; Wen, 2020). However, its applicability and effectiveness in karst areas are still unclear, especially at the scale of the minimum administrative unit.

Karst landforms are unique products formed under special geological and climatic background conditions. Its types include karst gorge, peak cluster depressions, plateaus, basins, trough valleys, etc. (Gutiérrez and Gutiérrez, 2016). The internal environment of different karst landforms has obvious differences (vegetation, slope, soil thickness, exposed bedrock, etc.). Landform plays a macro-control role on surface processes and phenomena by influencing geographical environmental factors (Corenblit et al., 2011; Lara et al., 2018). As a surface process, soil erosion shows obvious differences in erosion characteristics in different karst landforms (Bridge and Demicco, 2008; Zeng et al., 2018; Zhou et al., 2020; Li et al., 2021). However, the current research on soil erosion in karst areas has mostly focuses on the process mechanism, spatiotemporal evolution and influencing factors at the slope, watershed or regional scale (Dai et al., 2017; Feng et al., 2020; Zhou et al., 2020). Environmental differences between karst landforms are rarely considered, which makes it difficult to gain a complete understanding of soil erosion characteristics and its drivers in karst areas.

Elucidating soil erosion drivers and its contribution rates is a prerequisite for the implementation of targeted control measures. Soil erosion is usually affected by various natural and anthropogenic factors, such as rainfall erosivity, vegetation, soil properties, and land management practices (Eniyew et al., 2021; Senanayake et al., 2022). However, unlike non-karst areas, the bedrock in karst areas is soluble carbonate rock, which is easy to be dissolved by rainfall, thus making the surface environment complex and variable (Zeng et al., 2018). This makes the soil erosion mechanism in karst areas more complicated (Dai et al., 2017; Yang et al., 2022). Peng and Dai (2022) found through simulated rainfall experiments that bedrock exposure rate (Due to topsoil loss) was an important factor affecting the surface erosion rate in karst areas (Critical value = 30%). Different karst lithologies have different effects on the fertility, water retention and erosion resistance of surface soils, and these properties are closely related to soil erosion processes (Wang et al., 2019). The high erosion and low formation rate of soil in karst areas lead to generally thinner topsoil (Zhang et al., 2022b) and the soil layer thickness is directly related to the rate and amount of erosion (Fu et al., 2011; Huang et al., 2021). In addition, slope position, aspect, altitude and socio-economic development also have an important impact on soil erosion (Lan et al., 2022b; Huang et al., 2019; Niu and Shao, 2020; Zeng et al., 2017). Compared with natural factors, human interference is easier to adjust and control (Guo et al., 2021). Therefore, strengthening the understanding of human interference will help us better understand the soil erosion in this region. Notably, these factors tend to interact with each other and are expected to be influenced by geomorphological features (Gao et al., 2018). However, few studies have focused on the interactive effects of driving factors on soil erosion in different karst landforms.

Empirical regression equations and geographic detector are common methods for quantifying the drivers of soil erosion (Ke and Zhang, 2022; Gao and Wang, 2019b). However, the application of empirical regression equation has limitations because of the need for a large number of input data and cumbersome parameter calibration. By contrast, geographic detector has simpler data and operational requirements, and allow quantitative analysis of factor interactions. Therefore, it has achieved a series of research results in the analysis of soil erosion drivers (Li et al., 2022; Xu and Zhang, 2020; Liang and Fang, 2021). However, many studies have overlooked the fact that erosion estimation models (e.g., the revised universal soil loss equation (RUSLE), water erosion prediction project model (WEPP), and Soil & Water Assessment Tool (SWAT)) already take into account a number of environmental factors (Land use type, slope, etc.). This means that some driving factors are repeatedly detected (Li et al., 2022; Chu et al., 2020). This may compromise the accuracy of the study results, which in turn affects the efficiency and cost of soil erosion control.

The present study will systematically consider natural and socioeconomic factors that are not involved in the RUSLE model. Then, based on the minimum administrative units and karst landform types, the spatial clustering method and geographic detector will be used to identify the priority areas for soil erosion control in the karst area of Guizhou in 2018. The driving factors and the interactions of soil erosion will be quantitatively analyzed in the priority area. Our specific objectives include: revealing the spatial distribution characteristics of soil erosion in karst areas of Guizhou; identifying priority areas for soil erosion control; quantifying the drivers and its interactions of soil erosion under different karst landform types in priority areas. The results of this research will provide valuable technical and theoretical support for soil erosion management to mitigate the widespread and severe hydraulic erosion in karst areas of Southwest China.

II Materials and methods

1 Description of the study area

Guizhou Province is located in the center of the southwest karst region (Figure 1), and it is one of the regions with the most typical karst development and most serious risks of rock desertification hazards in the world. According to its landform, lithology, geological structure conditions, and field survey, Yuan (2014) divided Guizhou Province into five karst macro landform areas (i.e., karst trough valleys, peak cluster depressions, karst plateaus, karst basins, and karst gorges) and one non-karst area (Figure 1(a)). On the basis of this geomorphic division combined with lithological data, we extracted the distribution area of soluble rock in this karst area and considered this to be the study area (Figure 1(b)).

Figure 1.

Study area: Location and features. [Note: a is a karst macro landform type division; b is reclassified karst landform type zoning based on the distribution of soluble rocks (i.e., the study area)].

The study area (10.91 × 10⁴ km²) accounts for 61.9% of the total area of the province. It has an average altitude of 1171 m and has a subtropical humid monsoon climate, which is warm and humid. Precipitation mainly occurs during May to October, with the average perennial precipitation being between 963 mm and 1460 mm. The main soil types are yellow soil and lime soil, and the main vegetation is subtropical evergreen broad-leaved forest. Its ecological environment is fragile, with the discord between humans and the land being prominent; long-term unsustainable human activities have caused serious soil erosion and exposed a large area of the bedrock. The study area is located in the watershed that serves as a main water supply area of the Pearl River and Yangtze River and thus has a prominent ecological status.

2 Data

Rainfall data were obtained from the National Meteorological Science Data Center (http://data.cma.cn/). These data, including the rainfall data from 46 rainfall stations in Guizhou Province in 2018 (Figure 1), were used to calculate rainfall erosivity. Soil data, including soil texture (i.e., sand, silt, and clay content) and soil organic carbon content, were obtained from the China Soil Database (http://vdb3.soil.csdb.cn/) and were used to calculate soil erodibility factors. DEM data were obtained from the geospatial data cloud (https://www.gscloud.cn/), with a spatial resolution of 30 m; the elevation, slope, slope length, and aspect of the study area were extracted using ArcGIS.

Data on the lithology, land-use type (2018; 30 m × 30 m), and administrative divisions of Guizhou Province were obtained from the Resource and Environment and Data Center (http://www.resdc.cn/). Data on the slope position, bedrock exposure rate, and soil thickness in the Guizhou karst areas were all derived from the database of the third rocky desertification monitoring results in the karst region of China. The soil layer thickness was calculated as the average thickness of the patch soil layer in the field investigations and divided into four grades: medium, thin, thin, and extremely thin. The per capita gross domestic product (GDP) and population density data of Guizhou Province in 2018 were derived from the Guizhou Statistical Yearbook and used to build the regional development index.

3 RUSLE model

The RUSLE model, which comprehensively considers the main factors affecting soil erosion, has been widely applied in soil erosion prediction and soil and water conservation planning (Olorunfemi et al., 2020); it can be defined as follows:

A = R \times K \times L \times S \times C \times P

where A is the amount of soil loss (t·ha⁻¹), R is the rainfall erosivity coefficient (MJ·mm·h⁻¹·ha⁻¹), K is the soil erodibility factor (t·ha·h·mm·MJ⁻¹·ha⁻¹), LS is the slope length and slope factor (dimensionless), C is the vegetation cover and management factor (dimensionless), and P is the factor of soil and water conservation measures (dimensionless).

Rainfall erosion force (R)

At present, there are various types of rainfall erosion force calculation models commonly used in the southwest karst region, including the erosivity estimation model based on daily or monthly rainfall established by power law, exponential or nonlinear multi-parameter equations (Silva, 2004; Wischmeier and Smith, 1965; Richardson et al., 1983; Xie et al., 2016; Yu and Rosewell, 1996; Shi et al., 2006; Zhang et al., 2002). These rainfall erosion estimation models are widely used in the karst region of southwest China and have been recognized by many researchers (Cao et al., 2020a; Zhu et al., 2021; Dai et al., 2015; Pan et al., 2022; Gao and Wang, 2019b). Zhu et al. (2021) compared the applicability of the above model in the southwest karst region using relative error (RE) and absolute relative error (ARE) as measurement standards. The results show that the derivative version based on Yu and Rosewell (1996). Xie et al. (2016) has better computational performance for the R factor than other estimation models (RE=0.54%; ARE=8.27%). At the same time, 12 mm was verified as the optimal threshold for erosive rainfall in the karst region of southwest China. The model and threshold will be used in this study to calculate the R factor, and the formula is as follows:

R_{i} = 0.2686 [1 + 0.5412 \cos (\frac{π}{6} j - \frac{7 π}{6})] P_{d}^{1.7265}

where R_i is the monthly rainfall erosive force (MJ·mm·h⁻¹·ha⁻¹) and P_d is the daily erosive rainfall (≥12 mm), j is from January to December.

Soil erodibility factor (K)

K was estimated using the Erosion Productivity Impact Calculator model based on soil texture and soil organic matter data:

K = {0.2 + 0.3 e^{[- 0.0256 W_{d} (1 - W_{i} / 100)]}} \times {(\frac{W_{i}}{W_{i} + W_{t}})}^{0.3} \times [1 - \frac{0.25 W_{c}}{W_{c} + e^{(3.72 - 2.95 W_{c})}}] \times [1 - \frac{0.25 W_{n}}{W_{n} + e^{(- 5.51 + 22.9 W_{n})}}]

W_{n} = 1 - \frac{W_{d}}{100}

Where W_d is the sand content (%), W_i is the powder content (%), W_t is the clay content (%), and W_c is the organic carbon content (%).

Slope length and slope factors (LS)

L and S are quantified by calculations of topography and flow accumulation (Moore and Burch, 1986). Zhang et al. (2013) optimized the traditional flow path and cumulative cell length method (FCL) by means of a channel network and convergent flow algorithm, and developed a convenient LS factor calculation tool (LS-TOOL) using C language based on the NET environment. LS-TOOL locates soil erosion and sedimentation zones, channel networks, and confluence areas through slope faults identified from digital elevation model (DEM) (Gao and Wang, 2019b). Compared with the traditional unit contribution area method (UCA) or the flow path and cumulative cell length method (FCL), LS-TOOL is more suitable for LS factor extraction of complex and fragmented landforms (Zhang et al., 2013). The expressions for the L and S factors are as follows (Mccool et al., 1987, 1989):

L = {(\frac{λ}{22.13})}^{m}

m = \frac{β}{β + 1}

β = \frac{\sin θ}{[3 \cdot {(\sin θ)}^{0.8} + 0.56]}

S = 10.80 \cdot \sin θ + 0.03 (s l o p e l e n g t h > 4.6 m, S < 9 %)

S = 16.80 \cdot \sin θ - 0.50 (s l o p e l e n g t h > 4.6 m, S \geq 9 %)

S = 3.0 \cdot {(\sin θ)}^{0.8} + 0.56 (s l o p e l e n g t h < 4.6 m)

where λ is the slope length in the horizontal direction, m is the variable length-slope index, β is the factor that changes with slope, and θ is the slope angle. LS-TOOL calculates λ for each slope section by means of a unit area contribution algorithm. At the same time, it can also be the accumulated area of the deposition location or the cut-off point where the accumulated slope reaches the specified threshold. The minimum set of its cumulative area thresholds is the same as the unit grid area. Details of LS-TOOL can be found in Zhang et al. (2013).

Cover-management and support practice factor (C and P)

Referring to many previous studies in the southwest karst areas (Zhang et al., 2022a; Gao and Wang, 2019b), we used the assignment method to calculate the C and P factors. C-factors for paddy land, dry land, forest, open forest, shrub, grassland, water body, construction land, and bare rock were assigned values of 0.1, 0.22, 0.006, 0.01, 0.01, 0.04, 0, 0, and 0, respectively; P-factors for paddy land, dry land, forest, open forest, shrub, grassland, water body, construction land, and bare rock were assigned values of 0.15, 0.4, 1, 1, 1, 1, 0, 0, and 0, respectively.

In accordance with the simulation results and the grading standard of soil erosion intensity in the karst area (Ministry of Water Resources of the People’s Republic of China, 2009), the soil erosion degree in the study area was divided into six levels: slight, light, moderate, high, very high, and severe, corresponding to soil erosion of <0.5, 0.5–3, 3–15, 15–30, 30–60, and >60 t·ha⁻¹·a⁻¹, respectively. The soil erosion grade values of different karst landform regions and administrative regions were extracted using ArcGIS.

4 Spatial clustering

Getis-Ord Gi^* is widely used to analyze spatial attribute data distribution (Cho and Park, 2013). It calculates a Gi^* statistic for each feature in the dataset, provides the location of spatial clustering of high or low value features through the resulting z score and p value, and calculates the statistically significant hot and cold spots with high and low values and surrounded by other high or low values. Here, we used this method to identify the priority areas for soil erosion control at the minimum administrative unit scale, as follows:

G_{i}^{*} = \frac{\sum_{j = 1}^{n} w_{i, j} x_{j} - \bar{x} \sum_{j = 1}^{n} w_{i, j}}{S \sqrt{(\sum_{j = 1}^{n} w_{i, j}^{2} - {(\sum_{j = 1}^{n} w_{i, j})}^{2}) / n - 1}},

\bar{x} = \frac{\sum_{j = 1}^{n} x_{j}}{n},

S = \sqrt{\frac{\sum_{j = 1}^{n} x_{j}^{2}}{n} - {(\bar{x})}^{2}},

where G^* statistics is the Z score, X_i is the attribute value of factor j, W_i,j is the spatial weight between factors i and j, and n is the total number of elements.

5 Geographic detector

Geographic detector, which is widely used in the fields of the natural and social sciences (Ren et al., 2022; Yue and Zhu, 2019), is a statistical model used to detect spatial heterogeneity and reveal underlying driving forces (Wang et al., 2016). Here, we analyzed the degree of contribution of environmental factors to soil erosion and the multifactor gradient areas at risk of soil erosion by using the three functions of the model: factor detection, risk detection, and interaction detection.

Factor detection

We detected the spatial differentiation of soil erosion and the contributions of control factors to soil erosion, the sizes of which are measured as q:

q = 1 - \frac{\sum_{h = 1}^{L} N_{h} {σ_{h}}^{2}}{N σ^{2}} = 1 - \frac{S S W}{S S T},

S S = \sum_{h}^{L} N_{h} {σ_{h}}^{2},

S S T = N σ^{2},

where h = 1, 2…; L is the stratification of independent variable X;

N_{h}

and N are the numbers of units in the layer H and the whole area, respectively;

σ_{h}

and

σ

are the variances in the layer H and Y values of the whole region, respectively; SSW is the intralayer variance sum; SST is the total regional variance; and q is the explanatory power of the independent variable for the dependent variable, with the range of [0,1]. The closer the q is to 1, the stronger the explanatory power of the independent variable is for the dependent variable. After being subjected to a simple transformation, the q value has a noncentral distribution, and the geographical detector can test the significance of the q value. We used this model to detect the contribution of the driving factors to the spatial differentiation of soil erosion.

Risk detection

Determine whether the means of the properties between two subregions were significantly different by t statistics, the calculation formula is as follows:

t = \frac{{\bar{Y}}_{h = 1} - {\bar{Y}}_{h = 2}}{{[\frac{V a r (Y_{h = 1})}{n_{h = 1}} + \frac{V a r (Y_{h = 2})}{n_{h = 2}}]}^{1 / 2}}

where

{\bar{Y}}_{h}

is the average value of soil erosion in the subregion h,

n_{h}

is the number of samples in the subregion h, and Var is the variance. The t statistic is partially subject to a Student t distribution, where the degree of freedom is calculated as:

d f = \frac{\frac{V a r ({\bar{Y}}_{h = 1})}{n_{h = 1}} + \frac{V a r ({\bar{Y}}_{h = 2})}{n_{h = 2}}}{\frac{1}{n_{h = 1} - 1} {[\frac{V a r ({\bar{Y}}_{h = 1})}{n_{h = 1}}]}^{2} + \frac{1}{n_{h = 2} - 1} {[\frac{V a r ({\bar{Y}}_{h = 2})}{n_{h = 2}}]}^{2}}

Zero hypothesis $H_{0}$ : ${\bar{Y}}_{h = 1} - {\bar{Y}}_{h = 2}$ . If $H_{0}$ is rejected at confidence level α, the attribute mean between the two subregions is considered to be significant. We used this model to identify multifactor gradient risk areas for soil erosion by analyzing whether there were significant differences in the amount of soil erosion between the layers of driving factors.

Interaction detection

We use it to evaluate the influence of interaction between two environmental variables on spatial differentiation of soil erosion; the judgment basis is shown in Table 1.

Table 1.

Types of interaction between independent variables.

Criterion	Interaction
q (X1∩X2) < Min (q (X1), q (X2))	Nonlinear weakening
Min (q (X1), q (X2)) < q (X1∩X2) < Max (q (X1), q (X2))	Single factor nonlinear weakening
q (X1∩X2) > Max (q (X1), q (X2))	Two-way enhancement
q (X1∩X2) = q (X1) + q (X2)	Independent
q (X1∩X2) > q (X1) + q (X2)	Nonlinear enhancement

Note: X1 and X2 represent different independent variables.

6 Selection and testing of identification factors of gradient risk zones

According to the characteristics of the soil erosion processes in karst areas, seven natural and socioeconomic factors, namely elevation, slope position, aspect, lithology, soil thickness, the development index, and bedrock exposure rate, were selected to identify the gradient areas at risk of soil erosion.

On the basis of the evaluation method of the socioeconomic development index recommended by the United Nations Environment Programme (Jin et al., 2020), we used MinMaxScaler standardization and the entropy weight method to calculate the development index, which reflects the impact of socioeconomic development on soil erosion.

We first used MinMaxScaler to standardize GDP and population density:

Z_{i, j} = \frac{x_{i, j} - \min {x_{1}, x_{2 j}, \dots x_{m j}}}{\max {x_{1}, x_{2 j}, \dots x_{m j}} - \min {x_{1}, x_{2 j}, \dots x_{m j}}}

Z_{i, j} = \frac{\max {x_{1}, x_{2 j}, \dots x_{m j}} - x_{i, j}}{\max {x_{1}, x_{2 j}, \dots x_{m j}} - \min {x_{1}, x_{2 j}, \dots x_{m j}}}

Entropy value $E_{j}$ was then calculated on the basis of the standardized data:

E_{j} = - k \sum_{i = 1}^{m} Z_{i, j} \times \ln Z_{i, j}

k = \frac{1}{I n (m)}

We next calculated the index weight; W is the index weight:

W_{j} = \frac{1 - E_{j}}{\sum_{j = 1}^{n} (1 - E_{j})}

The greater the weight, the greater the contribution of the index to the measurement result. The development index was then calculated according to the index weight determined by the entropy weight method:

S_{i} = \sum_{j = 1}^{n} W_{j} \times Z_{i, j}

Finally, we validated the calculation results. The correctness of the step calculation results (MinMaxScaler standardization and weighting) and the final calculation results (regional development index) were tested using the Pearson correlation coefficient and the partial correlation, respectively. When r > 0 and r < 0, the correlation between the two variables is positive and negative, respectively. |r| reflects the degree of linear correlation between the two variables. When r = 1, the two variables have a completely consistent trend of change; in other words, one variable can accurately reflect the change in the other variable. These methods have been widely used in variable analysis and data validation (Jiang et al., 2004; Zuo and Kita, 2012). The Pearson correlation coefficient and partial correlation were calculated using R (version 4.1.3; https://cran.r-project.org/web/packages/Hmisc/; https://cran.r-project.org/web/packages/ppcor/).

The validation results showed that the correlation values between the calculated results (standardization, weight, regional development index) and the observed values were all equal to 1 (p < 0.01). This suggests that the calculation process and results of the regional development index are accurate, and it can well reflect GDP and population density indicators.

We next performed a multicollinearity test on the seven selected factors of natural and socioeconomic development. The results revealed that the values of the variance inflation factors of the selected factors were all <2, with tolerances >0.1. No collinearity was noted, the selected environmental factors could be used to identify the gradient risk area of soil erosion.

III Results

1 Spatial pattern of soil erosion

Soil erosion grades in the study area were predominantly moderate grade, and the average erosion rate was 16.49 t·ha⁻¹·a⁻¹ (Figure 2(a)). The spatial heterogeneity of soil erosion intensity was very pronounced. Slight and light erosion was mainly distributed in the plateaus and trough valleys in the central and northeastern parts of the study area, accounting for 22.37% of the total soil loss area. Moderate erosion was distributed discretely in the study area, accounting for 33.74% of the total soil loss area. Intense and above erosion was mainly distributed in the gorges and basins in the southwest of the study area and in the plateaus and trough valleys in the eastern and northeastern margin of the study area, accounting for 43.89% of the total soil loss area.

Figure 2.

Spatial distribution (a) and spatial clustering (b) of soil erosion.

2 Priority areas for soil erosion control

Spatial clustering

Spatial clustering based on the minimum administrative unit (i.e., town) was used to identify priority areas for soil erosion control in the study area. The results showed that the priority areas with statistical significance accounted for 14.83% of the total area (Figure 2(b), Confidence coefficient=90%). Compared with the areas with intense and above erosion, the priority areas demonstrated a higher degree of spatial agglomeration. Priority areas were continuously distributed in the southwest, southeast, and northeast of the study area, and were mainly located in the plateaus, basins, and trough valleys. It overlapped the areas with relatively strong erosion, and the average soil erosion rate was 45.79 t·ha⁻¹·a⁻¹, which belongs to extremely intense erosion.

Multifactor gradient risk zones

Based on karst landform types, the geographic detectors were used to detect multifactor gradient risk zones in the priority areas identified by spatial clustering. As shown in the detection results (Table 2, Confidence coefficient=95%), there are significant differences in gradient risk zones for different karst landform type areas. But they all belong to extremely intense levels of erosion (Figure 3). Soil erosion rate in karst trough valleys and plateaus first increased and then decreased with elevation, and the turning point of erosion rate was the elevation risk zone (Figure 4); soil erosion rates in the other karst landform areas fluctuated with elevation, and the elevation risk zone was the highest peak (Figure 4). Slope position risk zone of the karst basin was ridge, whereas the other karst landforms are hillsides. Soil thickness was closely related to bedrock exposure rate; the higher the bedrock exposure rate, the thinner the soil layer, and vice versa. For instance, the bedrock exposure rates of the karst gorges were 82%–100%, and therefore, their corresponding soil thickness risk zone was extremely thin (<10 cm).

Table 2.

Multifactor gradient risk zone.

	Karst gorge	Peak cluster depression	Karst plateau	Karst basin	Karst trough valley	Karst gorge
Soil thickness	Extremely thin	Thin	Thin	Thin	Extremely thin	Extremely thin
Aspect	South slope	South slope	South slope	South slope	South slope	South slope
Slope position	Downhill	Mid-slope	Mid-slope	Ridge	Uphill	Downhill
Elevation	950 m–1131 m	1315 m–1525 m	1525 m–1785 m	770 m–950 m	91 m–569 m	950 m–1131 m
Lithology	Limestone sandwich clastic rock combination	Continuous limestone combination	Mix of limestone and dolomite	Limestone and clastic rock are interleaved	Limestone and clastic rock are interleaved	Limestone sandwich clastic rock combination
Development index	0.7%–3.7%	6.5%–10.1%	6.5%–10.1%	3.7%–6.5%	0.7%–3.7%	0.7%–3.7%
Bedrock exposure rate	82%–100%	52%–62%	82%–100%	72%–82%	82%–100%	82%–100%

Figure 3.

Soil erosion rate in risk zones.

Figure 4.

Change of soil erosion rate in elevation risk zones.

3 Quantitative attribution of soil erosion in different karst landform areas

Driver detection

To avoid repeated detection of driving factors, and thus improve the targeting and effectiveness of prevention and control measures, we focused on the impact of natural and socioeconomic factors not included in the RUSLE model on soil erosion. The results showed that lithology, soil thickness, aspect, slope position, and elevation significantly affected soil erosion in karst landform areas (p < 0.05). The development index and bedrock exposure rate demonstrated nonsignificant effects on soil erosion in karst basins and trough valleys, respectively (p > 0.05; Figure 5).

Figure 5.

Distribution of q values in different karst landforms. (Note: The square indicates that the influence of the control factor on soil erosion is not significant, p > 0.05; the sphere indicates that it has a significant effect, p < 0.05).

The dominant drivers of soil erosion and its explanatory power were different in different karst landform areas. Development index had a strong explanatory power for soil erosion in karst gorges and peak cluster depressions ( $\bar{q} = 13.1 %$ ); but its explanatory force for soil erosion in other karst landform areas was low ( $\bar{q} = 5.09 %$ ), and had no significant impact in karst basins (p > 0.05). The average explanatory power of aspect on soil erosion was 8.32%, while that in karst trough valleys was 17.47%, which was significantly higher than other environmental factors. Slope position had the highest explanatory power for soil erosion, and presented the following rules: the explanatory power for soil erosion was the highest in peak cluster depressions and karst basins with small undulations ( $\bar{q} = 11.86 %$ ), and gradual decreases were noted in karst trough valleys, gorges, and plateaus with large undulations ( $\bar{q} = 6.01 %$ ).

Interaction of driving factors

Overall, the interaction of environmental factors will enhance the explanatory power for soil erosion and was higher than the sum of the explanatory powers of single factors (Figure 6). The dominant interaction factors of soil erosion showed great differences in different karst landforms. For example, the dominant interaction factors for karst trough valleys and gorges were development index and lithology ( $\bar{q} = 27.89 %$ ), while karst plateau were elevation and slope position (q=12.81%). In addition, although the dominant interaction factor had geomorphological differences, the repetition rate of single factor was high; the repetition rate for elevation was 60%, and the repetition rate for lithology and development index was 40%. Bedrock exposure rate had no significant effect on soil erosion in karst trough valleys (p > 0.05). However, its explanatory power for soil erosion was significantly enhanced when interacted with other factors ( $\bar{q} = 10.16 %$ , p < 0.05); the development index was similar.

Figure 6.

Interaction detection between drivers.

IV Discussion

1 Identification of priority areas for soil erosion control

The identification of priority areas for soil erosion control is crucial for the implementation of soil and water conservation projects (Gao and Wang, 2019b), particularly in the southwest karst areas of China, where soil erosion is widespread and extremely severe (Jiang et al., 2021). In this study, the priority areas for soil erosion control identified were clustered in the southwest, southeast, and northeast marginal areas of the study area (Figure 2(b)); this also verifies the previous analysis conclusions of soil erosion in this area (Fang et al., 2022). Priority areas overlapped with the strong erosion zone (Figure 2), and the average soil erosion rate was 45.79 t·ha⁻¹·a⁻¹, which belonged to very high grade of erosion. Ideally, 24.28% of the soil erosion in the study area may be reduced by implementing water and soil conservation projects in the priority areas (Table 3). Obviously, the implementation of ecological conservation projects in priority areas will have more economic and soil and water conservation effects.

Table 3.

Priority areas of soil erosion control in different karst landforms.

	Karst trough valley	Peak cluster depression	Karst plateau	Karst basin	Karst gorge	Karst area
Percentage of area (%)	2.36	1.76	6.34	0.73	1.57	14.16
Percentage of erosion (%)	4.84	4.35	12.71	2.23	3.57	24.28
Average erosion rate (t·ha⁻¹·a⁻¹)	42.29	46.59	45.86	41.56	40.45	45.79
Grade	Very high	Very high	Very high	Very high	Very high	Very high

(Note: Percentage of area refers to the area ratio between the priority area and this region; Percentage of erosion refers to the ratio of total soil erosion between the priority area and this region).

In addition, we identified multifactorial gradient risk zones for soil erosion in priority areas using socioeconomic and natural factors (e.g., the development index, bedrock exposure rate, and slope position). It revealed high-risk areas for soil erosion in different karst landforms through multiple key factors. Therefore, it can help managers develop more targeted soil and water conservation measures. Meanwhile, the multifactorial gradient risk zone showed more severe erosion compared to the average soil erosion rate in the study area. This is indicative for the implementation of large-scale soil erosion control projects.

Soil erosion is influenced by a combination of environmental factors (Jomaa et al., 2012; Chu et al., 2020). The heterogeneity of environmental factors will lead to strong spatial variability in soil erosion, especially in karst areas with complex and fragmented landforms (Jiang et al., 2014). This is a serious constraint for managers to plan and manage soil erosion in the region. In this study, we identified priority areas for soil erosion control through spatial correlations and administrative boundaries. Compared with traditional methods, the priority areas identified in this study are closer to actual management needs. This may increase the ease of implementation of soil erosion control planning and soil and water conservation projects; this result is also in line with the current development trend of soil erosion control in China from slope scale to comprehensive control of small watersheds (Zhao et al., 2013; Jiang et al., 2018).

2 Response of soil erosion to karst landforms

Spatial heterogeneity of karst landforms and environmental factors may lead to local enhancement or weakening of soil erosion (Cao et al., 2020b), which has a major impact on the determination of priority areas for soil erosion control. The priority areas identified in this study showed significant variation across the different karst landforms. Among them, the erosion rate, area and amount of the priority area in the karst plateau were higher than those of other karst landforms. This was because karst plateaus have a well-developed traditional agricultural economy, and frequent irrational farming in these areas has led to severe anthropogenic soil loss (Ci et al., 2006). Moreover, frequent erosive rainfall due to quasistatic fronts affects soil erosion (Jiang et al., 2021). Therefore, soil erosion control in karst plateaus warrants further attention.

In addition to influencing the identification of priority areas, karst landforms also exert a macro-control on the drivers of soil erosion in priority areas by influencing surface processes (Zl et al., 2019). In this study, the dominant control factors for soil erosion in the five karst landforms differed considerably. This may be due to the special geochemical background and strong dissolution in the karst areas, which makes the surface of these areas crack and break and the geographical characteristics of different karst landforms significantly different (Lan et al., 2022a). Bedrock exposure rate was the dominant control in karst basins, whereas there was no significant effect on soil erosion in karst trough valleys. This indicates that although the bedrock exposure rate is a crucial factor affecting soil erosion in karst areas, it does not control soil erosion significantly in specific karst landform.

The explanatory power of driving factors for soil erosion also demonstrated various changes in the different karst landforms. Slope position showed a higher explanatory power for soil erosion in karst landforms with relatively flat terrains (i.e., peak cluster depressions and karst basins) than it did in karst landforms with relatively large reliefs (i.e., karst trough valleys, plateaus, and basins). This may be due to the complex topography, ecological fragility, and high climate variability in steep areas, which lead to more complex drivers of soil erosion (Huan et al., 2019); therefore, the selected factors have low explanatory power for the spatial distribution of soil erosion. Elevation demonstrated the highest explanatory power for soil erosion in karst gorges. This was possibly because these areas have deep and steep valleys; a large relative height difference; considerable vertical differentiation of the geographical environment. And the interlayer differentiation of altitude can reflect the gradient differences of climate, vegetation, soil, and other driving factors. The soil erosion studies carried out by some scholars from the perspective of large geomorphology also show similar patterns to ours (Gao and Wang, 2019a).

In conclusion, we find that karst landforms had a strong control on soil erosion, especially in terms of priority areas and driving factors (Li et al., 2019; Gao and Wang, 2019b). Therefore, it is necessary to consider the geographical environmental factors of soil erosion when implementing soil conservation and ecological restoration projects in karst areas. It also gives personalized management measures for different types of karst landforms in order to improve the efficiency of soil erosion control.

3 Response of soil erosion to socioeconomic factors

Soil erosion is a complex surface process affected by both natural conditions and anthropogenic activities (Manojlović et al., 2022; Udayakumara et al., 2010). And in some specific contexts, the influence of socio-economic factors on soil erosion may be stronger (Borrelli et al., 2017; Poesen, 2018). Therefore, we constructed a development index based on population density and GDP to explore the influence of socioeconomic factors on soil erosion.

We observed that the development index significantly affected soil erosion in all karst landform areas, with the exception of karst basins. Karst basins, located in the southwest of the Guizhou karst area, are rich in natural and human resources (coal, forests, and tourism), have developed services and industries, and have low-intensity agricultural activities that cause relatively little damage to the surface soil (Sun et al., 2020). This may be the cause of the nonsignificant impact of the regional development index on soil erosion in the region. In addition, the explanatory power of the development index for soil erosion in karst trough valleys and gorges was 8.14 times that of the other three karst landform areas. This may be due to the special geographical environment of the study area. Karst trough valleys and gorges have closed terrains, lack resources, and have weak external transportation capacities (Waltham, 2008); its development has lagged behind other regions for a long time. An excessive population has led to considerable discord between humans and the land in the study area. To obtain sufficient living materials, forest land and grassland have been largely reclaimed as farmland (Wang et al., 2020). However, inefficient traditional agricultural cultivation has aggravated the damage to the ecological environment, resulting in serious anthropogenic soil loss (Qiao et al., 2020). Thus, soil erosion prevention and control in karst trough valleys and gorges should focus on economic development along with strategic rural revitalization and gradual adjustment of the industrial structure. To reduce the pressure on the land, the agricultural population in the area could gradually shift to engage in rural tourism and other service industries. Moreover, depending on the characteristics of the natural environment, a mixed agriculture and forestry industry may be developed in karst trough valleys and gorges, and low-quality farmlands may be transformed into economic and ecological fruit forests (Zou et al., 2019).

4 Effects of driving factor interaction on soil erosion

Soil erosion is simultaneously affected by various natural and human driving factors (Boardman et al., 2003; Eniyew et al., 2021). Therefore, we introduced the geographical detector to analyze the effects of the interactions of these driving factors on soil erosion. The results showed that the interaction of the drivers will enhance the explanatory power for soil erosion in different karst landforms (Figure 6). Thus, considering the interaction between the driving factors is essential for formulating and implementing water and soil conservation measures.

Notably, the dominant interaction factors for soil erosion in different karst landscapes showed large differences (Figure 6). For karst trough valleys and gorges, the interaction of lithology and development index had greater explanatory power for soil erosion than other factors. This may be because the economic structure of the study area is mainly agricultural, and the irrational cultivation methods there may have led to extreme damage to the soil (Wang et al., 2020). Lithology is the material basis of soil formation, and different lithological backgrounds play a decisive role in determining the physical and chemical properties of surface soils (Duan et al., 2021). The interactions of lithology and development index can effectively explain the spatial distribution of soil erosion in this area. Thus, scientific agricultural production systems should be established in karst trough valleys and gorges to reduce soil erosion from irrational agricultural activities. And implement different protection measures for different lithological areas. Such measures include reducing the area of land for cultivation in the limestone sandwich clastic rock distribution zone, increasing the proportion of forests and grasslands in the region, and building terraces or implementing intercropping.

In addition, although the dominant interaction factors were different for karst plateaus, karst basins, and peak cluster depressions, they all contain elevation. Peak cluster depressions, which belong to the middle stage of karst landform evolution, have peaks that are undulating and connected, which lead to the formation of some small-scale depressions between the peaks (Vardanjani, 2015). Karst basins also belong to the middle stage; however, those in Guizhou are generally small, with the peaks around the basin being tall and occupying most of the geomorphological units (Sweeting, 2012). Karst Plateau is located mainly in the north of Guizhou Province, where geotectonic movements have created many undulating mountain ranges. In all three karst landform areas, peak clusters and mountains form the majority of the structures, with relatively few flat regions present within the structures. As such, many farmers perform farming activities on the slopes in these areas to obtain enough living resources (Peng and Wang, 2012). Under karstification, the slope increases sharply with an increase in elevation (Sweeting, 2012). Meanwhile, as elevation increases, vegetation gradually becomes sparse and the ecological environment tends to become fragile (Piao et al., 2011). This leads to the interaction between elevation and other environmental factors having a major impact on spatial distribution of soil erosion in the region. Therefore, managers and stakeholders should timely adjust their agricultural production mode and draw an elevation red line for land cultivation. Ecological restoration measures, such as restoring farmlands to forests and grasslands should be implemented above the red line. Gradually reduce the intensity of cultivation and develop ecological agriculture at low altitudes below the red line to reduce soil damage.

V Conclusion

In this study, we systematically considered natural and socioeconomic factors not covered in the RUSLE model, and then identified priority areas for soil erosion control based on minimum administrative units and karst landforms. Our analysis showed that the priority areas within the smallest administrative units were clustered, and overlapped with erosion zones of strong and above grades (Erosion rate=45.79 t·ha⁻¹·a⁻¹); it had strong erosivity and spatial correlation. Gradient risk zones had obvious differences in different karst landforms, but multiple factors had an indicative effect on the establishment of management measures. This indicated that the implementation of soil and water conservation and ecological restoration projects in priority areas may have higher economic and ecological benefits. We also highlighted the role of karst landforms in controlling soil erosion. The dominant driving factors for karst trough valleys, plateaus, basins, gorges and peak cluster depressions were lithology, slope position, bedrock exposure rate, development index and elevation, respectively (q = 0.82% – 17.73%, P < 0.05). Interaction of driving factors could enhance the explanation power for soil erosion, with increases ranging from 7.99% to 1653.3%; but the dominant interaction factors and its explanatory power were geomorphological differences. Our research can provide valuable decision support for soil and water resources management by the water and forestry sectors.

Supplemental Material

Supplemental Material - Identification of priority areas for soil erosion control based on minimum administrative units and karst landforms in karst areas of Guizhou

Identification of priority areas for soil erosion control based on minimum administrative units and karst landforms in karst areas of Guizhou by Jun Jing, Rui Li, Yushan Zhang, Qinglin Wu in Progress in Physical Geography: Earth and Environment

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Natural Science Foundation of China (NO. 32060372), Guizhou Province Department of Science and Technology (QianKeHe JiChu-ZK [2023] Key029), Guizhou Province Department of Science and Technology (QianKeHe ZhiCheng [2021] Yiban462), Natural Science Foundation of Guizhou Province (Grant No. Qiankehe Jichu -ZK [2022] 317)

ORCID iD

Rui Li

Supplemental Material

Supplemental material for this article is available online.

References

Asmamaw

Mohammed

(2019) Identification of soil erosion hotspot areas for sustainable land management in the Gerado catchment, North-eastern Ethiopia. Remote Sensing Applications: Society and Environment 13: 306-317.

Betelhem Fetene

Temesgen

Anwar

, et al. (2022) Identification of soil erosion hot-spot areas for prioritization of conservation measures using the SWAT model in Ribb watershed, Ethiopia. Resources, Environment and Sustainability 8: 100059.

Boardman

Poesen

Evans

(2003) Socio-economic factors in soil erosion and conservation. Environmental Science & Policy 6(1): 1-6.

Borrelli

Robinson

Fleischer

, et al. (2017) An assessment of the global impact of 21st century land use change on soil erosion. Nature Communications 8(1): 2013.

Bridge

Demicco

(2008) Earth Surface Processes, Landforms and Sediment Deposits. Cambridge, England: Cambridge University Press

Cao

Wang

Peng

, et al. (2020a) Monitoring of suspended sediment load and transport in an agroforestry watershed on a karst plateau, Southwest China. Agriculture, Ecosystems & Environment 299: 106976.

Cao

Zhang

, et al. (2020b) Identifying and estimating soil erosion and sedimentation in small karst watersheds using a composite fingerprint technique. Agriculture, Ecosystems & Environment 294: 106881.

Cao

Keli

Jianghu

, et al. (2021) Linking rocky desertification to soil erosion by investigating changes in soil magnetic susceptibility profiles on karst slopes. Geoderma 389: 114949.

Cheng

Ren

, et al. (2017) Integrated emergy and economic evaluation of three typical rocky desertification control modes in karst areas of Guizhou Province, China. Journal of Cleaner Production 161: 1104-1128.

10.

Cho

Park

(2013) Prioritizing subwatersheds for stormwater pollution to wachusett reservoir. Water Environment Research A Research Publication of the Water Environment Federation 85(2): 158-166.

11.

Chu

Sun

Wang

, et al. (2020) Temporal and spatial heterogeneity of soil erosion and a quantitative analysis of its determinants in the three gorges reservoir area, China. International Journal of Environmental Research and Public Health 17(22): 8486.

12.

Yang

Qing

, et al. (2006) Rocky land desertification and its controlling measurements in the karst mountainous region,southwest of China. Chinese Journal of Soil Science 5.

13.

Corenblit

Baas

ACW

Bornette

, et al. (2011) Feedbacks between geomorphology and biota controlling earth surface processes and landforms: a review of foundation concepts and current understandings. Earth-Science Reviews 106(3): 307-331.

14.

Dai

Liu

Shao

, et al. (2015) Karst bare slope soil erosion and soil quality: a simulation case study. Solid Earth 6(3): 985-995.

15.

Dai

Peng

Yang

, et al. (2017) Runoff and erosion processes on bare slopes in the karst rocky desertification area. Catena 152: 218-226.

16.

Duan

Sheng

Yuan

, et al. (2021) Land use conversion and lithology impacts soil aggregate stability in subtropical China. Geoderma 389: 114953.

17.

Eniyew

Teshome

Sisay

, et al. (2021) Integrating RUSLE model with remote sensing and GIS for evaluation soil erosion in Telkwonz Watershed, Northwestern Ethiopia. Remote Sensing Applications: Society and Environment 24: 100623.

18.

Fang

, et al. (2022) RUSLE model evaluation of the soil and water conservation ratio of the Guizhou province in China between 2000 and 2019. Sustainability 14(13): 8219.

19.

Feng

Chen

Polyakov

, et al. (2016) Soil erosion rates in two karst peak-cluster depression basins of northwest Guangxi, China: comparison of the RUSLE model with 137Cs measurements. Geomorphology 253: 217-224.

20.

Feng

Liang

, et al. (2020) Forestation does not necessarily reduce soil erosion in a karst watershed in southwestern China. Progress in Physical Geography: Earth and Environment 45(1): 82-97.

21.

Cai

, et al. (2011) Soil thickness effect on hydrological and erosion characteristics under sloping lands: a hydropedological perspective. Geoderma 167: 41-53.

22.

Gao

Wang

(2019a) Spatiotemporal tradeoff of Karst water yield and soil erosion based on GWR Model: a case study in Sancha River Basin of Guizhou Province, China. Mountain Research 37(4): 518-527.

23.

Gao

Wang

(2019b) Temporal analysis on quantitative attribution of karst soil erosion: a case study of a peak-cluster depression basin in Southwest China. Catena 172: 369-377.

24.

Gao

Wang

Zuo

(2018) Spatial gradient and quantitative attribution of karst soil erosion in Southwest China. Environmental Monitoring and Assessment 190(12): 730.

25.

Guo

Liu

Men

, et al. (2021) Multiscale spatiotemporal characteristics of landscape patterns, hotspots, and influencing factors for soil erosion. Science of the Total Environment 779: 146474.

26.

Gutiérrez

(2016) Karst landforms. In: Gutiérrez

Gutiérrez

(eds) Landforms of the Earth: An Illustrated Guide. Cham: Springer International Publishing, pp.59-101.

27.

Huan

Jiangbo

Wenjuan

(2019) Quantitative attribution analysis of soil erosion in different geomorphological types in karst areas: based on the geodetector method. Journal of Geographical Sciences 29(002): 271-286.

28.

Huang

Zhou

Wang

, et al. (2019) Occurrence mechanism and prediction of rocky land degradation in karst mountainous basins with the aid of GIS technology, a study case in Houzhai River Basin in southwestern China. Environmental Earth Sciences 78(6): 217.

29.

Huang

Zhang

Zhou

, et al. (2021) Spatial heterogeneity of soil thickness and factors controlling it in a karst basin. Eurasian Soil Science 54(4): 478-486.

30.

Jiang

Zhang

, et al. (2004) Progress in developing an ANN model for air pollution index forecast. Atmospheric Environment 38(40): 7055-7064.

31.

Jiang

Lian

Qin

(2014) Rocky desertification in Southwest China: impacts, causes, and restoration. Earth-Science Reviews 132: 1-12.

32.

Jiang

Zhang

, et al. (2018) Model-based assessment soil loss by wind and water erosion in China's Loess Plateau: dynamic change, conservation effectiveness, and strategies for sustainable restoration. Global and Planetary Change 172(JAN.): 396-413.

33.

Jiang

Gao

Yang

, et al. (2021) The interactive effects of elevation, precipitation and lithology on karst rainfall and runoff erosivity. Catena 207: 105588.

34.

Jin

Qian

Chin

, et al. (2020) A global assessment of sustainable development based on modification of the human development index via the entropy method. Sustainability 12.

35.

Jing

Yuan

, et al. (2023) Rainstorm sediment events in heterogeneous karst small watersheds: process characteristics, prediction modeling and management enlightenment. Science of the Total Environment 875: 162679.

36.

Jomaa

Barry

Brovelli

, et al. (2012) Rain splash soil erosion estimation in the presence of rock fragments. Catena 92: 38-48.

37.

Zhang

(2022) Interaction effects of rainfall and soil factors on runoff, erosion, and their predictions in different geographic regions. Journal of Hydrology 605: 127291.

38.

Lan

Ding

Dai

, et al. (2022a) Assessing the degree of soil erosion in karst mountainous areas by extenics. Catena: An Interdisciplinary Journal of Soil Science Hydrology-Geomorphology Focusing on Geoecology and Landscape Evolution 209(1).

39.

Lan

Ding

Dai

, et al. (2022b) Assessing the degree of soil erosion in karst mountainous areas by extenics. Catena 209: 105800.

40.

Lara

Nitze

Grosse

, et al. (2018) Reduced arctic tundra productivity linked with landform and climate change interactions. Scientific Reports 8(1): 2345.

41.

, et al. (2016) Quantifying the impacts of climate and human activities on water and sediment discharge in a karst region of southwest China. Journal of Hydrology 542: 836-849.

42.

Zhu

, et al. (2019) Effects of lithology and geomorphology on sediment yield in karst mountainous catchments. Geomorphology 343: 119-128.

43.

S-L

Liu

C-Q

Chen

J-A

, et al. (2021) Karst ecosystem and environment: characteristics, evolution processes, and sustainable development. Agriculture, Ecosystems & Environment 306: 107173.

44.

Guan

Sun

, et al. (2022) Spatiotemporal analysis of the quantitative attribution of soil water erosion in the upper reaches of the Yellow River Basin based on the RUSLE-TLSD model. Catena 212: 106081.

45.

Liang

Fang

(2021) Quantitative analysis of driving factors in soil erosion using geographic detectors in Qiantang River catchment, Southeast China. Journal of Soils and Sediments 21(1): 134-147.

46.

Manojlović

Sibinović

Srejić

, et al. (2022) Factors controlling the change of soil erosion intensity in mountain watersheds in Serbia. Frontiers in Environmental Science 10.

47.

Mccool

Brown

Foster

, et al. (1987) Revised slope steepness factor for the universal soil loss equation. Transactions of the ASAE - American Society of Agricultural Engineers (USA) 30(5): 1387-1396.

48.

Mccool

Foster

Mutchler

, et al. (1989) Revised slope length factor for the universal soil loss equation. Transactions of the ASAE 32: 1571-1576.

49.

Moore

Burch

(1986) Physical basis of the length-slope factor in the universal soil loss equation. Soil Science Society of America Journal 50(5): 1294-1298.

50.

Niu

Shao

(2020) Soil conservation service spatiotemporal variability and its driving mechanism on the Guizhou plateau, China. Remote Sensing 12(14).

51.

Olorunfemi

Komolafe

Fasinmirin

, et al. (2020) A GIS-based assessment of the potential soil erosion and flood hazard zones in Ekiti State, Southwestern Nigeria using integrated RUSLE and HAND models. Catena 194: 104725.

52.

Pan

Wen

(2014) Estimation of soil erosion using RUSLE in Caijiamiao watershed, China. Natural Hazards 71(3): 2187-2205.

53.

Pan

L-d

Shu

D-c

, et al. (2022) Effects of rainfall and rocky desertification on soil erosion in karst area of Southwest China. Journal of Mountain Science 19(11): 3118-3130.

54.

Panagos

Ballabio

Himics

, et al. (2021) Projections of soil loss by water erosion in Europe by 2050. Environmental Science & Policy 124: 380-392.

55.

Peng

Wang

S-j

(2012) Effects of land use, land cover and rainfall regimes on the surface runoff and soil loss on karst slopes in southwest China. Catena 90: 53-62.

56.

Peng

Dai

(2022) Drivers of soil erosion and subsurface loss by soil leakage during karst rocky desertification in SW China. International Soil and Water Conservation Research 10: 217-227.

57.

Piao

Cui

Chen

, et al. (2011) Altitude and temperature dependence of change in the spring vegetation green-up date from 1982 to 2006 in the Qinghai-Xizang Plateau. Agricultural and Forest Meteorology 151(12): 1599-1608.

58.

Poesen

(2018) Soil erosion in the anthropocene: research needs. Earth Surface Processes and Landforms 43(1): 64-84.

59.

Qiao

Chen

Jiang

(2020) Quantifying the impacts of lithology on vegetation restoration using a random forest model in a karst trough valley, China. Ecological Engineering 156: 105973.

60.

Rafiei

Ghahramani

An-Vo

D-A

, et al. (2020) Modelling hydrological processes and identifying soil erosion sources in a tropical catchment of the great barrier reef using SWAT. Water 12(8): 2179.

61.

Ren

, et al. (2022) Specific urban units identified in tuberculosis epidemic using a geographical detector in Guangzhou, China. Infectious Diseases of Poverty 11(1): 1-12.

62.

Richardson

Foster

Wright

(1983) Estimation of erosion index from daily rainfall amount. Transactions of the ASAE American Society of Agricultural Engineers 26(1): 0153-0156.

63.

Senanayake

Pradhan

Alamri

, et al. (2022) A new application of deep neural network (LSTM) and RUSLE models in soil erosion prediction. Science of the Total Environment 845: 157220.

64.

Shi

Guo

Zeng

(2006) Study on rainfall erosivity of the characteristics and daily rainfall erosivity model in Wuhan city. China Soil Water Conserv 1: 22–24.

65.

Silva

(2004) Rainfall erosivity map for Brazil. Catena 57(3): 251-259.

66.

Sun

Yang

, et al. (2020) Comprehensive partitions and different strategies based on ecological security and economic development in Guizhou Province, China. Journal of Cleaner Production 274: 122794.

67.

Sweeting

(2012) Karst in China: Its Geomorphology and Environment. Berlin, Germany: Springer Science & Business Media.

68.

Tamene

Adimassu

Ellison

, et al. (2017) Mapping soil erosion hotspots and assessing the potential impacts of land management practices in the highlands of Ethiopia. Geomorphology 292: 153-163.

69.

Tang

Qin

, et al. (2019) Karst rocky desertification progress: soil calcium as a possible driving force. Science of the Total Environment 649: 1250-1259.

70.

Udayakumara

EPN

Shrestha

Samarakoon

, et al. (2010) People's perception and socioeconomic determinants of soil erosion: a case study of Samanalawewa watershed, Sri Lanka. International Journal of Sediment Research 25(4): 323-339.

71.

Vardanjani

(2015) Karst Hydrogeology and Geomorphology. Hoboken, NJ, USA: John Wiley and Sons.

72.

Vijith

Seling

Dodge-Wan

(2018) Estimation of soil loss and identification of erosion risk zones in a forested region in Sarawak, Malaysia, Northern Borneo. Environment, Development and Sustainability 20(3): 1365-1384.

73.

Waltham

(2008) Fengcong, fenglin, cone karst and tower karst. Cave and Karst Science 35(3): 77-88.

74.

Wang

Zhang

(2016) A measure of spatial stratified heterogeneity. Ecological Indicators 67: 250-256.

75.

Wang

Zhang

Chen

, et al. (2019) Karst landscapes of China: patterns, ecosystem processes and services. Landscape Ecology 34(12): 2743-2763.

76.

Wang

Luo

(2020) Spatiotemporal change characteristics and driving mechanism of slope cultivated land transition in karst trough valley area of Guizhou Province, China. Environmental Earth Sciences 79(12): 284.

77.

Wen

(2020) Temporal and spatial relationships between soil erosion and ecological restoration in semi-arid regions: a case study in northern Shaanxi, China. GIScience and Remote Sensing 57(4): 572-590.

78.

Wen

Deng

(2020) Current soil erosion assessment in the Loess Plateau of China: a mini-review. Journal of Cleaner Production 276: 123091.

79.

Wischmeier

Smith

(1965) Predicting rainfall-erosion losses from cropland east of the Rocky Mountains. Agricultural Handbook 282.

80.

Xie

Yin

Liu

(2016) Models for estimating daily rainfall erosivity in China. Journal of Hydrology.

81.

Zhang

(2020) Change pathway and intersection of rainfall, soil, and land use influencing water-related soil erosion. Ecological Indicators 113: 106281.

82.

Yang

, et al. (2022) Runoff sediment characteristics affected by erosive rainfall patterns in a small watershed in karst areas of southwest China. Catena 219: 106591.

83.

Rosewell

(1996) An assessment of a daily rainfall erosivity model for New South Wales. Australian Journal of Soil Research 34(1): 139-152.

84.

Yuan

(2014) The research and countermeasures of major environmental geological problems in karst areas of southwest china. Science Press.

85.

Yue

Zhu

(2019) Exploring the relationship between urban vitality and street centrality based on social network review data in Wuhan, China. Sustainability 11(16): 4356.

86.

Zeng

Wang

Bai

, et al. (2017) Soil erosion evolution and spatial correlation analysis in a typical karst geomorphology using RUSLE with GIS. Solid Earth 8(4): 721-736.

87.

Zeng

Jiang

Shen

, et al. (2018) Assessment of multiple and interacting modes of soil loss in the karst critical zone, Southwest China (SWC). Geomorphology 322: 97-106.

88.

Zhang

W-B

Xie

Liu

B-Y

(2002) Rainfall erosivity estimation using daily rainfall amounts. Scientia Geographica Sinica/Dili Kexue 22(6): 711-716.

89.

Zhang

Ling

, et al. (2010) Identification of priority areas for controlling soil erosion. Catena 83(1): 76-86.

90.

Zhang

Yang

, et al. (2013) Extension of a GIS procedure for calculating the RUSLE equation LS factor. Computers & Geosciences 52: 177-188.

91.

Zhang

Jing

(2022a) Soil erosion gradient and quantitative attribution in southwest China based on karst development degree. Ecological Indicators 144: 109496.

92.

Zhang

, et al. (2022b) Modelling soil thickness using environmental attributes in karst watersheds. Catena 212: 106053.

93.

Zhao

(2017) Institutional Environment for Township Governments. The Politics Of Peasants. Berlin, Germany: Springer, pp.123-129.

94.

Zhao

Wen

, et al. (2013) Soil erosion, conservation, and eco-environment changes in the Loess Plateau of China. Land Degradation & Development 24(5): 499-510.

95.

Zhou

Wang

, et al. (2020) The challenge of soil loss control and vegetation restoration in the karst area of southwestern China. International Soil and Water Conservation Research 8(1): 26-34.

96.

Zhu

Xiong

Xiao

(2021) Multi-time scale variability of rainfall erosivity and erosivity density in the karst region of southern China, 1960–2017. Catena 197: 104977.

97.

, et al. (2019) Effects of lithology and geomorphology on sediment yield in karst mountainous catchments. Geomorphology 343: 119-128.

98.

Zou

Zeng

Wang

, et al. (2019) Emergy and economic evaluation of seven typical agroforestry planting patterns in the karst region of Southwest China. Forests 10(2): 138.

99.

Zuo

Kita

(2012) Stock price forecast using Bayesian network. Expert Systems with Applications 39(8): 6729-6737.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.17 MB