Multi-terrain soil erosion space-time driving force analysis system

Abstract

The invention provides a multi-terrain soil erosion space-time driving force analysis system. The analysis system comprises the following steps: collecting soil erosion related data; dividing the terrain of the research area; calculating a soil erosion modulus of the research area by using a RUSLE model, and grading the erosion modulus according to a soil erosion classification standard; comparing soil erosion modulus and erosion grade differences in different terrain areas and RUSLE model driving factor space distribution and time change differences; analyzing spatial differentiation characteristics of driving factors of the soil erosion RUSLE model in different terrains through a geographic detector in space; analyzing the time differentiation characteristics of the natural and human factors through random forest regression in time, and analyzing the driving intensity of the natural and human factors through a structural equation model; according to the system, spatial and time dimensions are integrated, attribution analysis is carried out on soil erosion differences under multiple topographic conditions, and enhancement of understanding of different topographic erosion differences and mechanisms thereof is facilitated.

Description

Technical Field

[0001] The present invention relates to the field of soil erosion analysis, and more specifically to an analysis system of spatiotemporal driving forces of soil erosion in multiple terrains. Background Art

[0002] Soil erosion is a global environmental problem. It not only leads to the degradation of land resources and the loss of soil fertility, but also may cause ecological problems such as river siltation, posing a serious threat to human survival and ecological sustainability. Its occurrence and development are not only directly affected by natural factors such as rainfall and vegetation cover, but also by obvious differences in soil erosion under different terrain conditions, especially in mountainous areas and plains. In the current field of soil erosion analysis, the analysis of the differences in soil erosion under different terrain conditions is still insufficient, especially in the identification and comparative analysis of the main driving factors of soil erosion. There is an obvious gap, and the existing analysis only selects data from some years in time, which is insufficient in time continuity and cannot fully reflect the long-term dynamic changes of soil erosion. Therefore, it is very important to conduct spatiotemporal attribution analysis of the differences in soil erosion under multiple terrain conditions. This will help to enhance the understanding of the differences and mechanisms of erosion in different terrains, thereby improving the accuracy of soil erosion research and the effectiveness of governance measures. Summary of the invention

[0003] The purpose of the present invention is to provide a spatiotemporal driving force analysis system for soil erosion under multiple terrain conditions. The system first divides the study area into terrain units such as mountainous areas, hills, and plains; then, based on the modified universal soil loss equation (RUSLE) model, the system calculates the soil erosion modulus of each terrain unit, and divides the erosion modulus into 6 levels according to the soil erosion classification standard; then, the system compares the differences in soil erosion modulus and erosion levels under different terrains, as well as the differences in spatial distribution and temporal changes of driving factors of the RUSLE model; then, in the spatial dimension, the system analyzes the spatial differentiation characteristics of driving factors of the RUSLE model of soil erosion in different terrains through geographic detectors; in the temporal dimension, the temporal differentiation characteristics of natural and human factors are analyzed through random forest regression, and the driving intensity of natural and human factors is analyzed through the structural equation model.

[0004] The specific scheme of the present invention is to provide an analysis system of the spatiotemporal driving force of soil erosion in multiple terrains, comprising the following steps:

[0005] Step 1, collect relevant data on soil erosion in the study area;

[0006] Step 2: Classify the terrain in the study area based on the digital elevation model (DEM) and slope of the study area;

[0007] Step 3, after preprocessing the collected data on soil erosion in the study area, the soil erosion modulus of different terrain areas in the study area is calculated using the modified universal soil loss equation model, namely the RUSLE model, and the calculation results are graded for erosion intensity;

[0008] Step 4: Based on the results of step 3, compare and analyze the differences in soil erosion modulus and soil erosion intensity levels in different terrain areas of the study area;

[0009] Step 5, extract the driving factors in the RUSLE model, and compare and analyze the differences in spatial distribution and temporal changes of the driving factors of the RUSLE model of soil erosion in different terrain areas of the study area;

[0010] Step 6, the spatial differentiation characteristics of the driving factors of the RUSLE model of soil erosion in different terrain areas of the study area were analyzed through geographic detectors;

[0011] Step 7, obtain human factors and natural factors, and analyze the temporal differentiation characteristics of natural and human factors of soil erosion in different terrain areas of the study area through random forest regression model;

[0012] Step 8: The driving intensity of natural and anthropogenic factors of soil erosion in different terrain areas of the study area was analyzed through the partial least squares structural equation model (PLS-SEM).

[0013] As a further improvement of the present invention, the step 2 comprises the following steps:

[0014] Step 2-1, load and clip DEM data: first load the DEM data and study area, then use spatial analysis tools to clip the DEM to the study area.

[0015] Step 2-2, calculate slope: Use the slope tool to calculate the slope of the study area.

[0016] Step 2-3, slope grading and reclassification: Divide the grade and reclassify the terrain according to the slope calculation results.

[0017] As a further improvement of the present invention, the step 3 comprises the following steps:

[0018] Step 3-1, calculate the soil erosion modulus by using the modified universal soil loss equation RUSLE model, specifically:

[0019]

[0020] Where: A is the soil erosion modulus; R is the rainfall erosivity factor; K is the soil erodibility factor; LS is the slope length and slope factor; C is the surface vegetation coverage and management factor; P is the soil and water conservation measures factor;

[0021] Among them, the rainfall erosivity factor R is the rainfall erosivity of each year, and the calculation method is:

[0022]

[0023] Where P i is the monthly rainfall; P is the annual rainfall;

[0024] The soil erodibility factor K is an indicator of the sensitivity of soil properties to erosion. The calculation method is:

[0025]

[0026] In the formula, It represents the soil erodibility factor value calculated by the EPIC model, i.e. the cognitive model. K represents the soil erodibility factor value after correction. SAN represents sand particles. SIL represents silt particles. CLA represents clay particles. OM represents organic matter. The K value reflects the ease with which the soil is separated, eroded and transported by rainfall erosion. The larger the K value of the soil, the worse the soil's ability to resist water erosion and the more susceptible it is to erosion.

[0027] The slope length factor LS is divided into the slope length factor L and the slope factor S. The calculation method is:

[0028]

[0029] Among them, λ is the projection length of the grid unit; m is the slope length effect index; θ is the slope extracted from DEM;

[0030] Vegetation coverage and management factor C is used to reflect the impact of vegetation coverage and management measures on soil loss. The value is between 0 and 1. The calculation method is:

[0031]

[0032]

[0033] Where f is the vegetation coverage; C is the vegetation coverage and management factor value; NDVI is the normalized vegetation index using the multi-year monthly average; MAX and NDVI MIN is the maximum and minimum value of the normalized difference vegetation index;

[0034] The soil and water conservation measures factor P is the ratio of soil loss after taking soil conservation measures to soil loss when planting along the slope. Its value is between 0 and 1, with 0 representing an area where no soil and water erosion occurs and 1 representing an area where no soil and water conservation measures are taken.

[0035] Step 3-2, grading the erosion intensity of the soil erosion modulus calculation result. As a further improvement of the present invention, in step 4, based on the soil erosion modulus and intensity grading of the continuous time series of different terrain areas in the study area obtained in step 3, the soil erosion modulus of each area for many years is counted, and a curve graph is drawn to calculate the average soil erosion modulus for many years, so as to compare the difference of soil erosion modulus in different terrain areas in the study area, and at the same time, a soil erosion intensity grading diagram of the study area is drawn to further compare the difference of soil erosion intensity grades in different terrain areas in the study area;

[0036] As a further improvement of the present invention, in step 5, the driving factors in the RUSLE model are extracted, and a grading map of the driving factors of the RUSLE model in the study area is prepared to compare the differences in the spatial distribution of the driving factors of soil erosion in different terrains, and the data of the driving factors of the RUSLE model in the study area for many years are statistically analyzed and a change map is drawn to compare the differences in the temporal changes of the driving factors of soil erosion in different terrains;

[0037] Among them, the driving factors of the RUSLE model include rainfall erosivity factor, soil erodibility factor, slope gradient and length factor, vegetation cover factor and soil and water conservation factor.

[0038] As a further improvement of the present invention, step 6 is specifically as follows: quantifying the influence of a single factor on the spatial heterogeneity of soil erosion under different terrains through factor detection, and exploring the interaction between the driving factors of the RUSLE model under different terrains through interactive detection, so as to obtain the mechanism by which these factors jointly affect soil erosion under different terrains.

[0039] As a further improvement of the present invention, in step 7, human factors and natural factors are obtained, and based on a continuous time series, the statistical data of each factor under different terrain conditions are standardized, and then substituted into the random forest model to obtain the importance ratio of each natural and human factors based on the time series, and analyze the temporal differentiation characteristics of natural factors and human factors of soil erosion under different terrains;

[0040] Among them, natural factors include rainfall, temperature, and vegetation coverage; human factors include population density, GDP, and urbanization area.

[0041] As a further improvement of the present invention, in step 8, the standardized continuous time series data of different terrain areas are imported into the constructed partial least squares structural equation model PLS-SEM to obtain the path coefficients of the influence of natural factors and human factors on the soil erosion modulus of different terrains, so as to analyze the driving intensity of natural factors and human factors of soil erosion in different terrains.

[0042] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention provides a system for in-depth analysis of spatiotemporal driving forces for soil erosion in multiple terrains, covering the entire process from terrain division, difference comparison to attribution analysis. By combining the RUSLE model and terrain division, the system realizes the accurate calculation and evaluation of soil erosion modulus and intensity grade under multiple terrain conditions, and then compares the differences in soil erosion and driving factors under different terrain conditions. At the same time, the system integrates time and space dimensions, and uses methods such as geographic detectors, random forest regression and structural equation models to reveal the spatiotemporal differentiation characteristics and driving intensity of soil erosion driving factors under different terrains. The system conducts spatiotemporal attribution analysis on the differences in soil erosion under multiple terrain conditions, which helps to enhance the understanding of the differences in erosion and its mechanisms in different terrains, thereby improving the accuracy of soil erosion research and the effectiveness of control measures. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 This is a workflow diagram of a system for analyzing spatiotemporal driving forces of soil erosion in multiple terrains according to the present invention;

[0044] Figure 2 A topographic division map of the research area of ​​the present invention;

[0045] Figure 3 It is the soil erosion modulus curve graph of the study area of ​​the present invention from 2000 to 2022;

[0046] Figure 4 It is the soil erosion classification map of the study area of ​​the present invention in 2000, 2011 and 2022;

[0047] Figure 5 The distribution map of rainfall erosivity factor R, soil and water conservation measure factor P, and vegetation and management factor C in the study area of ​​the present invention in 2000, 2011 and 2022;

[0048] Figure 6 This is a change diagram of the rainfall erosivity factor R, the soil and water conservation measures factor P, and the vegetation and management factor C in the study area of ​​the present invention from 2000 to 2022;

[0049] Figure 7 It is a statistical histogram of q values ​​of soil erosion driving factors in mountainous areas and plain areas in the study area of ​​the present invention;

[0050] Figure 8 It is a statistical diagram of the interactive detection results of soil erosion driving factors in mountainous areas and plain areas in the study area of ​​the present invention;

[0051] Fig. 9 The present invention obtains a bar chart of the importance ratios of natural and human factors through random forest calculation;

[0052] Fig.10This is a schematic diagram of the driving intensity of natural and human factors on soil erosion in mountainous and plain areas in the study area obtained through the partial least squares structural equation model PLS-SEM. DETAILED DESCRIPTION

[0053] The present invention is described in detail below in conjunction with the various embodiments shown in the accompanying drawings, but it should be noted that these embodiments are not limitations of the present invention, and any equivalent transformations or substitutions in functions, methods, or structures made by ordinary technicians in the field based on these embodiments are all within the scope of protection of the present invention.

[0054] The present application discloses a system for analyzing the spatiotemporal driving force of soil erosion in multiple terrains, which specifically includes the following steps:

[0055] Step 1: Collect relevant data on soil erosion in the study area;

[0056] First, relevant data on soil erosion in the study area were collected, covering all meteorological and geographical data related to soil erosion, such as topography, rainfall, temperature, vegetation, land use, soil, population density, GDP and other data in the study area.

[0057] Step 2: Classify the regional terrain based on the digital elevation model (DEM) and slope of the study area;

[0058] Step 2-1, load and clip DEM data: first load the DEM data and the study area, such as the study area range, and then use the spatial analysis tools to clip the DEM to the study area.

[0059] Step 2-2, calculate slope: Use the slope tool to calculate the slope of the study area.

[0060] Step 2-3, slope classification and reclassification: Based on the calculation results, the landform types and slope values ​​are divided into the following levels as shown in Table 1, and the terrain is reclassified.

[0061] Table 1

[0062]

[0063] Step 3: Calculate the soil erosion modulus of different terrains in the study area using the RUSLE model, and classify the erosion intensity of the calculation results according to the SL190-2007 standard;

[0064] Step 3-1, calculate the soil erosion modulus through the modified universal soil loss equation RUSLE model. This model fully considers factors such as rainfall erosivity R, soil erodibility K, slope length LS, vegetation coverage C and soil and water conservation P. It is an effective method for studying soil erosion and is also the most widely used model for quantifying soil erosion. The calculation formula is as follows:

[0065]

[0066] Where: A is the soil erosion modulus (t / (km 2 ·a)); R is the rainfall erosivity factor ((MJ·mm) / (km 2 ·h·a)); K is soil erodibility factor ((t·km²·h) / (km²·MJ·mm)); LS is slope length and slope factor; C is surface vegetation coverage and management factor; P is soil and water conservation measures factor.

[0067] The rainfall erosivity R is a dynamic index to evaluate soil separation and transport caused by rainfall, reflecting the potential of rainfall to cause soil erosion. The calculation method for a continuous time series such as the rainfall erosivity for each year from 2000 to 2022 is:

[0068]

[0069] Where P i is the monthly rainfall; P is the annual rainfall.

[0070] The soil erodibility factor K is an indicator of the sensitivity of soil properties to erosion. The K value reflects the ease with which the soil is separated, eroded and transported by rainfall erosion. Generally speaking, the larger the soil K value, the worse the soil's ability to resist water erosion and the more susceptible it is to erosion. The present invention calculates the K value based on the EPIC model and modifies the results based on existing research, wherein the calculation of the soil erodibility factor is only related to the content of silt, sand, clay and organic matter in the soil.

[0071]

[0072] In the formula, It represents the soil erodibility factor value calculated by the EPIC model, i.e. the cognitive model; K represents the corrected soil erodibility factor value; SAN represents sand particles; SIL represents silt particles; CLA represents clay particles; and OM represents organic matter. The units of the above four variables are all %.

[0073] Slope length factor LS, topography is a geographical feature that affects soil erosion, and its influence is mainly reflected in the slope and slope length. Slope length refers to the distance from the starting point of surface runoff to the runoff converging ditch valley on the slope. The greater the slope length, the greater the runoff speed, the greater the converged flow, and the stronger the erosion. The slope represents the degree of inclination of a certain point on the surface, which affects the permeability, runoff and other factors by affecting the speed of water flow. The present invention extracts the slope length and slope through the digital elevation DEM of the study area, and calculates the slope length factor L. The slope factor S is specifically:

[0074]

[0075] Where L is the slope length factor; λ is the grid unit projection length; m is the slope length effect index; θ is the slope extracted from DEM; and S is the slope factor.

[0076] Vegetation coverage and management factor C, in the RUSLE model, C represents the ratio of soil loss of land covered with vegetation or implemented field management to soil loss of cleared land under the same conditions under certain conditions, which is used to reflect the impact of vegetation coverage and management measures on soil loss, and the value is between 0 and 1. Normalized difference vegetation index NDVI is the most common factor for calculating C factor, and the specific calculation of C factor is:

[0077]

[0078] Where f is the vegetation coverage; C is the vegetation coverage and management factor value; NDVI is the normalized vegetation index using the multi-year monthly average; MAX and NDVI MIN is the maximum and minimum value of the normalized difference vegetation index.

[0079] The soil and water conservation measures factor P is the ratio of soil loss after taking soil conservation measures to soil loss when planting along the slope. Its value is between 0 and 1, with 0 representing an area where no soil and water erosion occurs and 1 representing an area where no soil and water conservation measures are taken. Combined with the land use and agricultural activities in the study area, the P values ​​of different land use types were determined, as shown in Table 2 for the P values ​​of different land cover types.

[0080] Table 2

[0081]

[0082] Step 3-2, according to the soil erosion classification standard SL190-2007, soil erosion is divided into 6 erosion levels, as shown in Table 3, which is the soil erosion intensity classification standard.

[0083] Table 3

[0084]

[0085] Step 4: Compare and analyze the differences in soil erosion modulus and soil erosion intensity grades in different terrains in the study area;

[0086] In step 3, the soil erosion modulus and intensity classification of the continuous time series in different terrain areas of the study area were obtained, the soil erosion modulus of each area over the years was statistically plotted, and the average soil erosion modulus over the years was calculated, so as to compare the differences in soil erosion modulus in different terrain areas of the study area; a soil erosion intensity classification map of the study area was drawn to further compare the differences in soil erosion intensity levels in different terrain areas of the study area.

[0087] Step 5: Extract the driving factors in the RUSLE model and compare and analyze the differences in spatial distribution and temporal changes of the driving factors of the RUSLE model for soil erosion in different terrains in the study area;

[0088] Among them, the driving factors of the RUSLE model include rainfall erosivity factor, soil erodibility factor, slope gradient and length factor, vegetation cover factor and soil and water conservation factor.

[0089] A grading map of driving factors of the RUSLE model in the study area was produced to compare the differences in the spatial distribution of driving factors of soil erosion in different terrains; the data of driving factors of the RUSLE model in the study area for many years were collected and a change map was drawn to compare the differences in the temporal changes of driving factors of soil erosion in different terrains.

[0090] Step 6: Use geographic detector technology to analyze the spatial differentiation characteristics of driving factors of soil erosion RUSLE model in different terrains of the study area;

[0091] Geographic detector is an effective spatial statistical method based on the analysis of spatial variation at the geographic level of variables. Factor detection can quantify the impact of a single factor on the spatial heterogeneity of soil erosion. The higher the q value, the stronger the explanatory power. Interaction detection can evaluate the interaction between different factors and reveal how they jointly affect soil erosion. This application uses factor detection to statistically analyze the q value in the detection results of the differentiation of RUSLE model driving factors under different terrains to quantify the degree of explanation of the spatial differentiation of soil erosion by the driving factors of the RUSLE model under different terrains. And through interactive detection, the interaction between the driving factors of the RUSLE model under different terrains is explored to reveal the mechanism by which these factors jointly affect soil erosion under different terrains.

[0092] Step 7: Obtain human factors and natural factors, and analyze the temporal differentiation characteristics of natural and human factors of soil erosion in different terrains in the study area through the random forest regression model;

[0093] Among them, natural factors include rainfall, temperature, and vegetation coverage; human factors include population density, GDP, and urbanization area.

[0094] Statistics and collection of soil erosion-related data under different terrain conditions in the study area were conducted, mainly including natural factors (vegetation cover, temperature, rainfall) and human factors (GDP, urbanization area, population density).

[0095] Next, based on the continuous time series, the statistical data of each factor under different terrain conditions were standardized to eliminate the influence of different dimensions and ranges so that the data could be compared under the same standard.

[0096] Each factor in each region was standardized, and the random forest regression model was applied to different terrain regions. The soil erosion modulus was used as the dependent variable, and each influencing factor was used as the independent variable. Through regression analysis, the influence of each influencing factor on the soil erosion modulus in different terrain regions was calculated, that is, the importance ratio. Finally, according to the results of the model output, a chart was drawn to intuitively display and compare the contribution of each factor to soil erosion over the consecutive years, thereby revealing the temporal differentiation characteristics of soil erosion influencing factors in different terrain regions.

[0097] Step 8: Analyze the driving intensity of natural and human factors of soil erosion in different terrains through partial least squares structural equation model (PLS-SEM).

[0098] The partial least squares structural equation model (PLS-SEM) is used to reveal the driving strength of natural and human factors under different terrains. PLS-SEM is a statistical analysis method that combines partial least squares regression and structural equation modeling, which is suitable for dealing with small samples and non-normal distribution data. The structural equation consists of measurement equations and structural equations. The measurement equation refers to the connection between observed variables and latent variables, while the structural equation refers to the relationship between latent variables. The measurement equation is expressed as follows:

[0099]

[0100] Where: and is the vector of observed variables; ξ and η are the vectors of latent variables, Λ X and Λ Y is the factor loading matrix, δ and ε are interference terms. The structural equation is expressed as:

[0101]

[0102] Where: B is the relationship between endogenous latent variables; Γ is the impact of exogenous latent variables on endogenous latent variables; ζ is the error term.

[0103] First, the standardized statistical data of various influencing factors in different terrain areas under continuous time series were imported. Then, a structural equation model was constructed to define the relationship between variables. Taking mountainous areas and plain areas as examples, the direct path relationship between natural factors in mountainous areas, including vegetation cover, temperature, and rainfall, and the erosion modulus in mountainous areas, as well as the path relationship between anthropogenic factors in mountainous areas, including GDP, urbanization area, and population density, and the erosion modulus in mountainous areas were constructed. At the same time, path relationships were established between natural factors in plains, including vegetation cover, temperature, and rainfall, and the erosion modulus in plains, as well as between anthropogenic factors in plains, including GDP, urbanization area, and population density, and the erosion modulus in plains. Model analysis revealed the path coefficients of the influence of various natural and anthropogenic factors on the soil erosion modulus in different terrain areas, and quantified the intensity of each factor driving soil erosion in different terrain areas.

[0104] Experimental Example 1

[0105] This experimental example takes the spatiotemporal driving force of soil erosion in different terrain areas of the research area as the analysis target, and further demonstrates the analysis system and implementation method of the present invention:

[0106] Step 1: Collect relevant data on soil erosion in the study area and preprocess the data;

[0107] The data types and sources are shown in Table 4. All data were unified to a spatial resolution of 250 m using the ArcGIS raster resampling tool.

[0108] Table 4

[0109]

[0110] Step 2: Classify the terrain according to the digital elevation model (DEM) and slope of the study area. Considering that the hilly landform in the basin is not obvious, the study area is divided into two parts: mountainous area and plain area. Figure 2 As shown;

[0111] Step 3: Calculate the soil erosion modulus of the entire watershed, mountainous area, and plain area of ​​the study area using the RUSLE model, and classify the erosion intensity of the calculation results;

[0112] The processed rainfall erosion factor R, soil erosion factor K, slope and slope factor LS, vegetation cover and management factor C, and soil and water conservation measures factor P were introduced into the RUSLE model, and the annual soil erosion modulus of the entire basin, mountainous area and plain area of ​​the study area from 2000 to 2022 were calculated respectively. Then, the annual average soil erosion modulus of the entire basin, mountainous area and plain area of ​​the study area in the past 23 years was obtained statistically, and the erosion modulus was divided into 6 erosion grades according to the soil erosion classification and grading standard SL190-2007.

[0113] Step 4: Compare and analyze the differences in soil erosion modulus and soil erosion intensity levels between the mountainous and plain areas of the study area;

[0114] In terms of soil erosion modulus, the annual average modulus of soil erosion in the study area, including the entire basin and its mountainous and plain areas, was calculated from 2000 to 2022. Figure 3 As shown in the figure, it is found that the annual average modulus of soil erosion in the study area has shown an upward trend over the past 23 years, indicating that the soil erosion problem in the study area is becoming increasingly serious. Among them, the annual average soil erosion modulus of the entire basin is 159t / (km 2 a), 386 t / (km in mountainous areas 2 a), and 1.1 t / (km in plain areas 2 a), the soil erosion modulus in mountainous areas is much greater than that in plains, accounting for the vast majority of the annual average soil erosion modulus in the entire basin.

[0115] In terms of soil erosion intensity, this application selected three key years, 2000, 2011 and 2022, and drew a classification map of erosion intensity to visually show the comparison of soil erosion levels in the mountainous and plain areas of the study area, such as Figure 4 As shown. By comparing the data from 2000, 2011 and 2022, it can be observed that the distribution of soil erosion intensity in the study area shows a low in the southeast and high in the northwest, and the soil erosion intensity in the mountainous area is much higher than that in the plain area. In 2000, the plain area mainly experienced slight erosion, while the mountainous area was mainly mild erosion. By 2011, the mild erosion area in the mountainous area expanded, and the moderate erosion area also began to appear. By 2022, the mild and moderate erosion areas expanded further, and some areas in the mountainous area even reached strong erosion.

[0116] Step 5: Compare and analyze the differences in spatial distribution and temporal changes of driving factors of soil erosion RUSLE model between mountainous areas and plain areas in the study area;

[0117] This experiment focuses on the time span from 2000 to 2022. By applying the RUSLE model, the evolution trends of the key influencing factors such as rainfall erosivity factor R, soil and water conservation measures factor P, and vegetation and management factor C in the study area are analyzed. Since the soil erodibility factor K and slope length factor LS have a long change cycle and are relatively stable during the study period, their changes are not discussed. A classification map of the R, P, and C driving factors in the RUSLE model of the study area is made, and then the differences in the spatial distribution of soil erosion driving factors in the mountainous and plain areas of the study area are compared. At the same time, the data of R, P, and C driving factors in the entire basin, mountainous areas, and plains of the study area from 2000 to 2022 are statistically analyzed and a change map is drawn to compare the differences in the time changes of soil erosion driving factors in different terrains.

[0118] like Figure 5As shown in the figure, in terms of spatial distribution, the rainfall erosivity factor R in the plain area is significantly higher than that in the mountainous area. The greater rainfall intensity and flat terrain in the plain area promote the formation of surface runoff, thereby exacerbating soil erosion. In contrast, due to the rich natural vegetation in the mountainous area, the vegetation coverage and management factor C are relatively high, which provides a certain degree of protection for the soil and slows down the erosion process. However, in recent years, the C factor in the mountainous area has declined, exacerbating the erosion problem. In addition, the soil and water conservation measures factor P in the mountainous area is significantly higher than that in the plain area, which indicates that there are deficiencies in soil and water conservation in the mountainous area, increasing the risk of soil erosion.

[0119] like Figure 6 As shown in the figure, in terms of temporal changes, the change diagram of soil erosion influencing factors in the study area from 2000 to 2022 shows that the soil and water conservation measures factor P in the study area has decreased from 2000 to 2022, indicating that the soil and water conservation measures implemented in the region in recent years have achieved positive results, which can effectively control soil erosion, protect soil resources, and enhance the sustainable use of land. In contrast, the vegetation coverage and management factor C in the mountainous areas showed a decreasing trend, resulting in more serious soil erosion in the mountainous areas, while the vegetation coverage in the plains increased.

[0120] Step 6: Analyze the spatial differentiation characteristics of the driving factors of the RUSLE model of soil erosion in the mountainous and plain areas of the study area through geographic detector technology;

[0121] like Figure 7 As shown in Table 5, the influence of the driving factors of the RUSLE model on the spatial heterogeneity of soil erosion is quantified by factor detection, and the q values ​​of the driving factors of soil erosion in the mountainous and plain areas of the study area in 2022 are statistically analyzed. As shown in Table 5, the q value of the rainfall erosivity factor R in the mountainous area is 0.31348, which is much higher than 0.09105 in the plain area. It can be seen that rainfall plays a leading role in soil erosion in the mountainous area. The q value of the soil erodibility factor K in the mountainous area is 0.1957, which is also higher than 0.13072 in the plain area, reflecting that the soil in the mountainous area is more susceptible to erosion. In contrast, the q value of the soil and water conservation measures factor P in the plain area is 0.41394, which is significantly higher than 0.21563 in the mountainous area, indicating that in the plain area, soil and water conservation measures are more effective in slowing down soil erosion. The q value of the terrain factor LS in the plain area is 0.88596, which is slightly higher than 0.86135 in the mountainous area. Therefore, the influence of the terrain in the plain area on soil erosion is more significant. In addition, the q value of vegetation cover factor C in mountainous areas is 0.18959, slightly higher than 0.14551 in plain areas, indicating that vegetation cover has a certain soil protection effect in both areas, but its influence is relatively small.

[0122] Table 5

[0123]

[0124] like Figure 8 As shown in Table 6, the interactive detection explores the different interactions of the driving factors of the RUSLE model in soil erosion in mountainous areas and plain areas. The results in Tables 6 and 7 show that in mountainous areas, the interactive explanatory power of the rainfall erosivity factor R and the soil erodibility factor K reaches 0.88284, which is significantly higher than 0.73678 in the plain area, indicating that rainfall has a stronger driving effect on soil erosion in mountainous areas. In addition, the interactive explanatory power of the terrain factor LS and the soil and water conservation measures factor P in mountainous areas is 0.93208, which is also higher than 0.92607 in the plain area. This phenomenon may be closely related to the complex terrain conditions in mountainous areas. The complex terrain makes the effects of soil and water conservation measures more susceptible to terrain factors. In contrast, the interactive explanatory power of the vegetation cover factor C and the soil and water conservation measures factor P in mountainous areas is 0.53978, which is lower than 0.59901 in the plain area. This reflects that mountainous areas may face more challenges when implementing vegetation restoration measures, such as terrain restrictions or adverse effects of climatic conditions. In addition, the independent value of soil erodibility factor K in mountainous areas was 0.1957, which was higher than 0.13072 in plain areas, which may be related to the physical and chemical characteristics of mountainous soils, such as soil structures that are more susceptible to erosion. In summary, soil erosion in mountainous areas is significantly affected by rainfall and topographic factors, while in plain areas it is more dependent on the implementation of soil and water conservation measures and the characteristics of soil erodibility. These differences provide an important scientific basis for the formulation of regional and targeted soil protection and soil and water conservation strategies.

[0125] Table 6

[0126]

[0127] Table 7

[0128]

[0129] Step 7: Obtain human factors and natural factors, and analyze the temporal differentiation characteristics of natural and human factors of soil erosion in the whole watershed, mountainous area, and plain area of ​​the study area through the random forest regression model;

[0130] The relevant data of soil erosion in the mountainous and plain areas of the study area were collected and counted, mainly including natural factors (vegetation cover, temperature, rainfall) and human factors (GDP, urbanization area, population density). Then, the statistical data of each natural and human factor in each region from 2000 to 2022 were standardized to eliminate the influence of different dimensions and ranges. The standardized data were brought into the random forest regression model to calculate the importance ratio of each natural and human factor in the mountainous and plain areas of the study area over the past 23 years. Through comparative analysis, the temporal differentiation characteristics of natural and human factors in soil erosion in the mountainous and plain areas of the study area can be revealed.

[0131] like Fig. 9 As shown in the figure, the results of random forest calculation of the relative importance of each factor show that rainfall is the primary natural factor affecting soil erosion in the entire basin, and its explanation of the erosion process is as high as 55.3%, which is dominant. Temperature and vegetation coverage are equally important, but their effects are different in mountainous and plain areas. In mountainous areas, the effects of temperature and vegetation coverage on soil erosion are more significant, while in plain areas, the impact of urbanization area is more prominent. This shows that soil erosion in plain areas is closely related to the rapid urbanization process and land use changes in the region, while climatic conditions and vegetation conditions have a more direct impact on soil erosion in mountainous areas. Overall, natural factors are the main driving force of soil erosion, especially rainfall, while the contribution rate of human factors to the erosion modulus in plain areas is significantly higher than that in mountainous areas. Under different terrains, the degree of explanation of each factor on the spatial heterogeneity of soil erosion is different: the degree of explanation of rainfall erosivity R and urbanization area in plain areas is higher than that in mountainous areas; while the degree of explanation of temperature and vegetation coverage in mountainous areas is higher than that in plains.

[0132] Step 8: The driving intensity of natural and human factors of soil erosion in the mountainous and plain areas of the study area was analyzed through the partial least squares structural equation model (PLS-SEM).

[0133] First, the standardized data was imported. Then, a structural equation model was constructed to define the relationship between the variables, including the direct path relationship between natural factors (vegetation cover, temperature, and rainfall) in mountainous areas and the erosion modulus in mountainous areas, and the path relationship between anthropogenic factors (GDP, urbanization area, and population density) in mountainous areas and the erosion modulus in mountainous areas. Similarly, path relationships were established between natural factors (vegetation cover, temperature, and rainfall) in plains and the erosion modulus in plains, as well as between anthropogenic factors (GDP, urbanization area, and population density) in plains and the erosion modulus in plains. The model calculated the path coefficients of the influence of various natural and anthropogenic factors on the soil erosion modulus in mountainous and plain areas of the study area, and these coefficients reflected the driving strength of each influencing factor.

[0134] like Fig.10As shown in the figure, in mountainous areas, the natural factor (0.674) has a significantly higher impact on soil erosion than the anthropogenic factor (0.015). Among them, climate factors, including rainfall and temperature, play a positive role in promoting soil erosion. This is closely related to the increase in extreme weather events under the trend of global climate change. These extreme weather events have significantly aggravated soil erosion in the mountainous areas of the study area. At the same time, the reduction in vegetation cover has a significant negative impact on soil erosion in mountainous areas. The decline in vegetation coverage in mountainous areas has weakened the protective capacity of the soil, thereby aggravating the erosion problem. In contrast, the impact of anthropogenic factors in mountainous areas is relatively weak, which may be related to the relatively small number of human activities in mountainous areas. In plain areas, natural factors also play a major role in soil erosion, and rainfall is the key factor. Due to the high rainfall intensity and flat terrain in plain areas, rainwater easily forms runoff on the surface, thereby accelerating soil erosion. Vegetation cover and temperature also indirectly affect soil erosion by affecting soil stability and water retention capacity. In addition, anthropogenic factors, especially urbanization, have a significant impact on soil erosion in plain areas. Urbanization activities, such as land use change and infrastructure construction, can lead to the destruction of soil structure and reduce vegetation cover, thereby increasing the risk of soil erosion.

[0135] In general, natural factors play a key role in soil erosion in both mountainous and plain areas, but the impact of human factors is more significant in plain areas. Therefore, when formulating soil erosion prevention and control strategies, it is necessary to take into account the specific conditions of different terrain areas and adopt comprehensive management measures.

[0136] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. An analysis system of spatiotemporal driving forces of soil erosion in multiple terrains, characterized in that: The following steps are involved: Step 1, collect relevant data on soil erosion in the study area; Step 2: Classify the terrain in the study area based on the digital elevation model (DEM) and slope of the study area; Step 3, after preprocessing the collected data on soil erosion in the study area, the soil erosion modulus of different terrain areas in the study area is calculated using the modified universal soil loss equation model, namely the RUSLE model, and the calculation results are graded for erosion intensity; Step 4: Based on the results of step 3, compare and analyze the differences in soil erosion modulus and soil erosion intensity levels in different terrain areas of the study area; Step 5, extract the driving factors in the RUSLE model, and compare and analyze the differences in spatial distribution and temporal changes of the driving factors of the RUSLE model of soil erosion in different terrain areas of the study area; Step 6, the spatial differentiation characteristics of the driving factors of the RUSLE model of soil erosion in different terrain areas of the study area were analyzed through geographic detectors; Step 7, obtain human factors and natural factors, and analyze the temporal differentiation characteristics of natural and human factors of soil erosion in different terrain areas of the study area through random forest regression model; Step 8: The driving intensity of natural and anthropogenic factors of soil erosion in different terrain areas of the study area was analyzed through the partial least squares structural equation model (PLS-SEM).

2. The analysis system of spatiotemporal driving force of soil erosion in multiple terrains according to claim 1 is characterized in that: The step 2 comprises the following steps: Step 2-1, load and clip DEM data: first load the DEM data and study area, then use spatial analysis tools to clip the DEM to the study area. Step 2-2, calculate slope: Use the slope tool to calculate the slope of the study area. Step 2-3, slope grading and reclassification: Divide the grade and reclassify the terrain according to the slope calculation results.

3. The analysis system of spatiotemporal driving force of soil erosion in multiple terrains according to claim 1 is characterized in that: The step 3 comprises the following steps: Step 3-1, calculate the soil erosion modulus by using the modified universal soil loss equation RUSLE model, specifically: Where: A is the soil erosion modulus; R is the rainfall erosivity factor; K is the soil erodibility factor; LS is the slope length and slope factor; C is the surface vegetation coverage and management factor; P is the soil and water conservation measures factor; Among them, the rainfall erosivity factor R is the rainfall erosivity of each year, and the calculation method is: Where P i is the monthly rainfall; P is the annual rainfall; The soil erodibility factor K is an indicator of the sensitivity of soil properties to erosion. The calculation method is: In the formula, It represents the soil erodibility factor value calculated by the EPIC model, i.e. the cognitive model. K represents the soil erodibility factor value after correction. SAN represents sand particles. SIL represents silt particles. CLA represents clay particles. OM represents organic matter. The K value reflects the difficulty of soil separation, erosion and transportation by rainfall erosion. The larger the soil K value, the worse the soil's ability to resist water erosion and the more susceptible it is to erosion. The slope length factor LS is divided into the slope length factor L and the slope factor S. The calculation method is: Among them, λ is the projection length of the grid unit; m is the slope length effect index; θ is the slope extracted from DEM; Vegetation coverage and management factor C is used to reflect the impact of vegetation coverage and management measures on soil loss. The value is between 0 and 1. The calculation method is: Where f is the vegetation coverage; C is the vegetation coverage and management factor value; NDVI is the normalized vegetation index using the multi-year monthly average; MAX and NDVI MIN is the maximum and minimum value of the normalized difference vegetation index; The soil and water conservation measures factor P is the ratio of soil loss after taking soil conservation measures to soil loss when planting along the slope. Its value is between 0 and 1. 0 represents an area where no soil and water erosion will occur, and 1 represents an area where no soil and water conservation measures are taken. Step 3-2, classify the erosion intensity based on the soil erosion modulus calculation results.

4. The analysis system of spatiotemporal driving force of soil erosion in multiple terrains according to claim 3 is characterized in that: In step 4, based on the soil erosion modulus and intensity classification of the continuous time series of different terrain areas in the study area obtained in step 3, the soil erosion modulus of each area over the years is counted, and a curve graph is drawn to calculate the average soil erosion modulus over the years, so as to compare the differences in soil erosion modulus in different terrain areas in the study area, and at the same time draw a soil erosion intensity classification map of the study area to further compare the differences in soil erosion intensity levels in different terrain areas in the study area.

5. The analysis system of spatiotemporal driving force of soil erosion in multiple terrains according to claim 4 is characterized in that: In step 5, the driving factors in the RUSLE model are extracted, and a grading map of the driving factors of the RUSLE model in the study area is prepared to compare the differences in the spatial distribution of the driving factors of soil erosion in different terrains, and the data of the driving factors of the RUSLE model in the study area for many years are counted and a change map is drawn to compare the differences in the temporal changes of the driving factors of soil erosion in different terrains; Among them, the driving factors of the RUSLE model include rainfall erosivity factor, soil erodibility factor, slope gradient and length factor, vegetation cover factor and soil and water conservation factor.

6. The analysis system of spatiotemporal driving force of soil erosion in multiple terrains according to claim 5 is characterized in that: The step 6 is specifically as follows: quantifying the influence of a single factor on the spatial heterogeneity of soil erosion under different terrains through factor detection, and exploring the interaction between the driving factors of the RUSLE model under different terrains through interactive detection, so as to obtain the mechanism by which these factors jointly affect soil erosion under different terrains.

7. The analysis system of spatiotemporal driving force of soil erosion in multiple terrains according to claim 1 is characterized in that: In step 7, human factors and natural factors are obtained, and based on a continuous time series, the statistical data of each factor under different terrain conditions are standardized, and then substituted into the random forest model to obtain the importance ratio of each natural and human factors based on the time series, and analyze the temporal differentiation characteristics of natural factors and human factors of soil erosion in different terrains; Among them, natural factors include rainfall, temperature, and vegetation coverage; human factors include population density, GDP, and urbanization area.

8. The analysis system of spatiotemporal driving force of soil erosion in multiple terrains according to claim 7 is characterized in that: In step 8, the standardized continuous time series data of different terrain areas are imported into the constructed partial least squares structural equation model PLS-SEM to obtain the path coefficients of the influence of natural factors and human factors on the soil erosion modulus of different terrains, so as to analyze the driving intensity of natural factors and human factors of soil erosion in different terrains.

Images