Level collapse occurrence risk prediction method based on connectivity and geographic detector

Abstract

The invention discloses a slope collapse occurrence risk prediction method based on connectivity and a geographic detector, and the method comprises the steps: converting a comprehensive soil erosion index and an impact factor into characteristic parameters of a type variable in a discretization mode, carrying out the multi-collinearity analysis of all characteristic parameters, and filtering out the collinearity characteristic parameters; calculating the explanation intensity of the collapse occurrence risk of the characteristic parameters in the target area and the collapse occurrence frequency of sample points in different grids by using a geographic detector, and accumulating the product of the collapse occurrence explanation intensity and the occurrence frequency corresponding to different characteristic parameters in each grid; and obtaining a slope collapse risk space prediction distribution map about the target area. The method has the beneficial effects that by combining soil erosion and sediment transportation potential, the interpretation strength of each characteristic parameter on collapse slope occurrence is analyzed by utilizing the geographic detector, so that a high-accuracy collapse slope risk assessment result is obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of soil erosion and disaster prevention and control, and in particular to a method for predicting the risk of landslides based on connectivity and geographic detectors. Background Art

[0002] As a common geological disaster, gully collapse often has a serious impact on human life, property and ecological environment. With climate change and the intensification of human activities, the frequency and intensity of gully collapse are gradually increasing, especially in mountainous and hilly areas. Traditional gully collapse risk assessment methods often focus on static factor analysis, such as using information volume method, bivariate entropy information method, logistic regression analysis method, etc. The factors considered mainly include climate (precipitation, temperature), geology, soil, vegetation, terrain (slope, elevation), and human activities (land use). There is a lack of comprehensive understanding of the interaction between soil erosion and sediment transport, and the accuracy and reliability of gully collapse risk assessment are too low. Summary of the invention

[0003] In view of the above problems, the present invention proposes a method for predicting the risk of landslides based on connectivity and geographic detectors, which mainly solves the problem of low accuracy of existing landslide risk assessment methods based on static factor analysis.

[0004] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0005] A method for predicting the risk of landslide based on connectivity and geographic detectors includes the following steps:

[0006] Obtaining a soil surface erosion spatial distribution map and a sediment connectivity index of a target area, and normalizing the soil surface erosion spatial distribution map and the sediment connectivity index to obtain a soil surface erosion layer and a connectivity index layer;

[0007] After the soil surface erosion layer and the connectivity index layer are superimposed, the comprehensive soil erosion index of each area in the target area is calculated;

[0008] Acquire the spatial distribution information of the ridge collapse area in the target area, randomly collect sample points in the target area, and mark the sample points according to preset rules, and define the sample points as two types: points with ridge collapse and points without ridge collapse;

[0009] Obtain influencing factors that are strongly correlated with landslides;

[0010] The comprehensive soil erosion index and the influencing factors are converted into characteristic parameters of type variables by discretization, all the characteristic parameters are subjected to multicollinearity analysis, and characteristic parameters with collinearity are filtered out;

[0011] The geographic detector is used to calculate the explanatory strength of the characteristic parameters of the risk of landslides in the target area and the frequency of landslides at sample points in different grids, and the product of the explanatory strength and frequency of landslides corresponding to different characteristic parameters in each grid is accumulated to obtain the spatial prediction distribution map of landslide risks in the target area.

[0012] In some embodiments, the process of obtaining the soil surface erosion spatial distribution map includes: taking a digital elevation model, land use data and meteorological data of a target area as input, using an empirical soil erosion estimation model as a framework, and outputting the soil surface erosion spatial distribution map.

[0013] In some embodiments, the process of obtaining the connectivity index includes: evaluating the sediment transport potential based on the acquired sediment formation area and sediment transport area data of the target area to obtain the connectivity index.

[0014] In some implementations, the preset rule is: assigning a value of 1 to the sample points with a risk of landslide, and assigning a value of 0 to the sample points without a risk of landslide.

[0015] In some embodiments, the influencing factors include altitude, slope, soil type, lithology and / or land use.

[0016] In some embodiments, the frequency of landslide occurrence represents the proportion of landslide points to the sample points.

[0017] In some embodiments, the explanatory strength is the importance of the impact factor.

[0018] The beneficial effect of the present invention is that by combining soil erosion and sediment transport potential, the geographic detector is used to analyze the interpretation intensity of each characteristic parameter on the occurrence of landslide, so as to obtain a highly accurate landslide risk assessment result. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A schematic flow chart of a method for predicting the risk of landslide based on connectivity and geographic detectors disclosed in an embodiment of the present invention;

[0020] Figure 2 A schematic diagram of a soil erosion comprehensive index disclosed in an embodiment of the present invention;

[0021] Figure 3 This is a spatial prediction distribution map of hillock risk based on a geographic detector disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present invention clearer and more specific, the content of the present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods. It is understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for the convenience of description, only the parts related to the present invention are shown in the accompanying drawings, rather than all the contents.

[0023] This embodiment proposes a method for predicting the risk of landslides based on connectivity and geographic detectors. Figure 1 As shown, the following steps 1-6 are included:

[0024] Step 1: obtain the soil surface erosion spatial distribution map and sediment connectivity index of the target area, and perform maximum absolute value normalization on the soil surface erosion spatial distribution map and sediment connectivity index to obtain the soil surface erosion layer and connectivity index layer.

[0025] In this embodiment, the digital elevation model, land use data and meteorological data (annual rainfall and annual temperature) of the target area are used as input, and the empirical soil erosion estimation model is used as the framework to output the spatial distribution map of soil surface erosion, wherein the empirical soil erosion estimation model can adopt the RUSLE model or the MUSLE model. RUSLE (Revised Universal Soil Loss Equation) is a revised universal soil loss equation, which is a recognized and widely used empirical soil erosion estimation model. MUSLE (Modified Universal Soil Loss Equation) is a watershed scale soil erosion calculation model developed by the Soil Conservation Institute of the United States. The RUSLE model mainly estimates the amount of soil erosion based on the rainfall erosivity factor, while the MUSLE model pays more attention to the impact of surface runoff on soil erosion. In this embodiment, the method for obtaining the spatial distribution map of soil surface erosion is not limited.

[0026] The acquisition process of the connectivity index includes: obtaining the connectivity index according to the acquired data of the sediment formation area (Dup) and the sediment transport area (Ddn) of the target area, and evaluating the sediment transport potential. Sediment connectivity refers to the physical connection of sediment from one area or location to another area or location through the channel system and the potential for specific particles to move in the system. It reflects the sediment cascade relationship between different landforms or landscape units in the basin, and is an important indicator for measuring the potential for sediment movement in geomorphic units. The calculation methods for the sediment connectivity index mainly include the index method, the model simulation method, and the graph theory method. Among them, the connectivity index (IC) in the index method is the most widely used, which can characterize the potential connection between the sediment source and sink in the basin. The model simulation rule simulates the sediment source and sink of the basin erosion by constructing a mathematical model, and quantitatively characterizes the sediment connectivity. The graph theory rule can intuitively display the spatial distribution and dynamic changes of the sediment cascade between different landscape units, but the applicable spatial scale is limited. In this embodiment, the method for obtaining the sediment connectivity index is not limited.

[0027] Step 2: After the soil surface erosion layer and the connectivity index layer are superimposed, the comprehensive soil erosion index of each area in the target area is calculated. The core of this step is to incorporate the sediment transport potential represented by sediment connectivity into the soil erosion system, such as Figure 2 shown.

[0028] Step 3, obtain the spatial distribution information of the ridge collapse area in the target area, randomly collect sample points in the target area, and mark the sample points according to preset rules, and define the sample points as two types: with ridge collapse points and without ridge collapse points.

[0029] The preset rule is: assign a value of 1 to sample points with landslide risk and a value of 0 to sample points without landslide risk. This method treats the occurrence of landslide as a binary variable. This approach helps to identify and quantify the key factors that affect the occurrence of landslide and provides a basis for decision-making.

[0030] Step 4: Obtain influencing factors that are strongly correlated with the occurrence of landslides.

[0031] Influencing factors include altitude, slope, soil type, lithology and / or land use.

[0032] Step 5: The comprehensive soil erosion index and influencing factors are converted into characteristic parameters of type variables by discretization, all characteristic parameters are subjected to multicollinearity analysis, and characteristic parameters with collinearity are filtered out.

[0033] In step 5, the instability of the model can be reduced and the robustness of the model can be improved through the presence of collinear characteristic parameters. In addition, the individual impact of each influencing factor on the subsequent output prediction distribution results can be more clearly understood, thereby improving the interpretability of the model.

[0034] Step 6: Use the geographic detector to calculate the explanatory strength of the characteristic parameters of the risk of landslides in the target area and the frequency of landslides at sample points in different grids, and accumulate the product of the explanatory strength and frequency of landslides corresponding to different characteristic parameters in each grid to obtain the spatial prediction distribution map of the risk of landslides in the target area, such as Figure 3 shown.

[0035] The above-mentioned landslide occurrence frequency (r) represents the proportion of landslide points to sample points, and the interpretation strength is the importance of the influencing factor (q value).

[0036] In this embodiment, by detecting the explanatory power of different characteristic parameters on geographical phenomena, the geographical detector is used to effectively identify the key factors that affect the occurrence of gully collapse. Specifically, the geographical detector is used to analyze the relationship between each factor and the occurrence of gully collapse, and the risk assessment result is obtained by accumulating the importance q value (explanatory strength) of each factor and the product of the gully collapse frequency of each factor at each grid position. The risk of gully collapse in each grid can be quantified by multiplying the q value and the gully collapse frequency, aiming to identify areas with higher risk of gully collapse spatially, and provide a scientific basis for gully collapse prevention and decision-making.

[0037] Taking Huacheng Town as an example, the phenomenon of ridge collapse occurs frequently in the area, which seriously affects the local ecological environment and economic development. Soil erosion in the area is mainly caused by hydraulic and gravity effects, resulting in significant soil loss. In order to effectively assess the risk of ridge collapse, the ridge collapse risk prediction method based on connectivity and geographic detectors provided in this embodiment is used for analysis. High-resolution remote sensing images are used to identify 2146 ridge collapse sample points through machine learning models, and 1806 sample points where ridge collapse did not occur are manually added to establish a database of samples where ridge collapse occurred and did not occur. The influencing factors include altitude, slope, aspect, lithology, rainfall, temperature, vegetation, and soil erosion comprehensive index. The results show that the factor coefficients of soil erosion comprehensive index, aspect, and elevation rank in the top three, and the area under the operator operating characteristic curve (AUC) value is 0.851. The research results can provide a reference for soil and water conservation and ridge collapse prevention and control in the area.

[0038] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable ordinary technicians in the field to understand the content of the present invention and implement it accordingly, and they cannot be used to limit the protection scope of the present invention. Any equivalent changes or modifications made based on the essence of the content of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for predicting the risk of landslide based on connectivity and geographic detectors, characterized in that: The following steps are involved: Obtaining a soil surface erosion spatial distribution map and a sediment connectivity index of a target area, and normalizing the soil surface erosion spatial distribution map and the sediment connectivity index to obtain a soil surface erosion layer and a connectivity index layer; After the soil surface erosion layer and the connectivity index layer are superimposed, the comprehensive soil erosion index of each area in the target area is calculated; Acquire the spatial distribution information of the ridge collapse area in the target area, randomly collect sample points in the target area, and mark the sample points according to preset rules, and define the sample points as two types: points with ridge collapse and points without ridge collapse; Obtain influencing factors that are strongly correlated with landslides; The comprehensive soil erosion index and the influencing factors are converted into characteristic parameters of type variables by discretization, all the characteristic parameters are subjected to multicollinearity analysis, and characteristic parameters with collinearity are filtered out; The geographic detector is used to calculate the explanatory strength of the characteristic parameters of the risk of landslides in the target area and the frequency of landslides at sample points in different grids, and the product of the explanatory strength and frequency of landslides corresponding to different characteristic parameters in each grid is accumulated to obtain the spatial prediction distribution map of landslide risks in the target area.

2. The method for predicting the risk of landslide based on connectivity and geographic detectors according to claim 1, characterized in that: The process of acquiring the soil surface erosion spatial distribution map includes: taking the digital elevation model, land use data and meteorological data of the target area as input, using the empirical soil erosion estimation model as the framework, and outputting the soil surface erosion spatial distribution map.

3. The method for predicting the risk of landslide based on connectivity and geographic detectors according to claim 1, characterized in that: The process of obtaining the connectivity index includes: evaluating the sediment transport potential according to the obtained sediment formation area and sediment transport area data of the target area to obtain the connectivity index.

4. The method for predicting the risk of landslide based on connectivity and geographic detectors according to claim 1, characterized in that: The preset rule is: assigning a value of 1 to the sample points with landslide risk, and assigning a value of 0 to the sample points without landslide risk.

5. The method for predicting the risk of landslide based on connectivity and geographic detectors according to claim 1, characterized in that: The influencing factors include altitude, slope, soil type, lithology and / or land use.

6. The method for predicting the risk of landslide based on connectivity and geographic detectors according to claim 1, characterized in that: The frequency of landslide occurrence indicates the proportion of landslide points to the sample points.

7. The method for predicting the risk of landslide based on connectivity and geographic detectors according to claim 1, characterized in that: The explanatory strength is the importance of the impact factor.

Images