Method for identifying service dominant factors of ecological system

Abstract

The invention discloses an ecological system service dominant factor identification method. The method comprises the following steps: 1) establishing a basic ecological system service influence factor database; 2) selecting key evaluation indexes for evaluating the target area and quantifying the key evaluation indexes; 3) constructing a multiple linear regression model, and introducing a nonlinear term correction model; 4) performing primary screening on influence factors in the correction model based on a variance expansion factor VIF; 5) further screening the influence factors by using optimal subset feature screening, and establishing a corresponding influence factor index system; and 6) carrying out local regression analysis by adopting an improved multi-scale geographically weighted regression MGWR model, and identifying the ecosystem service dominant factor according to an obtained regression coefficient value. The method disclosed by the invention not only can reveal the dominant factors of ecosystem services in different areas and on different scales, but also ensures that the identified dominant factors have statistical significance through significance test on the regression coefficient, and enhances the credibility and practicability of an analysis result.

Description

Technical Field

[0001] The present invention relates to the field of ecological technology, and in particular to a method for identifying leading factors of ecosystem services. Background Art

[0002] Ecosystem services are complex natural systems encompassing the relationships between organisms, the environment, and their interactions. The functions and services of ecosystem services (such as material provision, soil conservation, water conservation, carbon sequestration, and recreation) are crucial for maintaining biodiversity, regulating climate, and providing the resources and values ​​necessary for human life. However, ecosystem services are influenced by multiple factors over time, including climate change, human activities, and land-use change. These factors interact through complex interactions to determine the structure and function of ecosystem services. In recent years, with the continuous development of remote sensing technology, big data analysis, spatial statistics, and machine learning methods, the identification of the dominant factors of ecosystem services has become more precise and efficient.

[0003] However, existing methods for identifying the dominant factors of ecosystem services have significant limitations. Most rely on single models (such as linear regression or machine learning algorithms), which struggle to capture complex nonlinear relationships and overlook the significant differences in the dominant factors of ecosystem services across regions and scales. Furthermore, the factor screening process lacks systematicity and scientificity, failing to effectively analyze from the global to the local level. This results in limited regional applicability and reduced accuracy and credibility of the results. These limitations make existing technologies difficult to adapt to complex ecosystem environments, hindering their dissemination and practical application. Therefore, a systematic approach based on a combination of global and local analyses is urgently needed to address these issues. Summary of the Invention

[0004] In order to overcome the shortcomings of the existing technology, the present invention provides a method for identifying the dominant factors of ecosystem services. By capturing the nonlinear relationship between influencing factors, introducing nonlinear terms, and establishing an influencing factor index system based on global to local analysis, the improved MGWR model is adopted to identify the dominant factors in different regions and scales based on the significance test of its regression coefficient. This not only ensures the independence of the explanatory variables, but also ensures that the identified dominant factors are statistically significant, thereby enhancing the credibility and practicality of the analysis results.

[0005] In order to achieve the above-mentioned purpose, the present invention adopts the following technical solutions.

[0006] A method for identifying dominant factors of ecosystem services comprises the following steps: Step S1: determine the target area and research period, collect relevant data, and establish a database of factors affecting basic ecosystem services; The target area is selected from administrative units, river basins or urban agglomerations; Step S2: Select key evaluation indicators for the target area and quantify them, and use the quantified results as the dependent variable data set; Step S3: construct a multiple linear regression model to capture the nonlinear relationship between influencing factors, and introduce nonlinear terms to modify the multiple linear regression model; Step S4: Perform a preliminary screening of the influencing factors in the revised model based on the variance inflation factor (VIF) to eliminate influencing factors with multicollinearity; Step S5: further screen the impact factors using the optimal subset feature screening to establish a corresponding impact factor index system; Step S6: Based on the influencing factor indicator system established in step S5, an improved multi-scale geographically weighted regression (MGWR) model is used to perform local regression analysis. The dominant factors of ecosystem services are identified based on the obtained regression coefficient values, and their spatial distribution is visualized in visualization software to draw a spatial distribution map of the dominant factors of ecosystem services.

[0007] Specifically, the ecosystem service influencing factor database in step S1 includes meteorological data, remote sensing data, socioeconomic data, land use data and soil property data; the meteorological data includes annual average precipitation, annual average temperature and annual evaporation; the remote sensing data includes DEM digital elevation, normalized difference vegetation index NDVI and vegetation net primary productivity NPP; the socioeconomic data includes gross domestic product, population density, tertiary industry output value and night lights; the land use data includes cultivated land area, forest land area, grassland area, water area and construction land area; the soil property data includes soil depth and the content of sand, silt, clay and organic carbon.

[0008] Specifically, in step S2, key evaluation indicators for the target area are selected and quantified. The key evaluation indicators include the supply service, regulation service, cultural service and support service of ecosystem services. The physical quantity quantitative indicators corresponding to each type of service include: grain output, water conservation, soil conservation, carbon sequestration, water purification, cultural service suitability index and habitat quality index; based on the research target area, data availability and ecological characteristics of the target area determined in step S1, representative ecosystem service indicators are selected for quantification, and the quantification results are used as the dependent variable data set for local regression analysis in subsequent steps.

[0009] Specifically, the process of introducing nonlinear terms to modify the multiple linear regression model in step S3 is as follows: Step S31: construct a multiple linear regression model and capture the nonlinear relationship between influencing factors through scatter plots or residual plots; Step S32: Based on the linear and nonlinear relationships among the influencing factors, a nonlinear term is introduced as an influencing factor to correct the multiple linear regression model. The mathematical expression of the correction process is as follows: ; In the above formula, is the dependent variable; is the intercept; It is i The coefficient of the independent variable; It is i independent variables; i =1,2,3,…,n, j =1,2,3,…,n, where n is a positive integer; is the nonlinear term introduced, is the coefficient of the nonlinear term, is an independent random error term.

[0010] Specifically, in step S4, the VIF value between each influencing factor is calculated in R language. If the VIF value is greater than 10, it indicates that multicollinearity exists between the corresponding influencing factors, and the corresponding influencing factors are eliminated.

[0011] Specifically, in step S5, the influencing factors are further screened using optimal subset feature screening, and the Mallows Cp criterion is selected as the analysis indicator. The smaller the Cp value, the higher the accuracy of the model. The indicator with the smallest Cp value is selected to construct the influencing factor indicator system.

[0012] Specifically, in step S6, based on the impact factor index system established in step S5, the improved multi-scale geographically weighted regression MGWR model is used to perform local regression analysis, and the obtained regression coefficients are used to calculate the local regression coefficients. β Value and θ The process of identifying the dominant factors of ecosystem services is as follows: Step S61: using the physical quantity of ecosystem services quantified in step S2 as the dependent variable, inputting the improved multi-scale geographically weighted regression (MGWR) model to perform local regression analysis; Step S62: Based on the regression analysis results, use the if function to screen the significance level p value of each independent variable. If p < 0.05 is satisfied, the regression coefficient corresponding to the independent variable is returned. If not satisfied, a value smaller than the minimum regression coefficient is returned. Step S63: Sort the values ​​returned by each independent variable to obtain the importance ranking of the influencing factors, thereby identifying the dominant factors of ecological services.

[0013] Furthermore, the improved multi-scale geographically weighted regression (MGWR) model captures the nonlinear relationship between influencing factors and introduces nonlinear terms for comprehensive analysis. It can find the optimal bandwidth for each linear and nonlinear explanatory variable, thereby effectively capturing the spatial heterogeneity characteristics of each influencing factor at different geographic spatial scales and identifying the dominant influencing factors of ecosystem services. The expression of the improved multi-scale geographically weighted regression (MGWR) model is as follows: ; ; In the above formula, Y i is the spatial position The dependent variable at is the spatial position The independent variable at is the spatial position The nonlinear term at ; is the intercept term of the regression relationship; the coefficient and is the spatial position A continuous function of 、 in Indicates the k The optimal bandwidth for each independent variable; is an independent random error term.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. By introducing nonlinear terms in the process of identifying the dominant factors of ecosystem services, the method of the present invention can not only more accurately capture the complex interactions between influencing factors and potential nonlinear mechanisms, thereby improving the model's goodness of fit and explanatory power, but also enhance the model's adaptability to heterogeneity in different regions and spatial scales, effectively avoiding misjudgments and omissions that may be caused by linear models; in addition, the improved MGWR model can not only reveal the dominant factors of ecosystem services in different regions and scales, but also ensure that the identified dominant factors are statistically significant through significance tests of regression coefficients, thereby improving the credibility and practical value of the analysis results.

[0015] 2. The present invention adopts an analysis method from global to local to accurately and efficiently identify the dominant factors of ecosystem services, and then visualizes the spatial distribution of the dominant factors of ecosystem services to draw a spatial distribution map. It can provide accurate data support and scientific basis for ecological protection and governance, and also provides new ideas and methods for ecological research. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1This is a flow chart of a method for identifying dominant factors of ecosystem services according to the present invention; Figure 2 This is an example diagram of the ecosystem service impact factor database in an embodiment of the present invention; Figure 3 This is the spatial distribution map of the dominant factors of ecosystem services in the Yangtze River Economic Belt drawn according to the method of the present invention. DETAILED DESCRIPTION

[0017] In order to facilitate those skilled in the art to understand and implement the present invention, each step of the method proposed in the present invention is described in detail below. It should be understood that these embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms also fall within the scope defined by the claims appended hereto.

[0018] Example 1 like Figure 1 As shown, the present invention discloses a method for identifying the dominant factors of ecosystem services, comprising the following steps: Step S1: determine the target area and research period, collect relevant data, and establish a database of factors affecting basic ecosystem services; The target area is selected from administrative units, river basins or urban agglomerations; Step S2: Select key evaluation indicators for the target area and quantify them, and use the quantified results as the dependent variable data set; Step S3: construct a multiple linear regression model to capture the nonlinear relationship between influencing factors, and introduce nonlinear terms to modify the multiple linear regression model; Step S4: Perform a preliminary screening of the influencing factors in the revised model based on the variance inflation factor (VIF) to eliminate influencing factors with multicollinearity; Step S5: further screen the impact factors using the optimal subset feature screening to establish a corresponding impact factor index system; Step S6: Based on the influencing factor indicator system established in step S5, an improved multi-scale geographically weighted regression (MGWR) model is used to perform local regression analysis. The dominant factors of ecosystem services are identified based on the obtained regression coefficient values, and their spatial distribution is visualized in visualization software to draw a spatial distribution map of the dominant factors of ecosystem services.

[0019] Specifically, the ecosystem service influencing factor database in step S1 includes meteorological data, remote sensing data, socioeconomic data, land use data and soil property data; the meteorological data includes annual average precipitation, annual average temperature and annual evaporation; the remote sensing data includes DEM digital elevation, normalized difference vegetation index NDVI and vegetation net primary productivity NPP; the socioeconomic data includes gross domestic product, population density, tertiary industry output value and night lights; the land use data includes cultivated land area, forest land area, grassland area, water area and construction land area; the soil property data includes soil depth and the content of sand, silt, clay and organic carbon.

[0020] Specifically, in step S2, key evaluation indicators for the target area are selected and quantified. The key evaluation indicators include the supply service, regulation service, cultural service and support service of ecosystem services. The physical quantity quantitative indicators corresponding to each type of service include: grain output, water conservation, soil conservation, carbon sequestration, water purification, cultural service suitability index and habitat quality index; based on the research target area, data availability and ecological characteristics of the target area determined in step S1, representative ecosystem service indicators are selected for quantification, and the quantification results are used as the dependent variable data set for local regression analysis in subsequent steps.

[0021] Specifically, the process of introducing nonlinear terms to modify the multiple linear regression model in step S3 is as follows: Step S31: construct a multiple linear regression model and capture the nonlinear relationship between influencing factors through scatter plots or residual plots; Step S32: Based on the linear and nonlinear relationships among the influencing factors, a nonlinear term is introduced as an influencing factor to correct the multiple linear regression model. The mathematical expression of the correction process is as follows: ; In the above formula, is the dependent variable; is the intercept; It is i The coefficient of the independent variable; It is i independent variables; i =1,2,3,…,n, j =1,2,3,…,n, where n is a positive integer; is the nonlinear term introduced, is the coefficient of the nonlinear term, is an independent random error term.

[0022] Specifically, in step S4, the VIF value between each influencing factor is calculated in R language. If the VIF value is greater than 10, it indicates that multicollinearity exists between the corresponding influencing factors, and the corresponding influencing factors are eliminated.

[0023] Specifically, in step S5, the influencing factors are further screened using optimal subset feature screening, and the Mallows Cp criterion is selected as the analysis indicator. The smaller the Cp value, the higher the accuracy of the model. The indicator with the smallest Cp value is selected to construct the influencing factor indicator system.

[0024] Specifically, in step S6, based on the impact factor index system established in step S5, the improved multi-scale geographically weighted regression MGWR model is used to perform local regression analysis, and the obtained regression coefficients are used to calculate the local regression coefficients. β Value and θ The process of identifying the dominant factors of ecosystem services is as follows: Step S61: using the physical quantity of ecosystem services quantified in step S2 as the dependent variable, inputting the improved multi-scale geographically weighted regression (MGWR) model to perform local regression analysis; Step S62: Based on the regression analysis results, use the if function to screen the significance level p value of each independent variable. If p < 0.05 is satisfied, the regression coefficient corresponding to the independent variable is returned. If not satisfied, a value smaller than the minimum regression coefficient is returned. Step S63: Sort the values ​​returned by each independent variable to obtain the importance ranking of the influencing factors, thereby identifying the dominant factors of ecological services.

[0025] Furthermore, the improved multi-scale geographically weighted regression (MGWR) model captures the nonlinear relationship between influencing factors and introduces nonlinear terms for comprehensive analysis. It can find the optimal bandwidth for each linear and nonlinear explanatory variable, thereby effectively capturing the spatial heterogeneity characteristics of each influencing factor at different geographic spatial scales and identifying the dominant influencing factors of ecosystem services. The expression of the improved multi-scale geographically weighted regression (MGWR) model is as follows: ; ; In the above formula, Y i is the spatial position The dependent variable at is the spatial position The independent variable at is the spatial position The nonlinear term at ; is the intercept term of the regression relationship; the coefficient and is the spatial position A continuous function of 、 in Indicates the k The optimal bandwidth for each independent variable; is an independent random error term.

[0026] The following example of the process of identifying the leading factors of carbon storage services in the Yangtze River Economic Belt is used to verify the credibility and practicality of the method of the present invention.

[0027] Step S1: determine the target area and research period, collect relevant data, and establish a database of ecosystem service impact factors; In this example, the target area is the Yangtze River Economic Belt, the research period is 2000-2020, the specific research years are 2000, 2005, 2010, 2015 and 2020, and the research scale is the county scale; like Figure 2 The following figure shows the ecosystem service impact factor database established in this example, which includes meteorological data, remote sensing data, socioeconomic data, and land use data. Step S2: Select key evaluation indicators for the target area and quantify them; In this example, carbon storage services are used as key evaluation indicators for the target area, and these key evaluation indicators are quantified. The grid scale in the quantification process is 1km. Step S3: construct a multiple linear regression model to capture the nonlinear relationship between influencing factors, and introduce nonlinear terms to modify the multiple linear regression model; Step S4: Perform a preliminary screening of the influencing factors in the revised model based on the variance inflation factor (VIF) to eliminate influencing factors with multicollinearity; Step S5: further screen the impact factors using the optimal subset feature screening to establish a corresponding impact factor index system; The carbon sequestration influencing factor index system constructed in this example is shown in Table 1 below: Table 1. Index system of factors affecting carbon sequestration ; Step S6: Based on the carbon sequestration influencing factor index system established in step S5, the improved multi-scale geographically weighted regression MGWR model is used to perform local regression analysis. β and θ The dominant factors of ecosystem services were identified by using the value of 0.05 (satisfying p < 0.05), and their spatial visualization was performed in ArcGIS 10.8.2 software, as shown in Figure 2. Figure 3 As shown in the figure, the spatial distribution map of the dominant factors of carbon storage services is obtained.

[0028] Through spatial visualization, we can accurately identify the spatial distribution characteristics of the dominant factors in different regions and their impact ranges; Figure 3 The results clearly show that the primary influencing factors of carbon storage services vary across regions, demonstrating significant spatial heterogeneity. The primary influencing factor for carbon sequestration is the Normalized Difference Vegetation Index (NDVI). Using the second dominant factor as the classification basis, the middle and lower reaches of the Yangtze River are divided into two regions: the western region is primarily influenced by topography (DEM and slope), and the eastern region is primarily influenced by human activities (forest area changes and nighttime lighting). The spatial distribution maps of the dominant factors of ecosystem services generated using this method can provide accurate data support and a scientific basis for ecological protection and governance.

[0029] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any other manner. Any person skilled in the art may utilize the above-disclosed technical content to modify or modify the present invention into equivalent embodiments. However, any simple modifications, equivalent variations, and modifications to the above embodiments that do not depart from the technical content of the present invention and are based on the technical essence of the present invention remain within the scope of protection of the present invention.

Claims

1. A method for identifying dominant factors of ecosystem services, characterized in that: The following steps are involved: Step S1: determine the target area and research period, collect relevant data, and establish a database of factors affecting basic ecosystem services; The target area is selected from administrative units, river basins or urban agglomerations; Step S2: Select key evaluation indicators for the target area and quantify them, and use the quantified results as the dependent variable data set; Step S3: construct a multiple linear regression model to capture the nonlinear relationship between influencing factors, and introduce nonlinear terms to modify the multiple linear regression model; Step S4: Perform a preliminary screening of the influencing factors in the revised model based on the variance inflation factor (VIF) to eliminate influencing factors with multicollinearity; Step S5: further screen the impact factors using the optimal subset feature screening to establish a corresponding impact factor index system; Step S6: Based on the influencing factor indicator system established in step S5, an improved multi-scale geographically weighted regression (MGWR) model is used to perform local regression analysis. The dominant factors of ecosystem services are identified based on the obtained regression coefficient values, and their spatial distribution is visualized in visualization software to draw a spatial distribution map of the dominant factors of ecosystem services.

2. The method for identifying the dominant factors of ecosystem services according to claim 1, characterized in that: The ecosystem service influencing factor database in step S1 includes meteorological data, remote sensing data, socioeconomic data, land use data and soil property data; the meteorological data includes annual average precipitation, annual average temperature and annual evaporation; the remote sensing data includes DEM digital elevation, normalized difference vegetation index NDVI and vegetation net primary productivity NPP; the socioeconomic data includes gross domestic product, population density, tertiary industry output value and night lights; the land use data includes cultivated land area, forest land area, grassland area, water area and construction land area; the soil property data includes soil depth and the content of sand, silt, clay and organic carbon.

3. The method for identifying the dominant factors of ecosystem services according to claim 1, characterized in that: In step S2, key evaluation indicators for the target area are selected and quantified. The key evaluation indicators include the supply service, regulation service, cultural service and support service of ecosystem services. The physical quantity quantitative indicators corresponding to each type of service include: grain output, water conservation, soil conservation, carbon sequestration, water purification, cultural service suitability index and habitat quality index; based on the research target area, data availability and ecological characteristics of the target area determined in step S1, representative ecosystem service indicators are selected for quantification, and the quantification results are used as the dependent variable data set for local regression analysis in subsequent steps.

4. The method for identifying the dominant factors of ecosystem services according to claim 1, characterized in that: The process of introducing nonlinear terms to modify the multiple linear regression model in step S3 is as follows: Step S31: construct a multiple linear regression model and capture the nonlinear relationship between influencing factors through scatter plots or residual plots; Step S32: Based on the linear and nonlinear relationships among the influencing factors, a nonlinear term is introduced as an influencing factor to correct the multiple linear regression model. The mathematical expression of the correction process is as follows: ； In the above formula, is the dependent variable; is the intercept; It is i The coefficient of the independent variable; It is i independent variables; i =1,2,3,…,n, j =1,2,3,…,n, where n is a positive integer; is the nonlinear term introduced, is the coefficient of the nonlinear term, is an independent random error term.

5. The method for identifying the dominant factors of ecosystem services according to claim 1, characterized in that: In step S4, the VIF value between each influencing factor is calculated in R language. If the VIF value is greater than 10, it indicates that multicollinearity exists between the corresponding influencing factors, and the corresponding influencing factors are eliminated.

6. The method for identifying the dominant factors of ecosystem services according to claim 1, characterized in that: In step S5, the influencing factors are further screened using the optimal subset feature screening, and the Mallows Cp criterion is selected as the analysis indicator. The smaller the Cp value, the higher the accuracy of the model. The indicator with the smallest Cp value is selected to construct the influencing factor indicator system.

7. The method for identifying the leading factors of ecosystem services according to claim 1, characterized in that: In step S6, based on the impact factor index system established in step S5, the improved multi-scale geographically weighted regression MGWR model is used to perform local regression analysis, and the obtained regression coefficients are used to calculate the local regression coefficients. β Value and θ The process of identifying the dominant factors of ecosystem services is as follows: Step S61: using the physical quantity of ecosystem services quantified in step S2 as the dependent variable, inputting the improved multi-scale geographically weighted regression (MGWR) model to perform local regression analysis; Step S62: Based on the regression analysis results, use the if function to screen the significance level p value of each independent variable. If p < 0.05 is satisfied, the regression coefficient corresponding to the independent variable is returned. If not satisfied, a value smaller than the minimum regression coefficient is returned. Step S63: Sort the values ​​returned by each independent variable to obtain the importance ranking of the influencing factors, thereby identifying the dominant factors of ecological services.

8. The method for identifying the dominant factors of ecosystem services according to claim 7, characterized in that: The improved multi-scale geographically weighted regression (MGWR) model captures the nonlinear relationship between influencing factors and introduces nonlinear terms for comprehensive analysis. It can find the optimal bandwidth for each linear and nonlinear explanatory variable, thereby effectively capturing the spatial heterogeneity of each influencing factor at different geographic spatial scales and identifying the dominant influencing factors of ecosystem services. The expression of the improved multi-scale geographically weighted regression (MGWR) model is as follows: ； ； In the above formula, Y i is the spatial position The dependent variable at is the spatial position The independent variable at is the spatial position The nonlinear term at ; is the intercept term of the regression relationship; the coefficient and is the spatial position A continuous function of 、 in Indicates the k The optimal bandwidth for each independent variable; is an independent random error term.

Images