Geoprobing-based vegetation environment response relationship analysis and simulation system

Abstract

This invention discloses a vegetation-environment response relationship analysis and simulation system based on geographic surveying. The system is mainly composed of a data acquisition module, an adaptive partitioning module, a factor analysis module, a factor interaction evaluation module, and a response model construction module, organically integrating these components. It enables the scientific analysis and accurate simulation of complex response relationships between vegetation and environmental factors. Simultaneously, the system adaptively partitions environmental factors based on their spatial autocorrelation characteristics, quantitatively assesses factor explanatory power through the ratio of intra-layer variance to overall variance, innovatively identifies and quantifies nonlinear interactions between environmental factors, and constructs a weighted response model that combines predictive accuracy and ecological explanatory power. This invention can accurately identify the dominant environmental factors in vegetation distribution across different ecological regions, reveal complex synergistic relationships between factors, and provide scientific decision support for vegetation dynamics prediction, ecosystem protection and restoration, and sustainable land resource management under the background of climate change.

Description

Technical Field

[0001] This invention relates to the field of geographic information science and technology, and more specifically, to a vegetation environment response relationship analysis and simulation system based on geographic exploration. Background Technology

[0002] Existing vegetation-environment relationship analysis and simulation techniques face numerous unresolved issues. Spatial zoning of study areas often relies on researchers' subjective experience and judgment, lacking quantitative standards and theoretical support. This makes it difficult for analytical unit settings to accurately reflect the natural transitions and heterogeneity of environmental gradients, particularly evident in complex terrains or ecological transition zones. Regarding the quantification of the explanatory power of environmental factors on vegetation distribution patterns, traditional methods largely rely on simple linear relationship assumptions, failing to address the prevalent nonlinear responses in real ecosystems. This results in significant biases in identifying dominant factors in complex regions such as alpine-plain transition zones or monsoon margins. More critically, the complex interaction mechanisms among environmental factors in ecosystems are widely overlooked. Key ecological processes such as the synergistic effects of temperature and precipitation, and the mutual influence of topography and soil factors, cannot be effectively captured and quantified, leading to a severe lack of understanding of vegetation dynamics in arid and semi-arid regions or the stability of forest-steppe transition ecosystems. In the model building stage, the determination of environmental factor weights often relies on experience or simple statistical relationships, lacking scientific methods to integrate factor interaction effects. This results in models that, while able to fit observational data under local conditions, have extremely limited predictive ability in situations such as climate change or human-induced disturbances. Furthermore, while currently widely used machine learning models exhibit high statistical fitting advantages, their "black box" nature leads to insufficient explanation of ecological mechanisms, making it difficult to provide clear theoretical guidance and decision support for ecological protection, vegetation restoration, or climate change adaptation management.

[0003] In view of this, the present invention proposes a vegetation environment response relationship analysis and simulation system based on geographic survey to solve the above problems. Summary of the Invention

[0004] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution:

[0005] A vegetation-environment response relationship analysis and simulation system based on geographic surveying, comprising:

[0006] The data acquisition module is used to acquire raster data of vegetation cover and multi-source environmental factors in the study area.

[0007] The adaptive partitioning module is used to adaptively partition the study area based on the spatial autocorrelation degree and local variation intensity of environmental factors, and obtain multiple spatial analysis units.

[0008] The factor analysis module is used to obtain the factor explanatory power index of each environmental factor based on the ratio of the intra-level variance of vegetation cover in each spatial analysis unit to the total variance of each environmental factor classification layer; and to screen the dominant environmental factors based on the factor explanatory power index.

[0009] The factor interaction evaluation module is used to obtain the response contribution weight of each dominant environmental factor based on the interaction type of any two dominant environmental factors and the factor explanatory power index of each dominant environmental factor.

[0010] The response model building module is used to construct a weighted response model based on the response contribution weights for vegetation distribution simulation and prediction.

[0011] Compared with existing technologies, the adaptive zoning method established in this invention significantly enhances the ecological significance of spatial analysis units, enabling research results to better reflect the inherent laws of natural ecosystems. In practical applications across different ecological regions, the environmental factor analysis capabilities of this invention, particularly in complex environments such as climate transition zones, alpine ecosystems, and areas affected by human activities, have successfully revealed environmental factor interaction mechanisms that are difficult to capture using traditional methods. The weighted response model constructed in this invention overcomes the bottleneck of balancing prediction accuracy and ecological explanatory power, providing a clear ecological interpretation path while maintaining high-precision simulation capabilities, enabling research results to be directly transformed into practical management measures. The multi-scenario simulation capabilities of this invention provide a scientific tool for predicting vegetation dynamics in ecologically vulnerable areas, and have a positive impact on biodiversity conservation, ecosystem service assessment, and sustainable land resource management. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of the vegetation environment response relationship analysis and simulation system based on geographic exploration according to the present invention. Detailed Implementation

[0013] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0014] This application provides a system for analyzing and simulating vegetation-environment response relationships based on geographic surveys. The system's execution entities include, but are not limited to, those mounted on the system: ecological monitoring platforms, geographic information systems, environmental response analysis platforms, and vegetation distribution prediction systems, which can be considered general computing nodes in this application. The system includes, but is not limited to, at least one of the following: a distributed data processing engine, an adaptive spatial analyzer, and a multi-factor interactive evaluation model.

[0015] Please see Figure 1 This embodiment provides a vegetation environment response relationship analysis and simulation system based on geographic surveying, which includes:

[0016] The data acquisition module is used to acquire raster data of vegetation cover and multi-source environmental factors in the study area. The multi-source environmental factor raster data includes key information such as climate factors, topographic factors, soil factors, and anthropogenic disturbance factors, which are acquired in real time through a multi-source data acquisition interface. Climate factors include climatic elements such as precipitation, temperature, and solar radiation; topographic factors include surface features such as elevation, slope, and aspect; soil factors include soil properties such as soil type, organic matter content, and texture composition; and anthropogenic disturbance factors include indicators of human influence such as land use type, population density, and road density. This data provides comprehensive environmental background information for subsequent analysis, ensuring the completeness and accuracy of the model construction.

[0017] The adaptive zoning module is used to adaptively zonate the study area based on the spatial autocorrelation and local variability of environmental factors, obtaining multiple spatial analysis units. The spatial analysis unit is the basic unit for vegetation-environment relationship analysis. By considering the spatial autocorrelation characteristics and variability of environmental factors, the study area is divided into regional units with relative internal homogeneity. The adaptive zoning process calculates local spatial autocorrelation coefficients and environmental variability coefficients, identifies the characteristics of environmental gradient changes, determines the optimal zoning threshold, and achieves scientific zoning of the study area, providing a spatial basis for subsequent vegetation-environment response relationship analysis.

[0018] The factor analysis module is used to obtain the factor explanatory power index for each environmental factor based on the ratio of the intra-level variance of vegetation cover within each environmental factor classification layer to the overall variance within each spatial analysis unit. Dominant environmental factors are then selected based on this factor explanatory power index. The factor explanatory power index is a core indicator for quantifying the ability of environmental factors to explain the spatial distribution of vegetation. By calculating the variance decomposition of vegetation cover within each environmental factor classification layer, key environmental factors that can significantly reduce the spatial heterogeneity of vegetation are identified. The calculation process is based on the geospatial detector theory, quantitatively assessing the explanatory power of each environmental factor by comparing the relationship between conditional variance and overall variance. Dominant environmental factors are then selected by setting thresholds, providing a factor basis for subsequent interactive evaluations.

[0019] The factor interaction assessment module is used to obtain the response contribution weight of each dominant environmental factor based on the interaction type between any two dominant environmental factors and the factor explanatory power index of each dominant environmental factor. The response contribution weight reflects the relative importance of the dominant environmental factors in the vegetation response process, and is comprehensively assessed by analyzing the interaction type and intensity between factors. The interaction assessment process first determines the interaction type between any two dominant environmental factors (nonlinear enhancement, two-factor enhancement, or nonlinear weakening), then calculates the interaction enhancement ratio of each factor based on the interaction results, and finally converts it into a normalized response contribution weight, providing a weighting basis for constructing a weighted response model.

[0020] The response model construction module is used to build a weighted response model based on response contribution weights for vegetation distribution simulation and prediction. The weighted response model is the core component for vegetation distribution simulation. By integrating dominant environmental factors and their response contribution weights, a quantitative response relationship between the environment and vegetation is established. The model construction process first calculates the weighted contribution of each dominant environmental factor, then obtains the predicted vegetation cover value through response function mapping, and optimizes the model parameters through residual analysis. Finally, a weighted response model that can accurately simulate the spatial distribution pattern of vegetation is formed, providing scientific decision support for ecological environment assessment and planning.

[0021] The modules are connected via wired and / or wireless means to enable data transmission between them.

[0022] In this embodiment of the invention, the detailed implementation steps for obtaining the factor explanatory power index of each environmental factor include:

[0023] Based on the numerical distribution characteristics of environmental factors, the corresponding environmental factor raster data is divided into several strata. Stratification is a fundamental step in factor explanatory power analysis, discretizing continuous environmental factor data into meaningful classification intervals. The process first analyzes the statistical distribution characteristics of environmental factors, including frequency distribution, kurtosis, and skewness, to determine a suitable classification method. Then, an appropriate classification algorithm is selected based on data characteristics, such as the equal interval method, natural discontinuity method, equal frequency method, or cluster analysis. Finally, several strata are generated, each representing a specific numerical interval of the environmental factor. The number of strata is usually dynamically determined based on data complexity and sample size, ensuring sufficient sample size within each stratum and significant environmental gradient differences between strata. This adaptive classification method based on data characteristics ensures the scientific validity and effectiveness of subsequent intra-stratum variance analysis.

[0024] The vegetation cover values ​​corresponding to all grid cells within each type layer are statistically analyzed, and the intra-layer variance of each type layer is calculated. Intra-layer variance calculation is a crucial step in quantifying the explanatory power of each environmental type layer for vegetation spatial differentiation. The calculation process first involves spatially overlaying the vegetation cover grid cells of the study area with the environmental factor type layers, extracting the vegetation cover values ​​corresponding to all grid cell locations within each type layer; then, the variance of vegetation cover within each type layer is calculated to quantify the degree of spatial heterogeneity of vegetation within the type layer. The formula for calculating intra-layer variance is:

[0025] ;in, For type layer Intra-layer variance, For type layer The number of grid cells within, For type layer The Middle The vegetation cover value of each grid cell. For type layer Average vegetation coverage.

[0026] The smaller the intra-layer variance, the stronger the explanatory power of the environmental factor for vegetation distribution within the corresponding numerical range, and the more homogeneous the spatial distribution of vegetation tends to be, providing basic data for subsequent calculation of the factor explanatory power index.

[0027] Calculate the overall variance of vegetation cover within the spatial analysis unit. Calculating the overall variance is a benchmark step in quantifying the overall level of spatial heterogeneity of vegetation within the study area. The calculation process is directly based on the vegetation cover values ​​of all grids within the spatial analysis unit, applying the variance calculation formula to derive the overall variance. The overall variance reflects the overall degree of variability in vegetation cover within the study area. As a reference benchmark for assessing the explanatory power of environmental factors, a larger value indicates stronger spatial heterogeneity of vegetation within the area, requiring stronger explanatory power from environmental factors. The calculation of overall variance takes into account the characteristics of the spatial analysis unit, ensuring the regional relevance and scientific rigor of the factor explanatory power analysis.

[0028] The intra-layer variances of all types of layers are weighted and summed according to the number of rasters. The ratio of the difference between the weighted sum and the overall variance to the overall variance is calculated to obtain the factor explanatory power index for the corresponding environmental factor. Calculating the factor explanatory power index is a core step in quantitatively assessing the explanatory power of environmental factors on vegetation distribution, based on the variance decomposition principle of a geographic detector. The calculation process first involves weighting and summing the intra-layer variances of each type of layer according to the proportion of rasters within that type of layer to the total number of rasters, obtaining the conditional variance. Then, the difference between the overall variance and the conditional variance is calculated, representing the variance explained by the environmental factor. Finally, the ratio of this difference to the overall variance is used as the factor explanatory power index. The formula for calculating the factor explanatory power index is:

[0029] ;in, As the explanatory power index of factors, For type layer The number of grid cells within, This represents the total number of grid cells within the spatial analysis unit. For type layer Intra-layer variance, This represents the overall variance of vegetation cover within the spatial analysis unit.

[0030] The factor explanatory power index ranges from [0,1], with values ​​closer to 1 indicating a stronger explanatory power of environmental factors on vegetation distribution. This index directly reflects the degree to which environmental factors control the spatial distribution of vegetation, providing a scientific basis for the selection of dominant factors and subsequent analysis.

[0031] In this embodiment of the invention, the detailed implementation steps of the method for obtaining the response contribution weight include:

[0032] The interaction factor explanatory power index is calculated for any two dominant environmental factors. The numerical relationship between this index and the factor explanatory power indices of the corresponding two dominant environmental factors is determined to identify the type of interaction between them. Calculating the interaction factor explanatory power index is a fundamental step in analyzing the synergistic effects between environmental factors, assessing the joint explanatory power of factor combinations on vegetation distribution. The calculation process first cross-combines the type layers of the two dominant environmental factors to generate a joint type layer; then, the in-layer variance and conditional variance are calculated based on the joint type layer; finally, the factor explanatory power index calculation formula is applied to derive the interaction factor explanatory power index. The type of interaction is determined by comparing the numerical relationship between the interaction factor explanatory power index and the single-factor explanatory power index to identify mutually reinforcing or weakening effects between factors. Interactions are classified into three basic types: nonlinear reinforcing, two-factor reinforcing, and nonlinear weakening, providing a basis for subsequent calculation of response contribution weights.

[0033] Based on the statistical results of the interaction types between each dominant environmental factor and all other dominant environmental factors, the interaction enhancement ratio of each dominant environmental factor is obtained. Calculating the interaction enhancement ratio is a crucial step in evaluating the synergistic effect of environmental factors in a multi-factor system. The calculation process first statistically analyzes the proportion of enhancing effects (non-linear or bi-factor enhancement) in all interaction combinations involving each dominant environmental factor; then, it weights the interactions according to their strength, assigning higher weights to non-linear enhancements; finally, the interaction enhancement ratio of each factor is derived, reflecting the synergistic ability of factors in complex environmental systems. The formula for calculating the interaction enhancement ratio is:

[0034] ;in, As a factor The interaction enhancement ratio, The interaction strength weights are 2 for the non-linear enhancement type and 1 for the two-factor enhancement type, where I is the indicator function. With factors Interaction type The value is 1 if it belongs to the enhanced type, and 0 otherwise. The total number of dominant environmental factors.

[0035] The interaction enhancement ratio directly reflects the synergistic potential of environmental factors in complex ecosystems. A higher ratio indicates that the factor is more able to enhance the overall explanatory power of vegetation distribution through synergistic effects with other factors.

[0036] The response contribution weight of each dominant environmental factor is obtained by normalizing the product of its explanatory power index and the corresponding interaction enhancement ratio. Calculating the response contribution weight is a core step that comprehensively considers both the factor's own explanatory power and its synergistic interaction capabilities. The calculation process first multiplies the factor's explanatory power index by the interaction enhancement ratio to obtain the factor's overall effect value; then, the overall effect values ​​of all dominant environmental factors are normalized to ensure that the sum of all weights is 1; finally, the response contribution weight of each dominant environmental factor is obtained. The formula for calculating the response contribution weight is:

[0037] ;in, As a factor Response contribution weight, As a factor The explanatory power index, As a factor The interaction enhancement ratio is calculated, with the denominator being the sum of the combined effects of all dominant environmental factors.

[0038] The response contribution weights comprehensively reflect the relative importance of environmental factors in the formation of vegetation distribution patterns. They take into account both the explanatory power of the factors themselves and the synergistic effects of the factors in complex environmental systems, providing a scientific basis for constructing weighted response models.

[0039] In this embodiment of the invention, the detailed implementation steps of the adaptive partitioning method include:

[0040] The local spatial autocorrelation coefficient of each grid cell within the study area is calculated, along with the environmental factor variation coefficient of surrounding grid cells within a predefined neighborhood. Local spatial autocorrelation calculation is a fundamental step in assessing the spatial clustering characteristics of environmental factors, identifying spatial structure by quantifying the similarity between a grid cell and its neighboring units. The calculation process employs the local Moran's I index to evaluate the spatial correlation between the environmental factor value at each grid cell location and the values ​​of surrounding grid cells. Simultaneously, the environmental factor variation coefficient is calculated to quantify the intensity of environmental gradient changes within the local area. The variation coefficient is calculated as the ratio of the standard deviation to the mean, reflecting the relative variability of environmental factors within the local area. These two indicators together describe the spatial distribution characteristics of environmental factors, providing a spatial structural basis for subsequent zoning.

[0041] Based on the joint distribution characteristics of local spatial autocorrelation coefficients and coefficients of variation, a spatial clustering threshold is determined. Determining the spatial clustering threshold is a crucial step in scientifically dividing spatial units, identifying the optimal segmentation point by analyzing the distribution characteristics of these two indicators. The process first constructs a two-dimensional scatter plot of local spatial autocorrelation coefficients and coefficients of variation to analyze the distribution pattern of the points; then, density clustering algorithms, such as DBSCAN or an improved K-means, are applied to identify the natural clustering structure of the points; finally, the optimal spatial clustering threshold is determined based on the principle of maximizing cluster gaps. The threshold selection process considers the scale characteristics and environmental complexity of the study area, ensuring that the partitioning results can maintain the natural structure of the environmental gradient to the greatest extent, providing reasonable spatial units for vegetation-environment relationship analysis.

[0042] Spatial clustering of the study area is performed based on a spatial clustering threshold, aggregating adjacent rasters with similar spatial autocorrelation characteristics and similar degrees of environmental variability into the same spatial analysis unit. Spatial clustering is an implementation step of adaptive partitioning, organizing similar rasters into continuous spatial units through a region growing algorithm. The clustering process first selects seed rasters that meet the threshold conditions, and then gradually incorporates similar rasters into the same unit through neighborhood expansion until all rasters are assigned to specific spatial analysis units. The clustering algorithm considers spatial continuity constraints, ensuring that the formed spatial analysis units are geographically continuous and relatively homogeneous in environmental characteristics. The final generated spatial analysis units maintain the spatial autocorrelation structure of environmental factors while controlling the degree of internal environmental variability, providing a scientific spatial basis for subsequent vegetation-environment relationship analysis.

[0043] In this embodiment of the invention, the detailed implementation steps of the method for screening dominant environmental factors include:

[0044] Calculate the mean and standard deviation of the factor explanatory power indices for all environmental factors. Statistical calculation is a fundamental step in the initial screening of dominant factors, determining screening criteria by analyzing the overall distribution characteristics of factor explanatory power. The calculation process first summarizes the explanatory power indices of all environmental factors, then applies statistical formulas to calculate the mean and standard deviation, quantifying the central tendency and dispersion of the explanatory power distribution. The mean reflects the overall explanatory power level of the environmental factors, while the standard deviation reflects the degree of difference in explanatory power among different factors. These two statistics together constitute the reference benchmark for screening dominant factors, providing a statistical basis for setting scientific screening thresholds.

[0045] Environmental factors whose explanatory power index is greater than the sum of the standard deviations of the mean and a preset multiple are identified as initial dominant factors. Determining the initial dominant factors is the first step in the screening process. This involves identifying environmental factors with significantly higher explanatory power than the average level by setting statistical thresholds. The threshold is set using the mean plus the standard deviation, where the multiple of the standard deviation is typically set between 0.5 and 1.5, dynamically adjusted according to the environmental complexity of the study area and the analytical objectives. This statistically significant screening method effectively identifies environmental factors with strong explanatory power, avoids including weakly correlated factors in the model, and improves the model's conciseness and explanatory power. The initial selection results typically include 3-8 dominant environmental factors, providing a preliminary set for subsequent collinearity diagnosis.

[0046] Collinearity diagnosis is performed on all initially selected dominant factors. Factors with variance inflation exceeding a preset inflation threshold are removed, and the remaining factors are designated as dominant environmental factors. Collinearity diagnosis is a crucial step in ensuring model stability, avoiding information redundancy by assessing the correlation structure between factors. The diagnosis process first constructs a correlation matrix among the initially selected dominant factors and calculates the variance inflation factor (VIF) for each factor. Then, based on a preset inflation threshold (usually set to 5-10), factors exhibiting severe collinearity are identified. Finally, a stepwise elimination method is used to remove factors with VIF values ​​exceeding the threshold until the VIF values ​​of all remaining factors are below the threshold. During the elimination process, factors with higher explanatory power are prioritized for retention, ensuring that the final selected dominant environmental factors possess strong explanatory power while being independent and free from severe redundancy, laying a factor foundation for building a robust response model.

[0047] In this embodiment of the invention, the detailed implementation steps of the interaction type determination method include:

[0048] Calculate the explanatory power index of interaction factors after joint classification of any two dominant environmental factors. Calculating the explanatory power index of interaction factors is a fundamental step in determining the type of factor interaction and assessing the joint explanatory power of factor combinations for vegetation distribution. The calculation process first cross-combines the type layers of the two dominant environmental factors to generate a joint type matrix; then, it calculates the intra-layer variance of vegetation cover based on the joint type matrix; finally, it applies the formula for calculating the explanatory power index of factors to obtain the explanatory power index of interaction factors. The calculation considers the joint distribution characteristics of the two factors, comprehensively reflects the integrated explanatory power of factor combinations for vegetation spatial patterns, and provides a quantitative basis for subsequent determination of interaction types.

[0049] If the explanatory power index of the interaction factor is greater than the sum of the explanatory power indices of the two dominant environmental factors, it is classified as a nonlinear enhancement type. Determining a nonlinear enhancement type is a crucial step in identifying strong synergistic effects, indicating that the factor combination produces a nonlinear enhancement effect beyond simple addition. The determination process compares the magnitude of the explanatory power indices of the interaction factor with the sum of the explanatory power indices of the individual factors. When the interaction index is significantly greater than the sum of the two factor indices, it is classified as a nonlinear enhancement type. This type of interaction indicates a strong synergistic effect between the two environmental factors, jointly producing a superadditive regulatory effect on vegetation distribution. Special attention should be paid to the significant impact of such interaction combinations on vegetation in model construction.

[0050] If the explanatory power index of the interaction factor is greater than the larger of the explanatory power indices of the two factors but less than their sum, it is classified as a two-factor enhancement type. The two-factor enhancement type determination is a step in identifying moderate synergy, indicating that the factor combination produces a synergistic effect intermediate between single-factor and nonlinear enhancement. The determination process compares the relationship between the maximum values ​​of the interaction factor explanatory power index and the single-factor explanatory power index, and their sum. When the interaction index is greater than the maximum single-factor index but less than the sum of the two factors, it is classified as a two-factor enhancement type. This interaction type indicates a positive synergistic effect between the two environmental factors, jointly enhancing their explanatory power for vegetation distribution, but the enhancement effect is not as significant as in the nonlinear enhancement type, reflecting a moderate degree of synergy between environmental factors.

[0051] If the explanatory power index of the interaction factor is less than the smaller of the explanatory power indices of the two factors, it is classified as a nonlinearly weakening type. The nonlinearly weakening type determination is a step in identifying antagonistic effects, indicating that the combination of factors produces a negative interaction effect of mutual inhibition. The determination process compares the minimum explanatory power index of the interaction factor with the minimum explanatory power index of the individual factors. When the interaction index is less than the minimum of the individual factors, it is classified as a nonlinearly weakening type. This type of interaction indicates that there is mutual inhibition or interference between the two environmental factors, and the explanatory power of the combination is lower than that of the individual factors, reflecting the antagonistic relationship between the environmental factors. Such negative interaction effects of factors should be considered in model construction.

[0052] In this embodiment of the invention, the detailed implementation steps of the method for constructing the weighted response model include:

[0053] The weighted contribution of each dominant environmental factor is calculated based on its response contribution weight and the normalized value of the corresponding environmental factor. Calculating the weighted contribution is a fundamental step in constructing the response model, integrating the numerical value of the environmental factor with its importance into a comprehensive impact value. The calculation process first normalizes the original values ​​of each dominant environmental factor to eliminate dimensional differences and map them to the [0,1] interval; then, the normalized value is multiplied by the response contribution weight to obtain the weighted contribution of each dominant environmental factor. Normalization typically employs min-maximum standardization or Z-fractional standardization, selecting an appropriate method based on the factor distribution characteristics. The weighted contribution directly reflects the actual impact of environmental factors on vegetation, comprehensively considering both the numerical intensity and relative importance of the factors, providing a basic input for vegetation cover prediction.

[0054] The weighted contributions of all dominant environmental factors are summed, and the predicted vegetation cover value is obtained by mapping the summation through a preset response function. Calculating the predicted value is the core step in generating vegetation distribution simulation results. The response function transforms the comprehensive impact of environmental factors into vegetation cover. The calculation process first sums the weighted contributions of each dominant environmental factor to obtain the comprehensive environmental impact value; then, the comprehensive impact value is mapped to the predicted vegetation cover value through a preset response function. The selection of the response function considers the actual response characteristics of vegetation to environmental factors; commonly used function forms include logistic functions, exponential functions, or polynomial functions. The vegetation cover prediction formula is:

[0055] ;in, This is the predicted value for vegetation cover. For the response function, As a factor Response contribution weight, As a factor The normalized value.

[0056] The predicted values ​​intuitively reflect the theoretical distribution level of vegetation under given environmental conditions, providing a benchmark for subsequent model optimization and validation.

[0057] Based on the residual distribution characteristics between measured and predicted vegetation cover, the parameters of the response function are iteratively optimized to obtain the final weighted response model. Model optimization is a key step in improving prediction accuracy, and the model parameters are adjusted by analyzing residual characteristics. The optimization process first calculates the residuals between measured and predicted values ​​and analyzes the distribution characteristics of the residuals, such as skewness and autocorrelation. Then, based on the residual characteristics, the parameters of the response function are iteratively adjusted using optimization algorithms such as gradient descent or simulated annealing to minimize the sum of squared residuals. Finally, the optimized weighted response model is obtained. The optimization process pays special attention to the spatial distribution pattern of the residuals. Systematic errors are identified through spatial autocorrelation analysis, and the model structure is adjusted accordingly to ensure that the final model can accurately capture the vegetation-environment response relationship, providing a high-precision prediction tool for vegetation distribution simulation.

[0058] In this embodiment of the invention, the detailed implementation steps of the accuracy verification method for the weighted response model include:

[0059] Spatial analysis units within the study area are randomly divided into training and validation sets. Dataset partitioning is a fundamental step in model validation, and stratified spatial sampling ensures the representativeness of both training and validation data. The partitioning process employs a stratified random sampling strategy, dividing the spatial analysis units into layers based on environmental gradient characteristics, and then randomly sampling within each layer to ensure that the training and validation datasets have similar distributions in terms of environmental features. Typically, 70%-80% of the spatial analysis units are allocated to the training set for model building and parameter estimation, while 20%-30% are allocated to the validation set for model accuracy evaluation. This stratified spatial random partitioning method ensures both the sufficiency of the training process and the reliability and representativeness of the validation results.

[0060] A weighted response model was constructed using vegetation cover data and environmental factor data from the training unit set. Model training is the core step in accuracy validation, where model parameters are fitted using training data. The training process first calculates the explanatory power index and interaction type of each environmental factor based on the environmental factor data from the training unit set; then, it determines the dominant environmental factor and its response contribution weight; finally, it constructs a weighted response model based on the weights and normalized environmental factor values, and optimizes the response function parameters through residual analysis. The training process strictly follows the aforementioned model construction method to ensure that the model can capture the vegetation-environment response relationship in the training data to the greatest extent, providing a benchmark model for subsequent validation.

[0061] Environmental factor data from the validation set are substituted into a weighted response model to obtain simulated vegetation cover values. Simulation prediction is a crucial step in evaluating the model's generalization ability, testing its performance with new data. The prediction process first normalizes the environmental factor data from the validation set using a standardization method consistent with the training data. Then, based on the response contribution weights obtained during training, the weighted contribution of each dominant environmental factor is calculated. Finally, the optimized response function generates simulated vegetation cover values ​​for each spatial location within the validation set. This step ensures the objectivity of model validation, testing the model's predictive ability with entirely new data, avoiding spurious accuracy caused by overfitting, and providing reliable prediction results for model evaluation.

[0062] The coefficient of determination (R²) and root mean square error (RMSE) between simulated and measured vegetation cover values ​​within the validation unit set are calculated to evaluate the simulation accuracy of the weighted response model. Accuracy evaluation is the final step in the validation process, quantifying the model's predictive performance using statistical indicators. The evaluation process first pairs simulated and measured vegetation cover values ​​within the validation unit set; then, the R² is calculated to assess the proportion of variance explained by the model; simultaneously, the RMSE is calculated to assess the absolute error level of the prediction; and, if necessary, relative error and systematic bias can also be calculated to comprehensively evaluate model performance. A higher R² and a smaller RMSE indicate higher model prediction accuracy. The accuracy evaluation results provide objective evidence for the model's practical application value and also point the way for further optimization and improvement of the model.

[0063] This invention achieves precise analysis and simulation of the response relationship between vegetation and environmental factors through data acquisition, adaptive zoning, factor analysis, factor interaction assessment, and response model construction. The geographic exploration method of this invention can quantify the explanatory power and interaction of environmental factors on vegetation distribution, construct a highly interpretable weighted response model, and provide a scientific tool for ecological environment assessment and management.

[0064] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0065] It should be noted that all formulas in this manual are calculated by removing dimensions and taking their numerical values. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0066] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A vegetation-environment response relationship analysis and simulation system based on geographic surveying, characterized in that, include: The data acquisition module is used to acquire raster data of vegetation cover and multi-source environmental factors in the study area. The adaptive partitioning module is used to adaptively partition the study area based on the spatial autocorrelation degree and local variation intensity of environmental factors, and obtain multiple spatial analysis units. The factor analysis module is used to obtain the factor explanatory power index of each environmental factor based on the ratio of the intra-level variance of vegetation cover in each spatial analysis unit to the total variance of each environmental factor classification layer. The dominant environmental factors were screened based on the explanatory power index of the factors. The factor interaction evaluation module is used to obtain the response contribution weight of each dominant environmental factor based on the interaction type of any two dominant environmental factors and the factor explanatory power index of each dominant environmental factor. The response model construction module is used to construct a weighted response model based on the response contribution weights to simulate and predict vegetation distribution.

2. The system according to claim 1, characterized in that, The process of obtaining the factor explanatory power index for each environmental factor includes: Based on the numerical distribution characteristics of environmental factors, the corresponding environmental factor raster data are divided into several types of layers; Collect vegetation cover values ​​for all grid cells within each type layer, and calculate the intra-layer variance for each type layer; Calculate the overall variance of vegetation cover within the spatial analysis unit; The intra-layer variances of all types of layers are weighted and summed according to the number of raster cells. The ratio of the difference between the weighted sum and the total variance to the total variance is calculated to obtain the factor explanatory power index of the corresponding environmental factor.

3. The system according to claim 1, characterized in that, The method for obtaining the response contribution weight includes: Calculate the explanatory power index of the interaction factor between any two dominant environmental factors, determine the numerical relationship between the explanatory power index of the interaction factor and the explanatory power index of the corresponding two dominant environmental factors, and determine the interaction type of the corresponding two dominant environmental factors. Based on the statistical results of the interaction types between each dominant environmental factor and all other dominant environmental factors, the interaction enhancement ratio of each dominant environmental factor is obtained; The product of the factor explanatory power index of each dominant environmental factor and the corresponding interaction enhancement ratio is normalized to obtain the response contribution weight of each dominant environmental factor.

4. The system according to claim 1, characterized in that, The adaptive partitioning method includes: Calculate the local spatial autocorrelation coefficient of each grid within the study area and the environmental factor variation coefficient of the grids within the surrounding preset neighborhood; Based on the joint distribution characteristics of the local spatial autocorrelation coefficient and the coefficient of variation, the spatial clustering threshold is determined; Based on the spatial clustering threshold, spatial clustering is performed on the study area, and adjacent grids with similar spatial autocorrelation characteristics and similar environmental variability are aggregated into the same spatial analysis unit.

5. The system according to claim 2, characterized in that, The method for screening the dominant environmental factors includes: Calculate the mean and standard deviation of the factor explanatory power index for all environmental factors; Environmental factors whose explanatory power index is greater than the sum of the standard deviations of the mean and a preset multiple are denoted as the initial dominant factors; Collinearity diagnosis is performed on all initially selected dominant factors. Initially selected dominant factors whose variance inflation factor exceeds the preset inflation threshold are removed, and the remaining initially selected dominant factors are used as dominant environmental factors.

6. The system according to claim 3, characterized in that, The method for determining the type of interaction includes: Calculate the explanatory power index of the interaction factors after joint classification of any two dominant environmental factors; If the explanatory power index of the interaction factor is greater than the sum of the explanatory power indices of the corresponding two dominant environmental factors, it is determined to be a nonlinear enhanced type. If the explanatory power index of the interaction factor is greater than the larger of the explanatory power indices of the two factors but less than the sum of the two, it is determined to be a two-factor enhanced type. If the explanatory power index of the interaction factor is less than the smaller of the explanatory power indices of the two factors, it is determined to be a nonlinear weakening type.

7. The system according to claim 1, characterized in that, The method for constructing the weighted response model includes: The weighted contribution of each dominant environmental factor is calculated based on the response contribution weight of each dominant environmental factor and the normalized value of the corresponding environmental factor. The weighted contributions of all dominant environmental factors are summed, and the predicted vegetation cover value is obtained by mapping through a preset response function. Based on the residual distribution characteristics between the measured vegetation coverage and the predicted vegetation coverage, the parameters of the response function are iteratively optimized to obtain the final weighted response model.

8. The system according to claim 7, characterized in that, The accuracy verification method for the weighted response model includes: The spatial analysis units within the study area are randomly divided into a training unit set and a validation unit set; A weighted response model was constructed using vegetation cover data and environmental factor data within the training unit set; Substitute the environmental factor data from the validation unit set into the weighted response model to obtain simulated vegetation cover values; The coefficient of determination and root mean square error between the simulated and measured vegetation cover values ​​within the validation unit set are calculated to evaluate the simulation accuracy of the weighted response model.

Images