Ecological risk evaluation and prediction evaluation method based on multi-source evaluation framework

Abstract

The invention provides an ecological risk evaluation and prediction evaluation method based on a multi-source evaluation framework, and relates to the technical field of ecological risk evaluation, and the method specifically comprises the steps: dividing a target region into grid units, and collecting landscape pattern and environmental social economic data; analyzing landscape data through a sliding window method, and calculating a landscape ecological risk value of each unit; constructing a regional ecological risk assessment index system based on social economic data, determining an index weight by adopting Bayesian BWM and CRITIC methods, and performing weighted calculation to obtain a regional ecological risk value; integrating the two types of risk values to form a spatial distribution diagram, and superposing to obtain a unit comprehensive ecological risk value; and performing risk grading by adopting a natural breakpoint method, generating a grading graph, predicting risk early warning time of each unit based on a historical trend, and realizing ecological risk dynamic evaluation and early warning.

Description

Technical Field

[0001] This invention relates to the field of ecological risk assessment technology, specifically to an ecological risk assessment and prediction method based on a multi-source assessment framework. Background Technology

[0002] Ecological risk assessment and early warning are important foundational tasks for regional environmental management and ecological protection. In recent years, with the rapid development of remote sensing technology, geographic information systems and data analysis methods, multi-source data fusion and spatial analysis have become the main trends in ecological risk assessment. Traditional ecological risk assessment methods mostly rely on single data sources or static indicators, lacking a refined characterization of dynamic changes in landscape patterns, spatial heterogeneity and the coupled effects of multiple factors, making it difficult to effectively support regional and forward-looking risk early warning and decision-making.

[0003] In the existing technology, a method for determining regional ecological risk early warning with publication number CN101882274A, although it proposes a comprehensive evaluation system based on regional environmental status, chemical risk pressure, and ecological risk management, still has the following limitations: On the one hand, this method mainly relies on static statistical indicators and expert weighting, lacking dynamic capture and adaptive analysis of landscape pattern spatial characteristics; on the other hand, its early warning process does not integrate landscape ecological response mechanisms at multiple spatiotemporal scales, especially failing to introduce a sliding window algorithm for local spatial correlation analysis, resulting in insufficiently refined identification of risk spatial differentiation characteristics; furthermore, this method does not fully integrate modern data mining techniques, lacking sufficient modeling and prediction capabilities for complex nonlinear risk evolution processes, making it difficult to achieve high-precision, time-series dynamic assessment of ecological risk impact time.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide an ecological risk assessment and prediction method based on a multi-source assessment framework to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The ecological risk assessment and prediction method based on a multi-source assessment framework includes the following steps: Step 1: Divide the target area into multiple grid evaluation units, collect landscape pattern data and regional environmental socio-economic data for each grid evaluation unit during the evaluation period, process the landscape pattern data using the sliding window method, obtain the interference, vulnerability and loss of the landscape pattern data during the evaluation period, and then determine the landscape ecological risk value of each grid evaluation unit. Step 2: Based on regional environmental and socio-economic data, determine the ecological pressure and ecological vulnerability of each grid assessment unit, and construct a regional ecological risk assessment index system for the grid assessment unit; Step 3: Convert the regional ecological risk assessment index system into a judgment matrix, use the Bayesian BWM method and the CRITIC method to determine the subjective and objective weights of each assessment index, and obtain the comprehensive weight of the assessment index through the distance function combination weighting method. Step 4: Use the comprehensive weight of the evaluation indicators to perform weighted calculation on the normalized values ​​of each evaluation indicator to obtain the regional ecological risk value of the grid evaluation unit. Integrate the landscape ecological risk values ​​and regional ecological risk values ​​of all grid evaluation units to construct the spatial distribution of landscape ecological risk values ​​and the spatial distribution of regional ecological risk values. Step 5: Overlay the spatial distribution of landscape ecological risk values ​​with the spatial distribution of regional ecological risk values ​​to obtain the comprehensive ecological risk value of the grid evaluation unit. Use the natural breakpoint method to cluster and classify the comprehensive ecological risk values ​​of all grid evaluation units, generate a comprehensive ecological risk classification map, and determine the risk warning time to achieve ecological risk assessment and prediction.

[0007] Furthermore, the landscape pattern data specifically includes the area, patch density, number of patches, and average patch area of ​​each landscape type, and the regional environmental socio-economic data specifically includes population density, land use intensity, and industrial emission intensity. Among them, the land use intensity Determined based on the following formula: In the formula, For the first grid evaluation unit Area of ​​land use type For the first The intensity weights corresponding to different land use types are determined based on the degree of land development and the intensity of human activities. The total number of land use types, An index for land use types; The total area of ​​the grid evaluation unit; The disturbance degree, vulnerability degree, and loss degree are quantified by the landscape disturbance degree index, landscape vulnerability index, and landscape loss degree index, respectively.

[0008] Furthermore, the specific processing procedure for the landscape pattern data using the sliding window algorithm is as follows: A window with a fixed size is constructed centered on each raster evaluation unit. A square moving window, in which It is an odd number greater than 1; Using each grid evaluation unit as the center of the window, the entire target area is covered by sliding row by row and column by column. For each window position, the landscape pattern data of all grid evaluation units within the window is extracted, and the landscape disturbance index, landscape vulnerability index, and landscape loss index of each grid evaluation unit within the current window range are calculated based on the landscape pattern data. The formulas used are as follows: in, For the current window range, the first Landscape disturbance index of each grid evaluation unit; For the current window range, the first The number of patches within each grid evaluation unit; For the current window range, the first The total area of ​​each grid evaluation unit This is the index of the raster evaluation cell within the current window range; For the current window range, the first The anthropogenic disturbance intensity coefficient for each grid evaluation unit is determined based on the population density and industrial emission intensity within the grid evaluation unit. in, For the current window range, the first Landscape vulnerability index for each grid assessment unit; Within the current window range, the first Ecologically sensitive landscape types in the first Area in each grid evaluation unit; The total number of landscape types, For the first The ecological sensitivity weights for different landscape types were determined through expert evaluation. An index for landscape types; In the formula, For the current window range, the first Landscape loss index of each grid evaluation unit; For the first Ecosystem service value per unit area of ​​landscape type For the first in the target area Historical loss rate of landscape types. Furthermore, the weighted interference index of the evaluation unit at the center of the window is calculated according to the following formula. Weighted Vulnerability Index and weighted loss index : in, This represents the total number of cells within the current window range. Indicates the first [number] within the current window range The influence weights of each grid evaluation unit on the window center grid evaluation unit are determined based on a Gaussian function: For the current window range, the first The Euclidean distance from the nth grid cell to the center of the window, when the nth grid cell is... When a grid cell is the center of the window, its Euclidean distance to itself is zero. This means that the central unit has the greatest influence on its own weight, reflecting the strengthening effect of local spatial features centered on itself within the window. This represents the standard deviation of the Gaussian function, and its range is determined based on the grid side length. It is a natural constant; Based on the weighted landscape disturbance index, weighted landscape vulnerability index, and weighted landscape loss index of the grid evaluation unit, the corresponding landscape ecological risk value is determined: in, Landscape ecological risk value; , , These are the preset weights for the corresponding indicators, and they satisfy... ; By traversing all grid evaluation units through a sliding window, the landscape ecological risk value of each grid unit is obtained, thereby constructing a spatial distribution map of the landscape ecological risk value.

[0009] Furthermore, ecological stress and ecological vulnerability are quantified using ecological stress and ecological vulnerability indices, respectively, based on the following formulas: In the formula, Indicates the first Ecological stress index of each grid evaluation unit For the first Population density of each grid evaluation unit, For the index of the raster evaluation unit; For the first Industrial emission intensity of each grid evaluation unit; For the first Land use intensity of each grid evaluation unit; , and These represent the maximum values ​​of population density, industrial emission intensity, and land use intensity within the target area, respectively. , and The preset weights for the corresponding indicators, ; In the formula, Indicates the first Ecological vulnerability index of each grid assessment unit; , The first The minimum and maximum values ​​of each vulnerability indicator within the target area; Indicates the first The weights of each vulnerability indicator; An index for vulnerability indicators, The number of vulnerability indicators; Indicates the first The first grid evaluation unit Vulnerability index value; The vulnerability indicators specifically include vegetation cover, soil erosion modulus, water scarcity index, habitat fragmentation, and biodiversity index. Among them, vegetation coverage is extracted by processing the normalized vegetation index of remote sensing images and using a pixel-division model. The water shortage index is defined as the ratio of the total water resources within a grid evaluation unit to the total water consumption of that unit. Habitat fragmentation is defined as the ratio of patch density to average patch area in a grid assessment unit. The biodiversity index is calculated using the Shannon-Wiener index, based on the following formula: In the formula, For biodiversity index, Indicates the first The relative abundance of each species, For species indexing, Number of species; The soil erosion modulus is calculated using the general soil loss equation; The regional ecological risk assessment index system specifically includes the following assessment indicators: population density, industrial emission intensity, land use intensity, vegetation cover, soil erosion modulus, water scarcity index, habitat fragmentation, and biodiversity index.

[0010] Furthermore, for each grid evaluation unit, the regional ecological risk assessment index system is used to construct an 8×8 judgment matrix by comparing each of the eight assessment indicators in pairs. , where matrix elements Indicates the first Evaluation indicator relative to the first The importance of each evaluation indicator is assigned a value using a 1-9 scale. The subjective weights of each evaluation indicator were determined using the Bayesian BWM method. First, the optimal indicator was determined from the eight evaluation indicators. and worst-case indicators Then construct the optimal index respectively. The relative importance vector of all other evaluation indicators and worst-case indicators The relative importance vector of all other evaluation indicators ;in, and Each component in the table uses a scale of 1 to 9, representing the optimal index relative to the 1st rank. The importance of the evaluation indicators, and the first The importance of each evaluation indicator relative to the worst-case indicator, among which It is an index of evaluation indicators other than the optimal indicator. It is an index of evaluation metrics other than the worst-case indicator; based on a Bayesian inference framework, it uses vectors. and For the observed data, a posterior probability model is constructed and solved iteratively by Markov chain Monte Carlo sampling to obtain the subjective weight vector of each evaluation index, and the sum of the subjective weight vectors of all evaluation indexes is 1. The objective weights of each evaluation indicator are determined using the CRITIC method, based on the following formula: In the formula, Indicates the first The objective weight of each evaluation indicator An index for evaluation metrics; Indicates the first The information content of each evaluation indicator is used to reflect the first The degree of fluctuation of the evaluation index across all grid evaluation units; It also serves as an index for evaluation indicators; For the first The standard deviation of each evaluation indicator across all grid evaluation units was obtained through statistical calculation. For the first Evaluation indicators and the first Pearson correlation coefficients among the evaluation indicators; Calculate the Euclidean distance between the subjective weight vector and the objective weight vector. The dynamic fusion coefficient is calculated based on this distance. Combining subjective and objective weights, the comprehensive weight of each evaluation indicator is obtained: In the formula, For the first The overall weight of each evaluation indicator; For the first Subjective weighting of each evaluation indicator; This is the dynamic fusion coefficient.

[0011] Furthermore, the raw values ​​of each assessment indicator in the regional ecological risk assessment indicator system are normalized and unified to... For positive indicators, maximum value normalization is used, based on the following formula: For negative indicators, minimum value normalization is used, based on the following formula: In the formula, Indicates the first The first grid evaluation unit The raw values ​​of each evaluation indicator; and Indicates the first The minimum and maximum values ​​of each evaluation indicator within the target area; For the first The first grid evaluation unit Normalized values ​​of the evaluation indicators; The higher the value of the positive index, the higher the ecological risk. These include population density, industrial emission intensity, land use intensity, soil erosion modulus, water scarcity index, and habitat fragmentation. The higher the value of the negative index, the lower the ecological risk. These include vegetation cover and biodiversity index. The normalized evaluation index values ​​are weighted and summed using a comprehensive weighting method to obtain the regional ecological risk value for each grid evaluation unit: In the formula, For the first Regional ecological risk value of each grid assessment unit Indicates the first The overall weight of each evaluation indicator; and For preset weights, and ; The landscape ecological risk values ​​and regional ecological risk values ​​of all grid evaluation units within the target area are integrated to generate a grid layer, and a spatial distribution map of landscape ecological risk values ​​and a spatial distribution map of regional ecological risk values ​​are constructed based on a geographic information system.

[0012] Furthermore, by overlaying and analyzing the landscape ecological risk value and regional ecological risk value of each grid evaluation unit, the comprehensive ecological risk value of that grid unit is calculated. The calculation formula is as follows: In the formula, Indicates the first The comprehensive ecological risk value of each grid assessment unit. , The first Landscape ecological risk value and regional ecological risk value of each grid assessment unit For the index of the raster evaluation unit; , For the preset weights of the corresponding items, And satisfy ; The natural breakpoint method is used to cluster and classify the comprehensive ecological risk values ​​of all grid assessment units. Specifically, the comprehensive ecological risk values ​​of all grid assessment units within the target area are arranged in ascending order to form an ordered sequence. Based on this ordered sequence, iterative optimization calculations are performed to find the result that minimizes the within-group variance and maximizes the between-group variance. The comprehensive ecological risk value is divided into several levels based on natural discontinuities. Raster assessment units with a comprehensive ecological risk value no greater than the first natural discontinuity are classified into one risk level; raster assessment units with a risk value greater than the first but not greater than the second natural discontinuity are classified into another risk level, and so on, thus classifying the comprehensive ecological risk value into several levels. A series of consecutive levels are used to generate a comprehensive ecological risk classification map; After generating the comprehensive ecological risk classification map, the specific method for determining the early warning time of the risk is as follows: Based on the comprehensive ecological risk value and its corresponding risk level of each grid evaluation unit, and based on the trend of the comprehensive ecological risk value of the grid evaluation unit in the past multiple evaluation periods, the linear trend extrapolation method is used to predict the time required for its comprehensive ecological risk value to reach the preset high-risk threshold, and this time is marked as the risk early warning time of the grid evaluation unit; finally, the risk early warning time of each grid evaluation unit is associated with its spatial location to form a comprehensive ecological risk spatiotemporal distribution map with time-series attributes.

[0013] Compared with the prior art, the beneficial effects of the present invention are: First, this invention constructs a multi-source assessment framework, comprehensively integrates landscape pattern data and regional environmental socio-economic data, and uses a sliding window algorithm to extract local spatial features, thereby achieving a refined and spatial assessment of ecological risks. This effectively overcomes the shortcomings of traditional methods in characterizing spatial heterogeneity and local correlation, and significantly improves the accuracy and spatial resolution of risk distribution identification. Secondly, this invention introduces a subjective and objective weight fusion mechanism that combines the Bayesian BWM method and the CRITIC method, as well as a dynamic weighting strategy based on the distance function, which makes the determination of the weights of the evaluation indicators more scientific and robust, taking into account both expert experience and fully mining the inherent information of the data, thereby improving the adaptability and reliability of the ecological risk assessment system. Furthermore, by superimposing landscape ecological risks and regional ecological risks, and combining natural fault point classification and risk impact time prediction models, this invention achieves a leap from static evaluation to dynamic prediction, which can provide time-series early warning support for regional ecological risk management and help to formulate more forward-looking and targeted risk prevention and ecological restoration strategies. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the overall method flow of the present invention; Figure 2 A dual Y-axis image showing the landscape ecological risk value, regional ecological risk value, and comprehensive ecological risk value; Figure 3 A 3D scatter image of landscape ecological risk value, regional ecological risk value and comprehensive ecological risk value. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0016] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0017] Example: Please see Figures 1-3 The present invention provides a technical solution: The ecological risk assessment and prediction method based on a multi-source assessment framework includes the following steps: Step 1: Divide the target area into multiple grid evaluation units, collect landscape pattern data and regional environmental socio-economic data for each grid evaluation unit during the evaluation period, process the landscape pattern data using the sliding window method, obtain the interference, vulnerability and loss of the landscape pattern data during the evaluation period, and then determine the landscape ecological risk value of each grid evaluation unit. In this embodiment, the target area is first divided into spatially continuous grid evaluation units. The size of each grid is set to 100 meters × 100 meters according to the regional spatial scale and data resolution, so as to achieve a detailed depiction of the landscape pattern and human activities.

[0018] The landscape pattern data specifically includes the area, patch density, number of patches, and average patch area of ​​each landscape type. The regional environmental socio-economic data specifically includes population density, land use intensity, and industrial emission intensity. Among them, the area, patch density, number of patches, and average patch area of ​​each landscape type are obtained by processing high-resolution remote sensing imagery and land use classification data; population density is obtained from statistical yearbooks and population census data; and industrial emission intensity data comes from environmental statistical yearbooks, industrial pollution source censuses, or air pollutant emission inventories. Among them, "patches" refer to geographical units that have the same or similar land cover types and are spatially continuous in remote sensing images or land use classification maps. They are the basic spatial units for landscape pattern analysis.

[0019] Among them, the land use intensity Determined based on the following formula: In the formula, For the first grid evaluation unit Area of ​​land use type For the first The intensity weights corresponding to different land use types are determined based on the degree of land development and the intensity of human activities. The total number of land use types, An index for land use types; The total area of ​​the grid evaluation unit; The land use types include construction land, cultivated land, forest land, grassland, water area, and unused land, and their corresponding weighting relationships are set as follows: Construction land arable land grassland woodland waters Unused land is assigned the highest weight to construction land because it drastically alters the natural structure of the land surface, and the concentrated high-intensity human activities lead to the basic loss of ecological functions and difficulty in natural recovery. Cultivated land is assigned the second highest weight because it requires frequent cultivation, irrigation, and the application of chemical fertilizers and pesticides, resulting in continuous and significant disturbances to soil, hydrology, and biodiversity. Grassland is assigned a higher weight than forest land and water area because it is mostly in a semi-natural or artificially managed state, easily affected by grazing, reclamation, and other activities, and has weak ecological stability. Forest land is assigned a lower weight because it usually retains a better natural vegetation structure, has strong ecological regulation and protection functions, and has relatively strong resistance to disturbances. Water area is assigned a lower weight, mainly because it has strong natural fluidity and a long ecological recovery cycle, but is far-reachingly affected by direct human interference (such as pollution and reclamation). Unused land is assigned the lowest weight because its ecological baseline functions are weak, such as bare land and sandy land, with little human activity, and its contribution to the overall pressure on the ecosystem is minimal.

[0020] Land use intensity Used to characterize the comprehensive pressure of land use patterns on the natural ecosystem within a target area of ​​the grid evaluation unit, when A larger value indicates that the land use structure of this unit is dominated by high-intensity development, resulting in significant changes in the natural attributes of the land, strong disturbance to its ecosystem service functions, and low ecosystem stability; conversely, when... A smaller value indicates that land use is mainly low-intensity or natural, the degree of land development is weak, the ecological pressure is small, and the ecosystem is relatively stable.

[0021] The disturbance degree, vulnerability degree, and loss degree are quantified by the landscape disturbance degree index, landscape vulnerability index, and landscape loss degree index, respectively.

[0022] The specific processing procedure for the landscape pattern data using the sliding window algorithm is as follows: A 3×3 square moving window is established centered on each raster evaluation unit, and it slides row by row and column by column to cover the entire target area; for each window position, the landscape pattern data of all raster evaluation units within the window is extracted, and based on the landscape pattern data, the landscape disturbance index, landscape vulnerability index, and landscape loss index of each raster evaluation unit within the current window range are calculated, using the following formulas: in, For the current window range, the first Landscape disturbance index of each grid evaluation unit; For the current window range, the first The number of patches within each grid evaluation unit; For the current window range, the first The total area of ​​each grid evaluation unit This is the index of the raster evaluation cell within the current window range; For the current window range, the first The anthropogenic disturbance intensity coefficient for each grid evaluation unit is determined based on the population density and industrial emission intensity within the grid evaluation unit. Sure The formula used is as follows: In the formula, For the first Population density of each grid evaluation unit; For the first Industrial emission intensity of each grid evaluation unit; These are preset weights used to reflect the differences in the contributions of population activity and industrial emissions to landscape disturbance. .

[0023] Used to characterize the degree of human activity disturbance to the landscape pattern within the current grid evaluation unit, reflecting the spatial structure changes and stability damage of landscape patches within the unit caused by human activities. A larger value indicates a greater number of landscape patches within the unit, a higher degree of spatial fragmentation, greater intensity of human disturbance, lower landscape stability, and susceptibility to ecological damage; conversely, a smaller value indicates a lower value. A smaller value indicates that the landscape patch structure within the unit is relatively intact and the distribution is stable, with weak human disturbance and the ecosystem maintaining a good natural state. The more patches per unit area, the more the landscape is fragmented into smaller and more scattered patches, the more complex the landscape structure tends to be, the lower the connectivity, and the lower the overall stability, thus increasing the degree of landscape disturbance. The human disturbance intensity coefficient is determined based on population density and industrial emission intensity combined with expert assessment, and it reflects the comprehensive intensity of human activities. In areas with high population density and high industrial emission intensity, human activities are frequent and intense, directly or indirectly changing the surface cover and disrupting the continuity of the landscape, thereby increasing the probability of patch formation and the instability of patch structure, further amplifying the degree of disturbance.

[0024] in, For the current window range, the first Landscape vulnerability index for each grid assessment unit; Within the current window range, the first Ecologically sensitive landscape types in the first Area in each grid evaluation unit; The total number of landscape types, For the first The ecological sensitivity weights for different landscape types were determined through expert evaluation. An index for landscape types; The landscape types include forest landscapes, grassland landscapes, farmland landscapes, wetland landscapes, built-up land landscapes, and bare land landscapes. The relative order of their corresponding ecological sensitivity weights is set as follows: Forest landscape = Wetland landscape > Grassland landscape = Farmland landscape > Built-up land landscape > Bare land landscape. This is because forests are the core carriers of terrestrial biodiversity conservation, possessing significant functions in water conservation, soil and water conservation, carbon sequestration, and climate regulation. Their complex structure and long recovery cycle mean that damage can easily lead to habitat fragmentation and species loss. Wetlands, as the "kidneys of the earth," play a crucial role in hydrological regulation and water purification. Grassland plays a crucial role in areas such as biodiversity, biological habitat, and carbon storage, and its hydrological and biological processes are extremely sensitive to external disturbances, making natural recovery difficult after degradation. Grassland and farmland landscapes are less sensitive: grasslands have some surface cover and soil conservation functions, but their ecosystem structure is relatively simple, their resistance to disturbance is weak, and they are easily affected by human activities such as grazing and reclamation, leading to degradation. Although farmland is an artificially managed landscape, its soil health, irrigation system, and surrounding habitat have a significant impact on regional ecological security and the sustainability of agricultural production, and farming activities easily cause soil erosion and non-point source pollution. Built-up land landscapes are mainly hardened surfaces, with largely lost natural ecological functions. Their response to external disturbances is mainly manifested in socio-economic benefits rather than ecological processes, thus exhibiting low sensitivity. Bare land landscapes, with sparse vegetation and weak ecological functions, are least affected by marginal impacts from human or natural disturbances, hence exhibiting the lowest sensitivity. This weighting system reflects the sensitivity gradient from natural high-function landscapes to artificial low-function landscapes, conforming to general ecological principles and providing a scientific basis for prioritizing regional ecological protection and risk management. The expert evaluation adopts the Delphi method or the analytic hierarchy process, inviting experts in ecology, geography, environmental planning and other fields to conduct multiple rounds of independent scoring and feedback adjustments on the sensitivity of each landscape type based on the actual situation of the region, and finally form a consensus on the weight assignment, so as to ensure that the weight system not only conforms to general ecological laws, but also fits the specific characteristics of the evaluation area.

[0025] Used to characterize the The degree to which landscape types within a grid assessment unit are susceptible to damage or difficult recovery from external disturbances reflects the sensitivity and potential ecological vulnerability of different ecologically sensitive landscape types within that unit to various stresses. A larger value indicates a higher proportion of highly sensitive landscape types within the unit, resulting in a weaker overall landscape structure that is less resistant to external disturbances, and a more vulnerable and slower-recovering ecosystem; conversely, a smaller value indicates a lower proportion of sensitive landscape types within the unit. A smaller value indicates that the landscape types within the unit are mainly of low sensitivity or high stability, with strong overall resistance to disturbance and relatively low ecological risk. If the area of ​​highly sensitive landscape types within the unit is large, their high weight in the weighted summation also increases, leading to an increase in the vulnerability index. This is because highly sensitive landscape types typically have important ecological functions, complex structures, are sensitive to environmental changes, and have weak resistance to disturbance. An increase in their area directly means a decrease in the overall ecological stability of the unit and an increase in potential vulnerability.

[0026] In the formula, For the current window range, the first Landscape loss index of each grid evaluation unit; For the first Ecosystem service value per unit area of ​​landscape type For the first in the target area Historical loss rate of landscape types; in, The determination method is as follows: First, referring to existing ecosystem service value assessment tables, a standard value coefficient is set for each landscape type. Then, considering the ecological function importance and restoration cost of the target area, the coefficient is adjusted through expert surveys to finally determine the coefficient. The value.

[0027] Among them, determine The method used is as follows: by collecting historical remote sensing images or land use data of the target area over the past 5 years, extracting the area change information of each landscape type at different time periods, calculating the area reduction ratio of each landscape type in continuous time intervals, and taking the 5-year average as the historical loss rate of that landscape type.

[0028] Used to characterize the degree of potential ecological function loss within the current raster evaluation unit due to historical loss of landscape types and loss of ecosystem service value, reflecting the cumulative impact of the decline in ecosystem service capacity of each landscape type within the unit caused by natural or anthropogenic factors; when A larger value indicates that the area of ​​landscape types with high ecosystem service value and severe historical losses is large within the unit, the ecosystem service function is significantly impaired, and the potential ecological restoration cost and risk are high; conversely, a smaller value indicates a larger area of ​​high ecosystem service value and severe historical losses within the unit, the greater the potential ecosystem service value and the higher the potential ecological restoration cost and risk. A lower value indicates that the landscape type within the unit has low ecosystem service value or slight historical loss, its ecological functions remain relatively intact, and its potential ecological risks are low. Ecosystem service value per unit area reflects the comprehensive value of this type of landscape in terms of ecological regulation, supply, support, and cultural services: landscape types with high ecosystem service value contribute more to the degree of ecological loss for the same area, as their loss or degradation significantly weakens the overall service function of the regional ecosystem; the historical loss rate is determined based on the proportion of area reduction of this type of landscape over a certain period, reflecting its disturbance and degradation trends: landscape types with high historical loss rates indicate that they continuously face strong degradation pressure or transformation risks in the region, and the ecosystem service function of their existing area is more likely to be unstable and continue to be lost.

[0029] The weighted interference index of the evaluation unit in the center of the window is calculated according to the following formula. Weighted Vulnerability Index and weighted loss index : in, This represents the total number of cells within the current window range. Indicates the first [number] within the current window range The influence weights of each grid evaluation unit on the window center grid evaluation unit are determined based on a Gaussian function: For the current window range, the first The Euclidean distance from each grid cell to the center of the window This represents the standard deviation of the Gaussian function, and its range is determined based on the grid side length. It is a natural constant.

[0030] In the calculation of the weighted disturbance index, weighted vulnerability index, and weighted loss index, the local contributions of grid cells within each window are weighted and summed using Gaussian function-based influence weights. This reflects the spatial attenuation effect, meaning that the closer a grid cell is to the center of the window, the greater its risk impact on the central cell. On the one hand, Gaussian weights can effectively capture the spatial continuity and local correlation of landscape patterns, making the landscape risk value of the central cell more realistically reflect the comprehensive impact of its surrounding environment. On the other hand, by introducing a distance attenuation mechanism, when calculating the contribution of each cell within the window, the direct impact of neighboring cells is emphasized, while the indirect effects of more distant cells are appropriately included. This maintains spatial smoothness while enhancing the robustness of risk assessment to local anomalies and edge effects.

[0031] Based on the weighted landscape disturbance index, weighted landscape vulnerability index, and weighted landscape loss index of the grid evaluation unit, the corresponding landscape ecological risk value is determined: in, Landscape ecological risk value; , , These are the preset weights for the corresponding indicators. And satisfy The reason for setting the weights in this way is as follows: Landscape vulnerability reflects the sensitivity and fragility of the ecosystem itself, and is the internal basic condition for the occurrence of risks. Its level directly determines the system's response to external disturbances. Therefore, it is given a higher weight to emphasize the dominant role of inherent vulnerability. Landscape disturbance represents the immediate impact of external pressures such as human activities on landscape structure. It is an important driving factor for risk triggering, but its role is often based on vulnerability. Therefore, its weight is slightly lower than that of vulnerability. Landscape loss is based on historical change data and reflects the trend of ecological function loss that has already occurred. It has a certain lag and cumulative nature and is more used for risk verification and trend indication. Therefore, it is given a relatively lower weight to highlight the assessment priority of the current state and immediate disturbances, so that the risk assessment is closer to the early warning and control needs in actual management.

[0032] Landscape ecological risk value Used to characterize the ecological risk level faced by the current grid assessment unit at the landscape scale, reflecting the possibility of potential ecosystem degradation or functional loss within the unit due to the combined effects of human activity disturbance, landscape inherent vulnerability, and historical ecological losses; when A larger value indicates that the landscape structure of this unit is highly disturbed, ecologically vulnerable, and has suffered significant historical losses, resulting in poor overall ecosystem stability and a high level of ecological risk; conversely, a smaller value indicates a lower value. A smaller value indicates that the landscape structure of the unit is relatively stable, has strong resistance to disturbance, and has a low degree of ecological loss, indicating a relatively healthy ecosystem and low ecological risk.

[0033] By traversing all grid evaluation units through a sliding window, the landscape ecological risk value of each grid unit is obtained, thereby constructing a spatial distribution map of the landscape ecological risk value.

[0034] Step 2: Based on regional environmental and socio-economic data, determine the ecological pressure and ecological vulnerability of each grid assessment unit, and construct a regional ecological risk assessment index system for the grid assessment unit; In this embodiment, ecological stress and ecological vulnerability are quantified using ecological stress index and ecological vulnerability index, respectively, based on the following formulas: In the formula, Indicates the first Ecological stress index of each grid evaluation unit For the first Population density of each grid evaluation unit, For the index of the raster evaluation unit; For the first Industrial emission intensity of each grid evaluation unit; For the first Land use intensity of each grid evaluation unit; , and These represent the maximum values ​​of population density, industrial emission intensity, and land use intensity within the target area, respectively. , and The preset weights for the corresponding indicators, ,and The reasons for setting the weights in this way are as follows: Population density, as a fundamental and continuous source of pressure, not only directly drives resource consumption and space occupation, but also indirectly affects industrial layout and land development intensity, making it a fundamental factor of ecological pressure. Therefore, it is given the highest weight. Although industrial emission intensity contributes significantly to environmental pollution, its spatial distribution is relatively concentrated and is greatly affected by industrial structure and governance level. Its prevalence and sustainability at the regional scale are slightly lower than that of population pressure, so its weight is second. Land use intensity reflects more the physical changes in land cover and spatial pattern. Its ecological impact is often manifested through long-term accumulation and is strongly coupled with population and industrial activities. Its independence as a pressure indicator is relatively weak. Therefore, it is given a relatively low weight to highlight the leading role of population and industrial activities in ecological pressure assessment.

[0035] In this formula, Used to characterize the The degree of comprehensive external stress on the natural ecosystem caused by human activities within each grid assessment unit reflects the ecological load level of that unit under multiple pressures such as population concentration, industrialization, and land development; when A larger value indicates that the unit is densely populated, has high industrial emission intensity, and a high degree of intensive land use, resulting in strong human interference with the ecosystem, significant environmental carrying capacity pressure, and a major challenge to ecosystem stability; conversely, a smaller value indicates a lower value. A smaller value indicates that the impact of human activities on the unit is weaker, the ecological pressure is relatively low, and the natural environment is less disturbed.

[0036] Population density directly reflects the intensity of human settlement and activity: densely populated areas are usually accompanied by higher resource consumption, waste discharge, and land occupation, leading to the compression of ecological space and damage to biological habitats, thereby exacerbating ecological pressure. Industrial emission intensity characterizes the pollution load of industrialization on the environment: waste gas, wastewater, solid waste, etc. generated by industrial activities directly or indirectly enter the environment, causing a decline in the quality of air, water, and soil, disrupting the ecological balance, and further increasing ecological pressure. Land use intensity reflects the degree of intensive land development and use: high-intensity use often leads to a reduction in natural surface cover, a decrease in ecological connectivity, and an increase in soil erosion, weakening the self-regulation capacity of the ecosystem and increasing the vulnerability of the ecosystem.

[0037] In the formula, Indicates the first Ecological vulnerability index of each grid assessment unit; , The first The minimum and maximum values ​​of each vulnerability indicator within the target area; Indicates the first The weights of each vulnerability indicator; An index for vulnerability indicators, The number of vulnerability indicators; Indicates the first The first grid evaluation unit Vulnerability index value; Used to characterize the The potential sensitivity of an ecosystem within a grid assessment unit to natural or anthropogenic disturbances, highlighting its vulnerability and difficulty in recovery, comprehensively reflects the unit's inherent deficiencies and resilience across multiple dimensions of natural attributes, including vegetation cover, soil conservation, water resources, habitat structure, and biodiversity. A large value indicates that the unit has sparse vegetation, severe soil erosion, water shortage, intensified habitat fragmentation, or low biodiversity. The ecosystem structure is unstable, its functions are easily damaged, its overall resistance to disturbance is weak, and it faces a high risk of ecological degradation. Conversely, a small value indicates a high risk of ecological degradation. A smaller value indicates that the unit has dense vegetation, well-preserved soil, abundant water resources, continuous and intact habitat, and rich biodiversity. The ecosystem has a strong self-sustaining and recovery capacity, low sensitivity to external disturbances, and relatively small ecological risks.

[0038] The vulnerability indicators specifically include vegetation cover, soil erosion modulus, water scarcity index, habitat fragmentation, and biodiversity index. Among them, vegetation coverage is extracted by processing the normalized vegetation index of remote sensing images and using a pixel-division model. The water shortage index is defined as the ratio of the total water resources within a grid evaluation unit to the total water consumption of that unit. Habitat fragmentation is defined as the ratio of patch density to average patch area in a grid assessment unit. The biodiversity index is calculated using the Shannon-Wiener index, based on the following formula: In the formula, For biodiversity index, Indicates the first The relative abundance of each species is obtained based on species distribution data; For species indexing, Number of species; The soil erosion modulus is calculated using the general soil loss equation; The regional ecological risk assessment index system specifically includes the following assessment indicators: population density, industrial emission intensity, land use intensity, vegetation cover, soil erosion modulus, water scarcity index, habitat fragmentation, and biodiversity index.

[0039] Step 3: Convert the regional ecological risk assessment index system into a judgment matrix, use the Bayesian BWM method and the CRITIC method to determine the subjective and objective weights of each assessment index, and obtain the comprehensive weight of the assessment index through the distance function combination weighting method. In this embodiment, for the regional ecological risk assessment index system of each grid evaluation unit, the eight assessment indicators it contains are compared pairwise to construct an 8×8 judgment matrix. , where matrix elements Indicates the first Evaluation indicator relative to the first The importance of each evaluation indicator is assigned a value using a 1-9 scale. The subjective weights of each evaluation indicator were determined using the Bayesian BWM method. First, the optimal indicator was determined from the eight evaluation indicators. and worst-case indicators Then construct the optimal index respectively. The relative importance vector of all other evaluation indicators and worst-case indicators The relative importance vector of all other evaluation indicators ;in, and Each component in the table uses a scale of 1 to 9, representing the optimal index relative to the 1st rank. The importance of the evaluation indicators, and the first The importance of each evaluation indicator relative to the worst-case indicator, among which It is an index of evaluation indicators other than the optimal indicator. It is an index of evaluation metrics other than the worst-case indicator; based on a Bayesian inference framework, it uses vectors. and For the observed data, a posterior probability model is constructed and solved iteratively by Markov chain Monte Carlo sampling to obtain the subjective weight vector of each evaluation index, and the sum of the subjective weight vectors of all evaluation indexes is 1. Using the Bayesian Best-Worst Method (BWM) to determine subjective weights effectively integrates expert judgment with a probabilistic statistical framework. By quantifying weight uncertainty through Bayesian inference and MCMC sampling, it enhances the robustness and interpretability of weight estimation. Compared to traditional point-based estimation methods such as the analytic hierarchy process (AHP), Bayesian BWM not only considers the consistency of expert preferences but also provides weight distribution information, enhancing the assessment process's adaptability to fluctuations in subjective judgment. This allows for a more scientific and reliable determination of subjective weights in complex multi-indicator systems.

[0040] The objective weights of each evaluation indicator are determined using the CRITIC method, based on the following formula: In the formula, Indicates the first The objective weight of each evaluation indicator An index for evaluation metrics; Indicates the first The information content of each evaluation indicator is used to reflect the first The degree of fluctuation of the evaluation index across all grid evaluation units; It also serves as an index for evaluation indicators; For the first The standard deviation of each evaluation indicator across all grid evaluation units was obtained through statistical calculation. For the first Evaluation indicators and the first Pearson correlation coefficients among the evaluation indicators; This invention employs the CRITIC method to determine objective weights. Based on the volatility of the indicators themselves and the conflict between indicators, it can quantify the information content and discrimination of each indicator in the dataset, thereby achieving objective weight allocation and reducing subjective bias. Compared with the traditional entropy weight method or standard deviation method, CRITIC not only considers the internal variation of indicators but also introduces the correlation between indicators, which more comprehensively reflects the comprehensive information contribution of indicators and enhances the scientificity and robustness of the weight system.

[0041] Calculate the Euclidean distance between the subjective weight vector and the objective weight vector. The dynamic fusion coefficient is calculated based on this distance. Combining subjective and objective weights, the comprehensive weight of each evaluation indicator is obtained: In the formula, For the first The overall weight of each evaluation indicator; For the first Subjective weighting of each evaluation indicator; The dynamic fusion coefficient; Used to characterize the The comprehensive importance of each assessment indicator in regional ecological risk assessment reflects its relative contribution after integrating expert experience and the inherent information in the data, and is a key weighting coefficient for the final calculation of the regional ecological risk value; when A large value indicates that the indicator is recognized as highly important in both subjective perception and objective data, has a significant overall impact on ecological risk, and plays a leading role in risk assessment; conversely, a small value indicates a large indicator. A smaller value indicates that the indicator is relatively less important from both expert experience and data volatility perspectives, and its contribution to overall risk is limited; Dynamic fusion coefficient The introduction of this technology allows the weight fusion process to adaptively adjust the balance between subjective and objective weights: when the difference between subjective and objective weights is large, i.e. When the value is large, When the weighting approaches 1, the overall weighting leans more towards subjective weighting, emphasizing expert experience; when the difference between subjective and objective weightings is small, i.e. When the value is small, The weighting is reduced, and the overall weighting is closer to the objective weighting, highlighting the data-driven approach.

[0042] Step 4: Use the comprehensive weight of the evaluation indicators to perform weighted calculation on the normalized values ​​of each evaluation indicator to obtain the regional ecological risk value of the grid evaluation unit. Integrate the landscape ecological risk values ​​and regional ecological risk values ​​of all grid evaluation units to construct the spatial distribution of landscape ecological risk values ​​and the spatial distribution of regional ecological risk values. In this embodiment, the original values ​​of each assessment indicator in the regional ecological risk assessment indicator system are normalized to be consistent with the standard values. For positive indicators, maximum value normalization is used, based on the following formula: For negative indicators, minimum value normalization is used, based on the following formula: In the formula, Indicates the first The first grid evaluation unit The raw values ​​of each evaluation indicator; and Indicates the first The minimum and maximum values ​​of each evaluation indicator within the target area; For the first The first grid evaluation unit Normalized values ​​of the evaluation indicators; The higher the value of the positive index, the higher the ecological risk. These include population density, industrial emission intensity, land use intensity, soil erosion modulus, water scarcity index, and habitat fragmentation. The higher the value of the negative index, the lower the ecological risk. These include vegetation cover and biodiversity index. The normalized evaluation index values ​​are weighted and summed using a comprehensive weighting method to obtain the regional ecological risk value for each grid evaluation unit: In the formula, For the first Regional ecological risk value of each grid assessment unit Indicates the first The overall weight of each evaluation indicator; and The preset weights for the corresponding indicators, and This is because population agglomeration, industrial emissions, and increased land use intensity are often the most significant sources of ecological risk in the short term. They directly alter land cover, emit pollutants, encroach on ecological space, and cause proactive impacts on the ecosystem.

[0043] In this formula, Used to characterize the Each grid assessment unit, at a regional scale, is evaluated based on the combined effects of eight indicators: population density, industrial emission intensity, land use intensity, vegetation cover, soil erosion modulus, water scarcity index, habitat fragmentation, and biodiversity index. This reflects the potential for ecological degradation or functional loss within the unit under the dual influence of human socio-economic activities and the vulnerability of the natural environment. A large value indicates that the unit is densely populated, heavily polluted by industry, and has a high intensity of land development. Simultaneously, it suffers from low vegetation cover, severe soil erosion, water scarcity, fragmented habitats, and poor biodiversity. The overall ecosystem faces significant external pressure and inherent vulnerability, resulting in prominent ecological risks. Conversely, a small value indicates a high level of ecological risk. A smaller value indicates less human disturbance, a healthy natural environment, higher ecosystem stability, and relatively lower ecological risk in the unit; furthermore... and An increase in all of these will lead to An increase indicates a heightened ecological risk; It reflects the direct driving force of human activities, and its increase often rapidly pushes up the risk in the short term; This reflects the system's inherent vulnerability; its increase signifies a decrease in the system's resistance and recovery capabilities, further amplifying the negative effects of stress.

[0044] The landscape ecological risk values ​​and regional ecological risk values ​​of all grid evaluation units within the target area are integrated to generate a grid layer, and a spatial distribution map of landscape ecological risk values ​​and a spatial distribution map of regional ecological risk values ​​are constructed based on a geographic information system.

[0045] Step 5: Overlay the spatial distribution of landscape ecological risk values ​​with the spatial distribution of regional ecological risk values ​​to obtain the comprehensive ecological risk value of the grid evaluation unit. Use the natural breakpoint method to cluster and classify the comprehensive ecological risk values ​​of all grid evaluation units, generate a comprehensive ecological risk classification map, and determine the risk warning time to achieve ecological risk assessment and prediction. In this embodiment, the comprehensive ecological risk value of each grid evaluation unit is calculated by superimposing and analyzing the landscape ecological risk value and the regional ecological risk value. The calculation formula is as follows: In the formula, Indicates the first The comprehensive ecological risk value of each grid assessment unit. , The first Landscape ecological risk value and regional ecological risk value of each grid assessment unit For the index of the raster evaluation unit; , For the preset weights of the corresponding items, And satisfy The reason for setting the weights in this way is that, in most regional-scale ecological risk assessments, the direct and indirect pressures of human socio-economic activities on ecosystems are often more dominant and driving in the short term. Their impact is wide-ranging and intense, and they are often rapidly transmitted to ecosystems through changes in landscape patterns and increased resource consumption. While changes in landscape patterns and the loss of their ecosystem services are important, their response is often delayed and they often exist as a result or intermediate state of human activities. Therefore, assigning higher weights to regional ecological risks is more in line with the actual risk formation mechanism, and helps to highlight the most active and controllable risk sources in the current and near term in the comprehensive assessment, thereby improving the timeliness and pertinence of risk management.

[0046] Used to characterize the The overall ecological risk level of each grid evaluation unit under the dual dimensions of comprehensive landscape pattern characteristics and regional socio-economic environment reflects the comprehensive ecological degradation probability and early warning level of the unit under the coupled effects of multiple factors such as landscape structure stability, historical loss of ecosystem service value, human activity pressure, and natural environment vulnerability; when A large value indicates that the unit not only suffers from severe landscape structure disturbance, high ecological vulnerability, and significant historical losses, but also faces significant socio-economic pressures, weak resilience of the natural system, and a highly complex ecological risk, resulting in a severe ecological security situation; conversely, a small value indicates a large ecological risk. A smaller value indicates that the landscape pattern of the unit is relatively complete, the ecological pressure is low, the natural environment is good, the ecosystem is relatively stable, and the ecological risks are controllable.

[0047] Landscape ecological risk values ​​influence overall risk from the perspective of changes in landscape spatial structure and historical ecosystem services: high landscape disturbance reflects patch fragmentation and decreased connectivity caused by human activities; high landscape vulnerability indicates that ecologically sensitive types are easily damaged; and large landscape loss indicates the continuous degradation of historical ecosystem service functions. These factors together weaken the ecological carrying capacity and recovery capacity of the landscape, thereby increasing overall risk. Regional ecological risk values ​​exert influence from the dimensions of human activity intensity and natural system vulnerability: dense population, high industrial emission intensity, and high land use intensity directly increase ecological pressure, while low vegetation cover, strong soil erosion, water shortage, habitat fragmentation, and low biodiversity reduce the system's self-regulation and resistance to disturbance. The combination of these two factors further exacerbates the risk of ecosystem imbalance.

[0048] Table 1: Statistical Table of Comprehensive Ecological Risk Values Analysis of the data in Table 1 shows that, from the overall trend of the data, the comprehensive ecological risk value, the landscape ecological risk value, and the regional ecological risk value all show a significant positive correlation. Moreover, due to the difference in weights, the regional ecological risk plays a more significant role in determining the comprehensive risk, reflecting the dominant position of regional socio-economic and human stress factors in this assessment.

[0049] The risk value gradually climbs from the low-risk range to the high-risk level, showing a smooth and continuous gradient change. This indicates that the risk structure of each grid evaluation unit in the assessment system is consistent, without any abnormal fluctuations or abrupt changes, which is consistent with the transitional characteristics of the actual spatial distribution of ecological risks. This consistent change also provides a reliable basis for subsequent risk level classification and spatial visualization, and can support risk contribution analysis based on weight differences and in-depth interpretation of the superposition effect of multiple risk sources.

[0050] The natural breakpoint method is used to cluster and classify the comprehensive ecological risk values ​​of all grid assessment units. Specifically, the comprehensive ecological risk values ​​of all grid assessment units within the target area are arranged in ascending order to form an ordered sequence. Based on this ordered sequence, iterative optimization calculations are performed to find the result that minimizes the within-group variance and maximizes the between-group variance. The comprehensive ecological risk value is divided into several levels based on natural discontinuities. Raster assessment units with a comprehensive ecological risk value no greater than the first natural discontinuity are classified into one risk level; raster assessment units with a risk value greater than the first but not greater than the second natural discontinuity are classified into another risk level, and so on, thus classifying the comprehensive ecological risk value into several levels. A series of consecutive levels are used to generate a comprehensive ecological risk classification map.

[0051] The natural breakpoint method can automatically determine the risk level classification threshold based on the inherent statistical distribution characteristics of the comprehensive ecological risk value, avoiding the subjectivity of artificially setting boundaries, ensuring that the differences within the same risk level are minimal and the differences between different levels are maximum, thereby objectively reflecting the spatial differentiation pattern of ecological risk, improving the scientificity and explanatory power of the classification results, and providing a reliable basis for the refined management of risks and spatially differentiated policies.

[0052] After generating the comprehensive ecological risk classification map, the specific method for determining the early warning time of the risk is as follows: Based on the comprehensive ecological risk value and its corresponding risk level of each grid evaluation unit, and based on the trend of the comprehensive ecological risk value of the grid evaluation unit in the past multiple evaluation periods, the linear trend extrapolation method is used to predict the time required for its comprehensive ecological risk value to reach the preset high-risk threshold, and this time is marked as the risk early warning time of the grid evaluation unit; finally, the risk early warning time of each grid evaluation unit is associated with its spatial location to form a comprehensive ecological risk spatiotemporal distribution map with time-series attributes; The method for determining the preset high-risk threshold is as follows: based on the historical comprehensive ecological risk value data of all grid evaluation units in the target area, calculate its statistical distribution characteristics; use K-means clustering to divide the comprehensive ecological risk value of all grid units into three categories: high, medium and low, and use the lower limit of the high-risk category as the preset high-risk threshold.

[0053] The above formulas are all dimensionless calculations. 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 in the formulas are set by those skilled in the art according to the actual situation.

[0054] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0055] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0056] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. An ecological risk assessment and prediction assessment method based on a multi-source evaluation framework, characterized in that, The method comprises the following steps: dividing the target area into a plurality of grid evaluation units, collecting landscape pattern data and regional environmental and social economic data of each grid evaluation unit in the evaluation period, processing the landscape pattern data by the sliding window method to obtain the interference degree, vulnerability degree and loss degree of the landscape pattern data in the evaluation period, and then determining the landscape ecological risk value of each grid evaluation unit; based on the regional environmental and social economic data, determining the ecological pressure and ecological vulnerability of each grid evaluation unit, and constructing a regional ecological risk assessment index system for the grid evaluation unit; converting the regional ecological risk assessment index system into a judgment matrix, using the Bayesian BWM method and the CRITIC method to determine the subjective and objective weights of each evaluation index, and using the distance function combination weighting method to obtain the comprehensive weight of the evaluation index; using the comprehensive weight of the evaluation index to weight the normalized value of each evaluation index, obtaining the regional ecological risk value of the grid evaluation unit, and integrating the landscape ecological risk value and the regional ecological risk value of all grid evaluation units to construct the spatial distribution of the landscape ecological risk value and the spatial distribution of the regional ecological risk value; superimposing and analyzing the spatial distribution of the landscape ecological risk value and the spatial distribution of the regional ecological risk value to obtain the comprehensive ecological risk value of the grid evaluation unit, using the natural breakpoint method to cluster and classify the comprehensive ecological risk value of all grid evaluation units, generating a comprehensive ecological risk classification map and determining a risk warning time to realize ecological risk assessment and prediction.

2. The ecological risk assessment and prediction evaluation method based on the multi-source evaluation framework according to claim 1, characterized in that: The landscape pattern data specifically includes the area, patch density, patch number and patch average area of each landscape type, and the regional environmental and social economic data specifically includes population density, land use intensity and industrial emission intensity; The land use intensity is obtained by multiplying the area of each type of land use in the grid evaluation unit by the corresponding intensity weight and then summing, and then dividing by the total area of the grid evaluation unit; wherein the intensity weight is determined based on the development degree of each type of land and the intensity of human activities; The interference degree, vulnerability degree and loss degree are quantitatively represented by the landscape interference degree index, the landscape vulnerability degree index and the landscape loss degree index respectively.

3. The ecological risk assessment and prediction evaluation method based on the multi-source evaluation framework of claim 2, characterized in that: The specific processing process of the landscape pattern data by the sliding window algorithm is as follows: taking each grid evaluation unit as a center, a square moving window with a fixed size of is constructed, wherein is an odd number greater than 1. Take each grid evaluation unit as the center of the window, and slide row by row and column by column to cover the entire target area; for each window position, extract the landscape pattern data of all grid evaluation units in the window, and based on the landscape pattern data, calculate the landscape interference degree index, the landscape vulnerability degree index and the landscape loss degree index of each grid evaluation unit in the current window range; The calculation of the landscape disturbance index is realized according to the following method: the number of patches of a grid evaluation unit in a current window range is divided by the total area of the grid evaluation unit, and then multiplied by the corresponding artificial disturbance intensity coefficient; wherein the artificial disturbance intensity coefficient is determined based on the population density and industrial emission intensity in the grid evaluation unit. the number of patches of a grid evaluation unit in a current window range is divided by the total area of the grid evaluation unit, and then multiplied by the corresponding artificial disturbance intensity coefficient; wherein the artificial disturbance intensity coefficient is determined based on the population density and industrial emission intensity in the grid evaluation unit. The landscape vulnerability index is calculated using the following method: Within the current window range, the first... The ratio of the area of ​​each ecologically sensitive landscape type in a grid evaluation unit to the total area of ​​that unit is calculated by multiplying the corresponding ecological sensitivity weight of the landscape type by the ratio of the ratio of the area of ​​each type to the total area of ​​the grid evaluation unit. The ecological sensitivity weight is determined by expert evaluation. The calculation of the landscape loss degree index is realized according to the following method: the area of each type of landscape in the evaluation unit of the grid in the current window range is multiplied by the unit area ecological service value of the corresponding landscape type and the historical loss rate of the landscape type in the target region, and then the sum of all landscape types is obtained. The area of each type of landscape in the evaluation unit of the grid in the current window range is multiplied by the unit area ecological service value of the corresponding landscape type and the historical loss rate of the landscape type in the target region, and then the sum of all landscape types is obtained.

4. The ecological risk assessment and prediction evaluation method based on the multi-source evaluation framework of claim 3, characterized in that: The method for determining the weighted interference degree index, the weighted vulnerability degree index and the weighted loss degree index is as follows: multiply the landscape interference degree index, the landscape vulnerability degree index or the landscape loss degree index of each grid evaluation unit in the window by the influence weight of each unit on the center unit of the window, sum them up, and then divide by the sum of all weights to obtain the weighted index corresponding to the center grid evaluation unit of the window; wherein the influence weight is determined based on the Gaussian function, and its value is negatively correlated with the Euclidean distance of each grid unit to the center of the window. The weighted landscape interference index, the weighted landscape vulnerability index and the weighted landscape loss index of the grid evaluation unit are evaluated, and then the corresponding landscape ecological risk value is determined, specifically: the weighted landscape interference index, the weighted landscape vulnerability index and the weighted landscape loss index are multiplied by the corresponding preset weight respectively, and then summed to obtain the landscape ecological risk value; wherein the sum of the preset weights is 1; The landscape ecological risk value of each grid unit is obtained by traversing all grid evaluation units through a sliding window, and a spatial distribution map of the landscape ecological risk value is constructed.

5. The ecological risk assessment and prediction evaluation method based on the multi-source evaluation framework of claim 2, wherein: Ecological stress and ecological vulnerability are quantified using ecological stress and ecological vulnerability indices, respectively. The method used is as follows: [The text abruptly ends here, likely due to an incomplete sentence or missing information.] The population density of each grid assessment unit is divided by its maximum population density within the entire target area to obtain the normalized population density contribution value. Similarly, the industrial emission intensity and land use intensity are divided by their respective maximum values ​​within the region to obtain their corresponding normalized contribution values. These three normalized values ​​are multiplied by preset weights and then summed to obtain the ecological pressure index of the grid assessment unit. For the index of the raster evaluation unit; For the first Each vulnerability index value in each grid evaluation unit is normalized. This is done by subtracting the minimum value of that vulnerability index within the target area from the original vulnerability index value, then dividing by the difference between the maximum and minimum values ​​to obtain the normalized result for each index. Next, each normalized index value is multiplied by its corresponding weight, and the weighted sum of all indexes is calculated and divided by the total number of indicators to obtain the final result. Ecological vulnerability index of each grid assessment unit; The vulnerability index specifically includes vegetation coverage, soil erosion modulus, water resource shortage index, habitat fragmentation and biodiversity index; The vegetation coverage is extracted by processing the normalized vegetation index of the remote sensing image using the pixel bisection model; The water resource shortage index is defined as the ratio of the total amount of water resources in the grid evaluation unit to the total amount of water used in the unit; The habitat fragmentation is defined as the ratio of the patch density to the average patch area of the grid evaluation unit; The biodiversity index is calculated using the Shannon-Wiener index; The soil erosion modulus is calculated using the Universal Soil Loss Equation; The regional ecological risk assessment index system specifically includes the following assessment indexes: population density, industrial emission intensity, land use intensity, vegetation coverage, soil erosion modulus, water resource shortage index, habitat fragmentation and biodiversity index.

6. The ecological risk assessment and prediction evaluation method based on the multi-source evaluation framework of claim 5, characterized in that: The regional ecological risk assessment index system of each grid evaluation unit is evaluated, eight evaluation indexes contained therein are compared with each other, and an 8×8 judgment matrix is constructed , wherein matrix elements represent the importance degree of the i th evaluation index relative to the j th evaluation index, and a 1-9 scale method is used for value assignment; The subjective weight of each evaluation index is determined by using the Bayesian BWM method. First, the optimal index and the worst index are determined from the eight evaluation indexes Then, the relative importance vector of the optimal index to all other evaluation indexes and the relative importance vector of the worst index to all other evaluation indexes are constructed, wherein each component in and is marked by a scale of 1-9, representing the importance of the optimal index relative to the first evaluation index and the importance of the first evaluation index relative to the worst index, wherein is the index of the evaluation index other than the optimal index, is the index of the evaluation index other than the worst index; based on the Bayesian inference framework, the posterior probability model is constructed with the observation data of the vector and , and the subjective weight vector of each evaluation index is obtained by Markov chain Monte Carlo sampling iteration, and the sum of the subjective weight vectors of all evaluation indexes is 1.​ 7. The ecological risk assessment and prediction evaluation method based on the multi-source evaluation framework of claim 5, characterized in that: When the CRITIC method is used to determine the objective weight of each assessment index, first, for each assessment index, the standard deviation in all grid evaluation units is calculated; then the Pearson correlation coefficient between the assessment index and each other assessment index is calculated, and then the information amount of the assessment index is calculated; finally, the information amount of each assessment index is divided by the sum of the information amounts of all assessment indexes to obtain the objective weight of the assessment index; When determining the comprehensive weight of each assessment index, first, the overall difference between the subjective weight vector and the objective weight vector is measured by calculating the Euclidean distance between them; based on the Euclidean distance, a dynamic fusion coefficient is determined, which tends to one as the distance increases and tends to zero as the distance decreases; finally, the subjective weight and the objective weight of each assessment index are combined by weighting according to the dynamic fusion coefficient to obtain the comprehensive weight of the assessment index.

8. The ecological risk assessment and prediction evaluation method based on the multi-source evaluation framework of claim 4, wherein: The original value of each evaluation index in the regional ecological risk evaluation index system is normalized to an interval; wherein, for a positive index, maximum value normalization is adopted, and for a negative index, minimum value normalization is adopted; The greater the positive index value, the higher the ecological risk, including population density, industrial emission intensity, land use intensity, soil erosion modulus, water resource shortage index and habitat fragmentation; the greater the negative index value, the lower the ecological risk, including vegetation coverage and biodiversity index; The normalized value of each assessment index is multiplied by its corresponding comprehensive weight, and the weighted results of all assessment indexes are summed; on this basis, the ecological pressure index and the ecological vulnerability index are introduced, and they are multiplied by their respective preset weights and added to the weighted sum to obtain the regional ecological risk value of the grid evaluation unit. The landscape ecological risk value and the regional ecological risk value of all grid evaluation units in the target area are integrated respectively to generate grid layers, and the spatial distribution map of the landscape ecological risk value and the spatial distribution map of the regional ecological risk value are constructed based on a geographic information system.

9. The ecological risk assessment and prediction evaluation method based on the multi-source evaluation framework of claim 1, wherein: The comprehensive ecological risk value of each grid evaluation unit is obtained by weighted superposition of the landscape ecological risk value and the regional ecological risk value, specifically, the landscape ecological risk value and the regional ecological risk value are multiplied by their corresponding preset weights respectively, and the two weighted results are added to obtain the comprehensive ecological risk value of the grid evaluation unit; wherein the weight of the regional ecological risk value is greater than the weight of the landscape ecological risk value, and the sum of the two weights is one; The comprehensive ecological risk values of all grid evaluation units are clustered and graded by using a natural breakpoint method. Specifically, the comprehensive ecological risk values of all grid evaluation units in the target region are arranged in ascending order to form an ordered series. Based on the ordered series, an iteration optimization calculation is performed to find the first natural breakpoint that minimizes the intra-group variance and maximizes the inter-group variance The grid evaluation units with a comprehensive ecological risk value not greater than the first natural breakpoint are classified into one risk level, the grid evaluation units with a risk value greater than the first and not greater than the second natural breakpoint are classified into one risk level, and so on. The comprehensive ecological risk values are divided into a continuous level, which is used to generate a comprehensive ecological risk grading map.

10. The ecological risk assessment and prediction evaluation method based on the multi-source evaluation framework of claim 9, wherein: After the comprehensive ecological risk classification map is generated, the specific method for determining the risk warning time is as follows: according to the comprehensive ecological risk value of each grid evaluation unit and the risk grade to which it belongs, based on the change trend of the comprehensive ecological risk value of the grid evaluation unit in the past multiple evaluation periods, the linear trend extrapolation method is used to predict the length of time required for the comprehensive ecological risk value to reach the preset high risk threshold, and the length of time is calibrated as the risk warning time of the grid evaluation unit; finally, the risk warning time of each grid evaluation unit is associated with its spatial position to form a comprehensive ecological risk spatio-temporal distribution map with time sequence attribute.

Images