A high spatial resolution optimization method for synergistically enhancing mountain biodiversity and water conservation

Abstract

This invention discloses a high-spatial-resolution optimization method for synergizing biodiversity and water conservation in mountainous areas. This method, which belongs to the field of ecological remote sensing, includes the following steps: obtaining a biodiversity index and a water conservation function index for a target area based on remote sensing image data; obtaining driving factors, constructing a spatial localization relationship model based on the biodiversity index, water conservation function index, and driving factors, and identifying the primary controlling factors for each sub-area based on the relationship model; determining the optimal threshold interval for each primary controlling factor based on its response characteristics; and classifying global spatial units based on the optimal threshold interval for each primary controlling factor to generate a global synergistic optimization solution. This invention uses remote sensing technology to rapidly obtain high-resolution regional biodiversity and water conservation information, overcoming the time-consuming and labor-intensive shortcomings of traditional methods.

Description

Technical Field

[0001] The present invention belongs to the field of ecological remote sensing technology, and in particular relates to a high spatial resolution optimization method for synergistically enhancing mountain biodiversity and water conservation. Background Art

[0002] Biodiversity and water conservation are two of the most prominent aspects of mountain ecosystem services. Biodiversity, encompassing genetic diversity, species diversity, and ecosystem diversity, is fundamental to maintaining ecosystem function. Water conservation is a crucial ecosystem regulating service, primarily manifesting in precipitation interception, runoff regulation, and water purification. Trade-offs and synergies often exist between biodiversity and water conservation: 1) Trade-offs occur when an increase in one ecosystem service leads to a decrease in another; 2) Synergies occur when two ecosystem services increase or decrease simultaneously. Generally speaking, natural ecosystems characterized by high biodiversity exhibit relatively high levels of water conservation and regulation services. However, in certain specific circumstances, trade-offs may exist between the two. For example, in arid regions, increased species diversity and vegetation cover can accelerate ground-to-air exchange processes such as evapotranspiration, thereby reducing ecosystem water conservation services. Currently, the trade-offs and synergies between ecosystem services are primarily assessed through spatial mapping and model simulation: 1) Spatial mapping involves spatial analysis by overlaying two ecosystem services, comparing the correlations between the two services at the same location, and ultimately identifying areas of trade-offs and synergies; 2) Model simulation establishes the coupling relationship between water conservation and biodiversity, analyzing the dynamics of these services by developing several development, construction, and ecological protection scenarios. Compared to simple spatial mapping, model simulation can quantitatively and dynamically track the spatiotemporal changes in ecosystem service functions, offering inherent advantages over long time series and large regional scales. Therefore, in ecosystem management and protection measures, exploring the trade-offs and synergies between biodiversity and water conservation, and establishing a spatial optimization model for their synergy, is a key scientific foundation for ensuring the overall function and long-term stability of ecosystems.

[0003] Existing ecosystem service synergy prediction models consider the overall regional context and fail to account for the impact of spatial heterogeneity. Furthermore, the key drivers and optimal combination scenarios of existing ecosystem service synergy are oriented toward the entire target region, failing to achieve precise and quantitative optimization of local areas. Therefore, this paper proposes a high-spatial-resolution optimization method for synergizing mountain biodiversity and water conservation. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention proposes a high spatial resolution optimization method for synergistically enhancing mountain biodiversity and water conservation to solve the problems existing in the above-mentioned prior art.

[0005] To achieve the above objectives, the present invention provides a high spatial resolution optimization method for synergistically enhancing mountain biodiversity and water conservation, comprising:

[0006] Obtain the biodiversity index and water conservation function index of the target area based on remote sensing image data;

[0007] Obtain driving factors, construct a relationship model in a spatial local scenario based on the biodiversity index, water conservation function index and the driving factors, and identify the main controlling factors of each sub-region based on the relationship model;

[0008] Determining the optimal threshold interval of each master control factor based on the response characteristics of the master control factor;

[0009] Based on the optimal threshold interval of each main control factor, the global spatial units are classified to generate a global synergistic efficiency optimization plan.

[0010] Optionally, the process of obtaining the biodiversity index and water conservation function index of the target area includes:

[0011] Based on the principal component analysis dimensionality reduction of Landsat multi-band images, the spectral diversity index was calculated as the biodiversity index;

[0012] Based on the reflectivity of the near-infrared and short-wave infrared bands, the humidity index is calculated as the water conservation function index.

[0013] Optionally, the calculation expression of the biodiversity index is:

[0014]

[0015] The calculation expression of the water conservation function index is:

[0016]

[0017] Where SDI is the biodiversity index, K is the number of spectral clustering categories, P k is the ratio of spectral pixels of the kth category, LSWI is the water conservation function index, NIR is the reflectance of the near-infrared band, and SWIR is the reflectance of the short-wave infrared band.

[0018] Optionally, the process of obtaining driving factors and constructing a relationship model in a spatial local scenario based on the biodiversity index, the water conservation function index, and the driving factors includes:

[0019] Based on the driving factors, an ecological spatial database of the target area is constructed, and a grid layer of a specified size is generated using GIS software to obtain an influencing factor raster dataset; wherein the driving factors include: climate factors, topographic factors, hydrological factors, soil factors, vegetation type factors, and socioeconomic factors;

[0020] Constructing a binary diagram of the trade-off or synergistic relationship between biodiversity and water conservation based on the biodiversity index and the water conservation function index;

[0021] Performing layer overlay processing on the impact factor grid dataset and the trade-off or synergy relationship binary map to obtain study area grid data;

[0022] Based on the raster data of the study area, the sliding window method was used in combination with the random forest algorithm to construct a relationship model in a spatial local scenario.

[0023] Optionally, the process of constructing the binary graph of trade-off or synergy relationships includes:

[0024] Calculate the changes of the biodiversity index and the water conservation function index in two periods respectively by using the difference comparison method;

[0025] If the product of the two changes is positive, it is marked as a synergistic relationship; if it is negative, it is marked as a trade-off relationship;

[0026] Based on the synergistic relationship and the trade-off relationship, a binary graph of the trade-off or synergistic relationship covering the entire domain is generated.

[0027] Optionally, the process of constructing a relationship model in a local spatial scenario and identifying the main controlling factors of each sub-region based on the relationship model includes:

[0028] The sliding window method was used to divide the local sub-regions into the same number of grids as the study area;

[0029] Based on the plurality of local sub-regions, establishing a random forest model corresponding to the local sub-regions;

[0030] The importance of each influencing factor is quantified by accumulating the reduction of the Gini coefficient of each split node in the random forest model;

[0031] The influencing factor with the largest reduction in Gini coefficient is selected as the main controlling factor of the current sub-region.

[0032] Optionally, after identifying the main controlling factors of each sub-region based on the sliding window and the relationship model, a prediction model of the trade-off and synergy between biodiversity and water conservation under different climate, topography, hydrology, soil, vegetation type and socioeconomic scenarios is obtained;

[0033] Wherein, the prediction model is:

[0034]

[0035] Where y wi is the trade-off or synergy probability value under the predicted i-th window, f RF () is the prediction model, x wik is the value of the kth (k=1,2...k) driving factor in the i-th window.

[0036] Optionally, the process of determining the optimal threshold interval of each master control factor based on the response characteristics of the master control factor includes:

[0037] Analyze the response characteristics of the main controlling factors and the synergistic or trade-off relationships based on partial dependence functions;

[0038] Based on the response characteristics of the main control factors and the synergy or trade-off relationship, the value range of the main control factors is divided, and the response curve is generated by traversing the values and calculating the predicted mean;

[0039] Based on the distribution of the synergistic regions in the response curve, the optimal value range of the main control factor is determined.

[0040] Alternatively, the partial dependence function is estimated as:

[0041]

[0042] Where, is the estimate of the partial dependence function, x i,c (i=1,2,…,n) is the window w i x in the sample c The value of x t are t main controlling factors, is the random forest model of each window, and n is the number of pixels in the study area.

[0043] Optionally, the process of classifying the global spatial units based on the optimal threshold intervals of the main control factors includes:

[0044] Areas that meet the optimal thresholds of all main control factors are designated as key protection areas;

[0045] The areas that meet the optimal threshold of at least one main control factor are divided into priority optimization areas;

[0046] The remaining area is classified as a restricted use area.

[0047] Compared with the prior art, the present invention has the following advantages and technical effects:

[0048] This paper uses a sliding window approach to construct a nonlinear relationship model between the trade-offs and synergies between biodiversity and water conservation and their driving factors under various spatial localization scenarios, enabling the identification of key controlling factors across different regions. The spatial optimization of the synergistic relationship between biodiversity and water conservation achieved by this invention quantitatively reflects the optimal threshold combinations of key controlling factors in each region, and can be used to guide ecological and environmental management in a quantitative and spatially explicit manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of this application. The exemplary embodiments and descriptions of this application are intended to explain this application and do not constitute an improper limitation on this application. In the accompanying drawings:

[0050] Figure 1 This is a technical roadmap for an embodiment of the present invention. DETAILED DESCRIPTION

[0051] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0052] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0053] Example 1

[0054] The present invention addresses the problem of the lack of a spatial optimization model for the synergistic enhancement between biodiversity and water conservation, and considers the advantages of high-precision dynamic monitoring and identification of remote sensing and the influence of spatial heterogeneity of mountainous areas to propose a new solution. Through remote sensing technology and machine learning algorithms, a trade-off and synergistic relationship model between biodiversity and water conservation in different climate, topography, hydrology, soil, and vegetation type environments is constructed to identify the optimal combination scenario, serving the high-precision spatial optimization of the synergistic enhancement of biodiversity and water conservation. This method takes into account the environmental combination scenarios of climate, topography, hydrology, soil, and vegetation types in each region, and with the help of the high-precision identification function of remote sensing technology, it meets the high-precision spatial identification capability of the synergistic enhancement between biodiversity and water conservation. Overall, the present invention provides a more refined and quantitative spatial optimization method for the synergistic enhancement of biodiversity and water conservation, which can be used for research in the fields of terrestrial ecological environment monitoring and management, and is of great significance in improving the level of regional ecological environment management and decision-making.

[0055] In response to the problems of low degree of refinement and lack of quantitative optimization in spatial optimization technologies for synergistic efficiency between biodiversity and water conservation, this paper proposes a quantitative spatial optimization method for synergistic efficiency between biodiversity and water conservation at a refined remote sensing scale, providing advice and reference for regional ecological management. This method is based on a localized machine learning algorithm to explore the nonlinear relationship between environmental factors such as climate-topography-hydrology-soil-vegetation type and the trade-offs and synergistic binary layers between biodiversity and water conservation, and identify the main controlling factors of each region. On this basis, by further conducting sensitivity analysis in each region, the optimal threshold combination scenario of the main controlling factors in each region is obtained, and an optimization scheme for synergistic efficiency between biodiversity and water conservation based on the optimal threshold of local factors is realized.

[0056] like Figure 1 As shown, this embodiment provides a high spatial resolution optimization method for synergistically enhancing the biodiversity and water conservation in mountainous areas, comprising the following steps: obtaining the biodiversity index and water conservation function index of the target area based on remote sensing image data; obtaining driving factors, constructing a relationship model in a spatial local scenario based on the biodiversity index, water conservation function index and driving factors, and identifying the main controlling factors of each sub-area based on the relationship model; determining the optimal threshold interval of each main controlling factor based on the response characteristics of the main controlling factor; classifying the global spatial units based on the optimal threshold interval of each main controlling factor, and generating a global synergistic efficiency optimization plan. The specific implementation process is as follows:

[0057] (1) Obtain high-precision biodiversity and water conservation function indices based on remote sensing. Furthermore, the process of obtaining the biodiversity index and water conservation function index of the target area includes: calculating the spectral diversity index as the biodiversity index based on principal component analysis dimensionality reduction of Landsat multi-band images; and calculating the humidity index as the water conservation function index based on the reflectance of the near-infrared and short-wave infrared bands.

[0058] 1) Principal component analysis was performed on the Landsat multi-band images at time t1 and t2 respectively to reduce the dimensionality of the multi-band high-dimensional remote sensing data. The Shannon diversity index (SDI) was applied to the spectral spatial distribution after dimensionality reduction and defined as the spectral diversity index:

[0059] (1)

[0060] Among them, K is the number of spectral clustering categories, P k is the proportion of spectral pixels of the kth category. A larger SDI indicates a richer spectral type in the region, corresponding to higher biodiversity.

[0061] 2) Based on Landsat images, the LSWI at time t1 and t2 is obtained, which is also known as the water conservation function index. This index is sensitive to soil moisture and vegetation moisture content and is suitable for monitoring the water conservation function of the ecosystem. The calculation formula of LSWI is:

[0062] (2)

[0063] Among them, NIR represents the reflectivity of the near-infrared band, and SWIR represents the reflectivity of the short-wave infrared band.

[0064] (2) Construction of a relationship model between the trade-off and synergistic relationship between biodiversity and water conservation and the influencing factors. Furthermore, the process of constructing the relationship model in a spatial local scenario includes: constructing an ecological spatial database of the target area based on the driving factors, and using GIS software to generate a grid layer of a specified size to obtain an influencing factor raster dataset; constructing a binary map of the trade-off or synergistic relationship between biodiversity and water conservation based on the biodiversity index and the water conservation function index; overlaying the influencing factor raster dataset and the trade-off or synergistic relationship binary map to obtain the study area raster data; based on the study area raster data, using the sliding window method combined with the random forest algorithm to construct the relationship model in a spatial local scenario.

[0065] 1) Based on the definitions of synergy and trade-off, a difference-in-difference method was used to compare changes in biodiversity and water conservation between periods t1 and t2 to determine their relationship. If the product of the changes in the two services is positive, a synergistic relationship is considered; if the product is negative, a trade-off relationship is considered. Synergistic or trade-off relationships are represented using a binary graph. The specific calculation formula is:

[0066] (3)

[0067] (4)

[0068] (5)

[0069] Among them, SDI t1 and SDI t2 are the SDI values at t1 and t2, LSWI t1 and LSWI t2 are the LSWI values at t1 and t2, respectively; △A and △B are the differences between the two indices at t1 and t2.

[0070] Furthermore, the construction process of the binary diagram of the trade-off or synergistic relationship includes: calculating the changes in the biodiversity index and the water conservation function index in two periods respectively through the difference comparison method; if the product of the changes of the two is positive, it is marked as a synergistic relationship; if it is negative, it is marked as a trade-off relationship; based on the synergistic relationship and the trade-off relationship, a binary diagram of the trade-off or synergistic relationship covering the entire domain is generated.

[0071] 2) Select relevant influencing factors from the perspectives of climate, topography, hydrology, soils, vegetation type, and socioeconomics to construct an ecological spatial database for the target region. Use the fishing net and clipping tools in professional GIS software to generate a grid layer of a specified size, resulting in a raster dataset of influencing factors.

[0072] 3) Overlay the synergy or trade-off binary map with the impact factor raster dataset to construct a relationship model between the trade-off and synergy between biodiversity and water conservation and the impact factors. When constructing the model, the synergy or trade-off binary map is the dependent variable Y, and the impact factor raster data is the independent variable X={X1,X2,…,X k}.

[0073] 4) Considering the impact of spatial heterogeneity, a sliding window method combined with the random forest (RF) algorithm was used to construct a relationship model between the trade-offs and synergies between biodiversity and water conservation and driving factors under various spatial local scenarios, thereby achieving dynamic identification of the main controlling factors in local areas.

[0074] Furthermore, the process of identifying the main controlling factors of each sub-region based on the relationship model includes: using the sliding window method to divide the local sub-regions into the same number as the number of grids in the study area; establishing a random forest model corresponding to the local sub-regions based on several local sub-regions; quantifying the importance of each influencing factor by accumulating the reduction of the Gini coefficient of each split node in the random forest model; and selecting the influencing factor with the largest reduction in the Gini coefficient as the main controlling factor of the current sub-region.

[0075] The window side length is w, and the first row and first column of the grid dataset is the center of the window (x i ,y i ) starting position, a sliding window method was used to divide the study area into the same number of local subregions as the number of grid cells, and corresponding random forest models were established for each subregion. The importance of variables was assessed based on the Gini index of each subregion, thereby characterizing the spatial heterogeneity of factors affecting water conservation. This process combines the advantages of spatial sensitivity and model nonlinearity and includes the following steps:

[0076] First, let the study area be D, and use the sliding window side length w and step length s to uniformly move the window center (x i ,y i ), each window W iContains several spatially adjacent sample points. The sliding window is covered as follows:

[0077] (6)

[0078] Among them, r is the window radius, (x i ,y i ) are the coordinates of the window center.

[0079] Then, local samples are extracted within the window, feature variables are constructed and sorted, and the random forest model is trained based on the local sample data. i Extract the sample points contained in the window W i The synergy or trade-off relationship binary graph sample y and K influencing factor samples x K , where each impact factor sample x=(x1,x2,...,x k ). All impact factors need to be standardized:

[0080] (7)

[0081] Among them, μ k and σ k are the mean and standard deviation of the kth factor, respectively, to ensure that the dimensions of each variable are consistent.

[0082] The random forest model is trained based on the extracted and sorted local samples. The model expression is:

[0083] (8)

[0084] where Y i is the predicted value of the i-th window, f m (X) is the output of the mth decision tree. In the random forest algorithm, the Gini coefficient, as a core indicator for measuring node purity, is an important theoretical basis for ranking feature importance. By accumulating the reduction in the Gini coefficient brought by each feature at each split node during the decision tree construction process, it is possible to accurately rank the importance of each influencing factor and identify the main controlling factor. The Gini coefficient can be expressed as:

[0085] (9)

[0086] Among them, p k is the proportion of samples of the kth class in node t. The value range of the Gini coefficient is [0,1]. The smaller the Gini coefficient, the higher the node purity and the better the splitting effect.

[0087] Finally, the sliding window is continued to train the random forest model for the next window and identify the main controlling factors until the last grid. Ultimately, a prediction model and corresponding main controlling factors for the trade-off and synergy between biodiversity and water conservation under different climate, topography, hydrology, soil, vegetation type, and socioeconomic scenarios can be obtained. The established prediction model expression is:

[0088] (10)

[0089] Among them, y wi is the trade-off or synergy probability value under the predicted i-th window, f RF () is the prediction model, x wik is the value of the kth driving factor in the i-th window.

[0090] (3) Quantitative spatial optimization based on the synergistic effect of biodiversity and water conservation.

[0091] Furthermore, the process of determining the optimal threshold interval of each master control factor based on the response characteristics of the master control factor includes: analyzing the response characteristics of the master control factor and the synergistic or trade-off relationship based on the partial dependence function; dividing the value range of the master control factor based on the response characteristics of the master control factor and the synergistic or trade-off relationship, and generating a response curve by traversing the values and calculating the predicted mean; and determining the optimal value interval of the master control factor based on the distribution of the synergistic area in the response curve.

[0092] 1) The relationship model constructed under each window is identified as the relationship model of the central pixel of the window, and the main control factor of each window is identified as the main control factor of the central pixel of the window. In the prediction model of formula (10), the window w i The k impact factors are X wi ={X wi,1 ,X wi,2 ,…,X wi,k}, t main controlling factors are x t ={X wi,1 ,X wi,2 ,…,X wi,t}, taking the main controlling factor as the feature subset of interest, and the remaining features as x c ={X wi,t+1 ,X wi,t+2 ,…,X wi,k}, establish partial dependency function:

[0093] (11)

[0094] Among them, Ex C Indicates x C The probability distribution P(x C) to find the expectation. Formula (11) can be estimated from the training data by the following formula:

[0095] (12)

[0096] Among them, x i,c (i=1,2,…,n) is the window w i x in the sample c The model eliminates the influence of all other influencing factors by averaging. is the random forest model of each window, and n is the number of pixels in the study area.

[0097] The value range of each main controlling factor in the entire study area [x t,min ,x t,max ] is the basis for the t main control factors x t Set the starting value (x t,min )、step size (step = (x t,max -x t,min ) / 100) and the end value (x t,max ), the value of each main control factor is {x t,min ,x t,min +(x t,max -x t,min ) / 100, x t,min +2(x t,max -x t,min ) / 100,…, x t,max Each main control factor starts with the starting value, traverses all samples in the training set, and sets the remaining features x C Use the actual observation value, input the model to predict, and finally calculate the average of n prediction results. Repeat the above steps to cover each main control factor x t All values of , so that the sample can be recorded in the target feature change (x t ) when the model output changes, revealing the prediction model's effect on each main control factor (x t ) response characteristics.

[0098] 2) Based on the response characteristics of each master control factor obtained, a relationship curve is drawn. The y-axis and x-axis represent the predicted value of trade-off and synergy and the value of the master control factor respectively. The curve represents the nonlinear correlation between the predicted value and each master control factor. The smaller the value of the y-axis, the closer it is to synergy, and the larger the value, the closer it is to trade-off. The value range of the dependent variable is divided into three parts on average. The value of the ideal dependent variable is closer to the origin of the axis, so the part closer to the origin is the ideal part. The x-axis value interval corresponding to the y-value of the ideal part curve of the dependent variable is the optimal value interval of the master control factor. The x-axis value interval corresponding to the violent fluctuation curve segment is the sensitive interval of the explanatory variable. At this point, the window w is obtained. iThe center pixel (x i ,y i ) is the optimal value of each main controlling factor.

[0099] 3) Repeat the above method on each window and pixel to obtain the optimal value range of each main control factor on each grid pixel.

[0100] 4) Classify the global spatial units using the optimal thresholds of each main control factor. Let the optimal threshold of the j-th driving factor of grid unit i be T ij , the attribute of the jth master driving factor of grid cell i is x ij , the spatial unit can be partitioned according to the following criteria:

[0101] (13)

[0102] Among them, z ij Indicates the threshold satisfaction of grid cell i on the main control factor j. The threshold criteria of multiple driving factors are integrated to form a multi-factor spatial partition diagram. When multiple main control factors are combined, "and" and "or" logic are used:

[0103] (14)

[0104] (15)

[0105] The areas that meet formula (14) are classified as key protection areas, the areas that meet formula (15) are classified as priority optimization areas, and the other areas are restricted utilization areas.

[0106] As another implementation of this embodiment, a biodiversity index calculated based on other vegetation indices (such as NDVI) or an index that can reflect water conservation (such as the vegetation moisture index (NDMI)) can also be used in this process to further enhance the diversity and applicability of the method.

[0107] As another implementation of this embodiment, for biodiversity and water conservation data acquisition, the present invention uses Landsat imagery with a spatial resolution of 30 meters as the data source. However, other imagery with the same or higher spatial resolution (such as Sentinel-2) can also be used as an alternative to meet the needs of different application scenarios.

[0108] As another implementation of this embodiment, in constructing a relationship model of trade-offs and synergies between biodiversity and water conservation and driving factors under various spatial local scenarios, the present invention adopts a random forest algorithm, but other machine learning methods can also be used, such as gradient boosting tree (GBDT), etc., to adapt to different data characteristics and modeling requirements.

[0109] Compared with existing technologies, this method offers significant advantages. First, remote sensing technology allows for direct and rapid acquisition of biodiversity and water conservation data. Traditional field biodiversity and water conservation monitoring is limited by time and space and is typically expensive and time-consuming. Remote sensing can provide large-scale, detailed data in a consistent and objective manner. Second, spatial heterogeneity is evident across regions, particularly in mountainous areas. By constructing binary maps of the trade-offs and synergies between biodiversity and water conservation in local subregions, modeling their relationships with climate, topography, hydrology, soils, vegetation types, and socioeconomic factors, the ability to fit complex nonlinear regression relationships can be greatly enhanced, allowing precise identification of the key controlling factors within each subregion. This approach can better capture the local coupling characteristics of biodiversity and water conservation with ecological and environmental factors, enhancing the ability to identify spatial heterogeneity. Finally, by further analyzing the local coupling characteristics of the trade-offs and synergies between biodiversity and water conservation with key controlling factors, the optimal thresholds for the key controlling factors affecting the synergistic effects of biodiversity and water conservation in each subregion can be identified, providing a refined, quantitative, and targeted approach for large-scale regional spatial optimization.

[0110] The above are merely preferred embodiments of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A high spatial resolution optimization method for synergistically enhancing mountain biodiversity and water conservation, characterized by: The following steps are involved: Obtain the biodiversity index and water conservation function index of the target area based on remote sensing image data; Obtain driving factors, construct a relationship model in a spatial local scenario based on the biodiversity index, water conservation function index and the driving factors, and identify the main controlling factors of each sub-region based on the relationship model; The process of constructing a relationship model under a spatial local scenario includes: constructing an ecological spatial database of the target area based on the driving factors, and using GIS software to generate a grid layer of a specified size to obtain an influencing factor raster dataset; wherein the driving factors include: climate factors, topographic factors, hydrological factors, soil factors, vegetation type factors, and socioeconomic factors; constructing a binary map of the trade-off or synergistic relationship between biodiversity and water conservation based on the biodiversity index and the water conservation function index; overlaying the influencing factor raster dataset and the trade-off or synergistic relationship binary map to obtain raster data of the study area; and constructing a relationship model under a spatial local scenario based on the study area raster data using a sliding window method combined with a random forest algorithm. Determining the optimal threshold interval of each master control factor based on the response characteristics of the master control factor; Based on the optimal threshold interval of each main control factor, the global spatial units are classified to generate a global synergistic efficiency optimization plan.

2. The high spatial resolution optimization method for synergistically enhancing mountain biodiversity and water conservation according to claim 1, characterized in that: The process of obtaining the biodiversity index and water conservation function index of the target area includes: Based on the principal component analysis dimensionality reduction of Landsat multi-band images, the spectral diversity index was calculated as the biodiversity index; Based on the reflectivity of the near-infrared and short-wave infrared bands, the humidity index is calculated as the water conservation function index.

3. The high spatial resolution optimization method for synergistically enhancing mountain biodiversity and water conservation according to claim 2, characterized in that: The calculation expression of the biodiversity index is: The calculation expression of the water conservation function index is: Where SDI is the biodiversity index, K is the number of spectral clustering categories, P k is the ratio of spectral pixels of the kth category, LSWI is the water conservation function index, NIR is the reflectance of the near-infrared band, and SWIR is the reflectance of the short-wave infrared band.

4. The high spatial resolution optimization method for synergistically enhancing mountain biodiversity and water conservation according to claim 1, characterized in that: The process of constructing the trade-off or synergy relationship binary graph includes: Calculate the changes of the biodiversity index and the water conservation function index in two periods respectively by using the difference comparison method; If the product of the two changes is positive, it is marked as a synergistic relationship; if it is negative, it is marked as a trade-off relationship; Based on the synergistic relationship and the trade-off relationship, a binary graph of the trade-off or synergistic relationship covering the entire domain is generated.

5. The high spatial resolution optimization method for synergistically enhancing mountain biodiversity and water conservation according to claim 1, characterized in that: The process of constructing a relationship model in a local spatial scenario and identifying the main controlling factors of each sub-region based on the relationship model includes: The sliding window method was used to divide the local sub-regions into the same number of grids as the study area; Based on the plurality of local sub-regions, establishing a random forest model corresponding to the local sub-regions; The importance of each influencing factor is quantified by accumulating the reduction of the Gini coefficient of each split node in the random forest model; The influencing factor with the largest reduction in Gini coefficient is selected as the main controlling factor of the current sub-region.

6. The high spatial resolution optimization method for synergistically enhancing mountain biodiversity and water conservation according to claim 5, characterized in that: After identifying the main controlling factors of each sub-region based on the sliding window and the relationship model, a prediction model of the trade-off and synergy between biodiversity and water conservation under different climate, topography, hydrology, soil, vegetation type and socioeconomic scenarios was obtained; Wherein, the prediction model is: Where y wi is the trade-off or synergy probability value under the predicted i-th window, f RF () is the prediction model, x wik is the value of the kth (k=1,2...k) driving factor in the i-th window.

7. The high spatial resolution optimization method for synergistically enhancing mountain biodiversity and water conservation according to claim 1, characterized in that: The process of determining the optimal threshold interval of each master control factor based on the response characteristics of the master control factor includes: Analyze the response characteristics of the main controlling factors and the synergistic or trade-off relationships based on partial dependence functions; Based on the response characteristics of the main control factors and the synergy or trade-off relationship, the value range of the main control factors is divided, and the response curve is generated by traversing the values and calculating the predicted mean; Based on the distribution of the synergistic regions in the response curve, the optimal value range of the main control factor is determined.

8. The high spatial resolution optimization method for synergistically enhancing mountain biodiversity and water conservation according to claim 7, characterized in that: The estimation formula of the partial dependence function is: Where, is the estimate of the partial dependence function, x i,c (i=1,2,…,n) is the window w i x in the sample c The value of x t are t main controlling factors, is the random forest model of each window, and n is the number of pixels in the study area.

9. The high spatial resolution optimization method for synergistically enhancing mountain biodiversity and water conservation according to claim 1, characterized in that: The process of classifying global spatial units based on the optimal threshold intervals of each main control factor includes: Areas that meet the optimal thresholds of all main control factors are designated as key protection areas; The areas that meet the optimal threshold of at least one main control factor are divided into priority optimization areas; The remaining area is classified as a restricted use area.

Images