Bayesian-based ecosystem service spatial trade-off collaborative optimization method and device

Abstract

This invention provides a Bayesian-based method and apparatus for spatial trade-off and synergistic optimization of ecosystem services, relating to the fields of ecology and statistics. The invention acquires an ecological spatial database of a target area, uses ArcGIS software's netting and clipping tools to generate a grid layer of a specified size for the target area, obtaining a pixel set. Based on the ecological spatial database and pixel set, the blue carbon storage, biodiversity maintenance capacity, and coastal protection capacity of the target area are calculated, representing three ecosystem services. Furthermore, the coupling coordination degree is obtained to analyze the overall coordination level among ecosystem services. A Bayesian-based ecosystem service synergy prediction model is constructed using Netica software to identify key driving factors and output the probability distribution of corresponding nodes, simulating and analyzing potential synergistic impacts under different scenarios. This invention can accurately identify and clarify the trade-off and synergistic relationships of key ecosystem services such as coastal wetlands, achieving synergistic spatial optimization of ecosystem services.

Description

Technical Field

[0001] This invention relates to the fields of ecology and statistics, and more specifically, to a Bayesian-based method and apparatus for collaborative optimization of spatial trade-offs in ecosystem services. Background Technology

[0002] Coastal wetlands, as vital natural resources for humankind, provide crucial ecosystem services such as blue carbon storage, biodiversity maintenance, and coastal protection. Exploring the trade-offs and synergies among these key ecosystem services in coastal wetlands and revealing the spatial distribution of synergistic benefits is crucial for optimizing resource allocation and improving conservation levels. However, current research lacks a national-scale map of key ecosystem services and their trade-offs in China's coastal wetlands. More importantly, the synergistic potential of key ecosystem services has not been fully explored, hindering the accurate identification of hotspots for ecosystem service synergies and trade-offs in China's coastal wetlands. This presents challenges to the precise protection and restoration of coastal wetlands. Furthermore, considering the complex interactions between ecosystem and social systems, complex models need to be developed to predict the synergistic potential of ecosystem services by coupling natural, socio-economic, and ecological factors.

[0003] Bayesian networks, as an important decision support tool, based on expert experience and quantitative analysis, can integrate multiple information sources, including ecological processes and socio-economic activities of ecosystem services, and their complex causal relationships. They can effectively infer the probability of future events and allow for graphical representation, thus helping to simulate the synergistic potential among ecosystem services. Therefore, mapping the spatial distribution of key ecosystem services in China's coastal wetlands—blue carbon storage, biodiversity conservation, and coastal protection—clarifies the spatial trade-offs and synergistic relationships of ecosystem services, explores the spatial pattern and driving forces of ecosystem service synergy in coastal wetlands, simulates the spatial distribution of potential synergistic areas, and optimizes the spatial pattern of ecosystem service synergy. This has significant theoretical and practical implications for the successful implementation of ecological protection and restoration of China's coastal wetlands and the sustainable management of coastal wetland resources.

[0004] In view of this, the applicant hereby submits this application after studying the existing technology. Summary of the Invention

[0005] This invention aims to provide a Bayesian-based method and apparatus for the coordinated optimization of spatial trade-offs in ecosystem services, in order to address the shortcomings of existing methods, such as their inability to accurately identify hotspots for the coordination and trade-offs of ecosystem services in coastal wetlands, and to achieve coordinated optimization of key ecosystem services.

[0006] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution:

[0007] A Bayesian-based method for spatial trade-off optimization of ecosystem services includes:

[0008] S1, Obtain the ecological spatial database of the target area, which includes a basic geographic dataset, a blue carbon storage dataset, a species spatial dataset, a climate and environment dataset, and a socio-economic dataset.

[0009] S2, Based on the ecological space database, the fishing net tool and clipping tool of ArcGIS software are used to generate a grid layer of a specified size for the target area, thereby obtaining the cell set of the target area;

[0010] S3. Based on the blue carbon storage dataset and the pixel set of the target area, the total biomass of the target area, i.e., the blue carbon storage, is calculated by combining the aboveground biomass and the underground biomass.

[0011] S4, based on the species spatial dataset and the pixel set of the target area, each pixel of the target area is regarded as a spatial unit, and the species richness and rarity range of the target area are obtained. By overlaying the species richness and rarity range in the layer space, the biodiversity maintenance capacity of the target area is obtained; wherein, the species richness is the total number of species distributed in each spatial unit; the rarity range is obtained by calculating the weighted sum of the proportions of the species distributed in the spatial unit to the global area.

[0012] S5. Based on the climate and environmental dataset and the socio-economic dataset, the exposure index under different scenarios is estimated using the InVEST model. The coastal protection capability of the target area is evaluated by the difference between the relative exposure indices under the two different scenarios.

[0013] S6. Based on the blue carbon storage, biodiversity maintenance capacity, and coastal protection capacity, i.e., the three ecosystem services, the coupling coordination degree is obtained to assess and analyze the overall coordination level among ecosystem services.

[0014] S7. Based on the ecological space database, exploratory variables affecting the ecosystem services are selected to obtain natural and social factors; then, the natural and social factors of the target area pixels are calculated using ArcGIS and zoning statistical tools.

[0015] S8. A Bayesian network-based ecosystem service collaborative prediction model is constructed using Netica software. The natural factors, social factors, ecosystem services, and coupling coordination degree are discretized using the natural breakpoint method and used as input variable nodes for the collaborative prediction model. The probability distribution of the ecosystem service and coupling coordination degree nodes is output.

[0016] S9. Based on the probability distribution of the ecosystem services and the coupling coordination nodes, the impact of natural and social driving factors on ecosystem services and their coordination is assessed through sensitivity analysis. Key driving factors for ecosystem service coordination and optimal combinations of environmental variables under different coordination scenarios are identified, and potential coordination areas of ecosystem services under different scenarios are simulated.

[0017] This invention also provides a Bayesian-based device for co-optimizing spatial trade-offs in ecosystem services, comprising:

[0018] The database acquisition unit is used to acquire the ecological spatial database of the target area, which includes a basic geographic dataset, a blue carbon storage dataset, a species spatial dataset, a climate and environment dataset, and a socio-economic dataset.

[0019] The pixel set generation unit is used to generate a grid layer of a specified size for the target area based on the ecological space database using the fishing net tool and clipping tool of ArcGIS software, thereby obtaining the pixel set of the target area;

[0020] The blue carbon storage calculation unit is used to calculate the total biomass of the target area, i.e., the blue carbon storage, based on the blue carbon storage dataset and the pixel set of the target area, combined with the aboveground biomass and the underground biomass.

[0021] A biodiversity maintenance capacity calculation unit is used to obtain the biodiversity maintenance capacity of a target area. Based on the species spatial dataset and the pixel set of the target area, each pixel of the target area is considered as a spatial unit. The species richness and rarity range of the target area are obtained. The biodiversity maintenance capacity of the target area is obtained by overlaying the species richness and rarity range in a layer space. The species richness is the total number of species distributed in each spatial unit. The rarity range is obtained by calculating the weighted sum of the proportions of the species distributed in the spatial unit to the global area.

[0022] The coastal protection capacity calculation unit is used to estimate the exposure index under different scenarios based on the climate environment dataset and socio-economic dataset using the InVEST model, and evaluate the coastal protection capacity of the target area by the difference between the relative exposure indices under two different scenarios.

[0023] The coupling coordination degree calculation unit is used to obtain the coupling coordination degree based on the blue carbon storage, biodiversity maintenance capacity and coastal protection capacity, i.e., the three ecosystem services, in order to assess and analyze the overall coordination level among ecosystem services.

[0024] The driving factor acquisition unit is used to screen out exploratory variables that affect the ecosystem services based on the ecological space database, and obtain natural and social factors; then, it uses ArcGIS and zoning statistics tools to calculate the natural and social factors of the target area pixels.

[0025] The Bayesian model building unit is used to construct a Bayesian-based ecosystem service co-prediction model. The model is constructed using Netica software based on a Bayesian network. The natural factors, social factors, ecosystem services, and coupling coordination degree are discretized using the natural breakpoint method and used as input variable nodes for the co-prediction model. The output is the probability distribution of the ecosystem service and coupling coordination degree nodes.

[0026] The analysis unit is used to assess the impact of natural and social driving factors on ecosystem services and their synergy through sensitivity analysis based on the probability distribution of the ecosystem services and the coupling coordination nodes, identify key driving factors of ecosystem service synergy and the optimal combination of environmental variables under different synergy scenarios, and simulate potential synergy areas of ecosystem services under different scenarios.

[0027] In summary, compared with the prior art, the present invention has the following beneficial effects:

[0028] This invention integrates existing ecological spatial databases and utilizes biodiversity species richness, rarity range indicators, and the InVEST model to obtain blue carbon storage, biodiversity maintenance capacity, and coastal protection capacity of target areas, in order to analyze the spatial distribution characteristics of ecosystem services.

[0029] Then, based on spatial autocorrelation analysis, self-organizing mapping network and coupling coordination degree analysis of the trade-off and synergy of ecosystem services, we elucidate the spatial clustering phenomenon between pairs of ecosystem services and the spatial combination characteristics and overall coordination level of the three ecosystem services.

[0030] Furthermore, by integrating natural and social factors, ecosystem services, and coupling coordination, a Bayesian network collaborative prediction model for ecosystem services is constructed to identify key driving factors affecting collaboration. By setting scenarios with different collaboration objectives, the spatial distribution of potential collaboration zones is identified, so as to comprehensively consider the coordination and potential collaboration of ecosystem services and achieve spatial optimization of ecosystem service collaboration.

[0031] The method of this invention can reveal the spatial distribution of key ecosystem services such as those in my country's coastal wetlands, clarify the trade-offs and synergies among ecosystem services, delineate the impact mechanisms of ecosystem service synergy, and simulate the distribution of potential synergy zones, thereby achieving spatial optimization of ecosystem service synergy. This will help provide policymakers with spatial management information for ecosystem construction and restoration, offer scientific references for the spatial management of my country's coastal wetlands, and promote national and local fulfillment of global climate change adaptation and biodiversity commitments. Attached Figure Description

[0032] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0033] Figure 1 This is a schematic diagram of a Bayesian-based collaborative optimization method for spatial trade-offs in ecosystem services, provided in Embodiment 1 of the present invention.

[0034] Figure 2 The flowchart illustrates a Bayesian-based collaborative optimization method for spatial trade-offs in ecosystem services, as provided in Embodiment 1 of the present invention.

[0035] Figure 3 A schematic diagram of a Bayesian network model for serving the national coastal wetland ecosystem, provided in Embodiment 1 of the present invention.

[0036] Figure 4 This is a schematic diagram of bivariate spatial autocorrelation analysis of pairwise ecosystem services of coastal wetlands in my country, provided in Embodiment 1 of the present invention.

[0037] Figure 5 This is a schematic diagram of a Bayesian-based ecosystem service spatial trade-off collaborative optimization device provided in Embodiment 2 of the present invention.

[0038] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0040] Example 1

[0041] Embodiment 1 of the present invention provides a Bayesian-based method for spatial trade-off optimization of ecosystem services, which can be implemented by a Bayesian-based device for spatial trade-off optimization of ecosystem services (hereinafter referred to as the optimization device), specifically, executed by one or more processors within the optimization device.

[0042] In this embodiment, the collaborative optimization computing device may be an electronic device equipped with a processor, the processor having a computer program for the Bayesian-based ecosystem service spatial trade-off collaborative optimization method, and the computer program being executable, such as a computer, smartphone, or smart tablet.

[0043] In recent years, exploring the trade-offs and synergies of ecosystem services at the national and regional scales has become an important research area. Trade-offs refer to the inverse interaction of ecosystem services, where an increase in one service leads to a decrease in others; synergies, on the other hand, manifest as a synchronous effect where ecosystem services increase (or decrease) together. Trade-off and synergistic phenomena among ecosystem services are widespread. For example, effective nutrient retention in wetlands may accelerate vegetation succession, leading to changes in species richness and a decrease in biodiversity; the denitrification capacity and water conservation potential of wetlands show a positive correlation and synergistic relationship.

[0044] Spatial mapping and statistical analysis are commonly used methods for assessing trade-offs and synergies among ecosystem services. Spatial mapping helps to represent the distribution characteristics and combination types of various ecosystem services in space, including spatial autocorrelation, ecosystem service clustering, and spatial overlay methods. Spatial autocorrelation analysis uses the magnitude of the Moran's index (reflecting positive or negative correlation) and LiSA clustering plots (reflecting high and low clustering characteristics) to measure the spatial trade-offs or synergies between pairs of ecosystem services.

[0045] Ecosystem service clusters are spatially defined using cluster analysis and principal component analysis to classify combinations of multiple ecosystem services. The co-occurrence of high and low values ​​of different ecosystem services within a cluster reflects trade-offs or synergies among services. Coastal wetlands, as important natural resources for humankind, provide critical ecosystem services such as blue carbon storage, biodiversity maintenance, and coastal protection. The dual pressures of climate change and human activities pose severe challenges to the sustainable supply and coordinated development of critical ecosystem services, necessitating strengthened protection and restoration of coastal ecosystems. Maximizing the protection of synergistically beneficial coastal wetlands is crucial for the sustainable management of wetland resources.

[0046] like Figures 1-2 As shown, a Bayesian-based method for spatial trade-off optimization of ecosystem services includes steps S1 to S9.

[0047] S1, Obtain the ecological spatial database of the target area, which includes a basic geographic dataset, a blue carbon storage dataset, a species spatial dataset, a climate and environment dataset, and a socio-economic dataset.

[0048] The coastal wetland spatial distribution data used in this embodiment, i.e., the ecological spatial database, comes from the study of Zhang et al. (2022), which extracted the spatial distribution of mangroves and salt marshes in China at a 10m resolution using Sentinel 2 satellite imagery from 2020. This invention can also use datasets from other wetlands for calculations, which will not be elaborated upon here.

[0049] Basic geographic data, including vector data of administrative units (provinces, cities, counties, and special administrative regions), were obtained from the National Bureau of Surveying and Mapping. Data on mangrove soil and biomass carbon storage, i.e., the blue carbon storage dataset, came from the Global MangroveWatch (GMW) dataset (https: / / www.globalmangrovewatch.org / ); the dataset on salt marsh soil carbon storage came from the study by Fan and Li (2024). Vector data on mangrove and salt marsh species distribution information, i.e., the species spatial dataset, came from the IUCN Red List of Threatened Species (https: / / www.iucnredlist.org / ).

[0050] Climate and environmental data include annual average temperature, annual average precipitation, minimum sea surface temperature, sea surface salinity, tidal range, elevation (DEM), and slope. Annual average temperature and annual average precipitation are from the China Meteorological Data Service Center. Minimum sea surface temperature is from the NASA multi-sensor ultra-high resolution daily sea surface temperature dataset. National coastal salinity data is from the GOFS 3.1 reanalysis dataset (https: / / www.hycom.org / dataserver / gofs-3pt1 / reanalysis), and Kriging spatial interpolation was used to obtain raster data of salinity in the nearshore waters of China. Tidal range data is from the National Marine Science Data Center, which collected hourly tide level data from 202 tide gauge stations along my country's coastline from 2020 to 2022 to estimate the average tidal range, and Kriging spatial interpolation was used to obtain the national coastal tidal range data. Slope data was calculated from DEM data in ArcGIS software; the DEM data is from the NASA DEM dataset.

[0051] Socioeconomic data included national GDP, population density, nighttime light data, national nature reserve boundaries, electricity consumption, and data on land reclamation and nearshore mariculture. GDP, nighttime light data, and global nature reserve boundaries were obtained from the Resource and Environmental Science and Data Center of the Chinese Academy of Sciences. Population density data was obtained from Worldpop (https: / / hub.worldpop.org / geodata / ). National electricity consumption raster data came from the study by Chen et al. (2022), with a spatial resolution of 1 km. National land reclamation and coastline geomorphology data came from the study by Sajjad et al. (2018), which used visual interpretation to map the spatial distribution of land reclamation in my country from 1995 to 2015 and identify three types of coastlines: artificial, natural, and tidal flats. Nearshore mariculture vector data came from the National Earth System Science Data Center. After obtaining the relevant data, they were uniformly converted to the WGS1984UTM Zone 49N projection coordinate system.

[0052] S2, Based on the ecological space database, the fishing net tool and clipping tool of ArcGIS software are used to generate a grid layer of a specified size for the target area, thus obtaining the pixel set of the target area.

[0053] In this embodiment, ArcGIS software is a series of Geographic Information System (GIS) software developed and maintained by Esri Corporation, designed to help users edit, use, and manage geographic information.

[0054] S3. Based on the blue carbon storage dataset and the pixel set of the target area, the total biomass of the target area, i.e., the blue carbon storage, is calculated by combining the aboveground biomass and the underground biomass.

[0055] In this embodiment, blue carbon storage refers to organic carbon captured and stored from the atmosphere by coastal and marine ecosystems, which can effectively mitigate and adapt to climate change.

[0056] Therefore, this embodiment quantifies the spatial distribution of blue carbon storage services in coastal wetlands of my country by integrating reliable and available blue carbon storage datasets from mangroves and salt marshes. Blue carbon storage in Chinese mangroves is calculated by spatially overlaying soil and biomass raster data. Aboveground biomass data is validated using lidar, a 30m resolution DEM, and ground observations to invert global mangrove canopy height, and then the aboveground biomass is estimated using the allometric growth equation. Belowground biomass is estimated using the allometric index and combined with aboveground biomass to calculate total biomass, as shown in the following formula:

[0057] c i =α×(b iabove +b ibe低值 );b ibe低值 =β×b iavove ;

[0058] Among them, c i For the total biogenic carbon storage of pixel i, b iabove b is the aboveground biomass of pixel i; ibe低值 Let α be the underground biomass of pixel i; α and β are preset weighting coefficients; based on experience, α is set to 0.5 and β to 0.38.

[0059] Considering that soil organic carbon is the main component of carbon sinks in salt marshes, and that spatial distribution data on salt marsh biomass in my country is lacking, soil carbon storage data was used to characterize the distribution of blue carbon storage in my country's salt marsh wetlands. Blue carbon storage maps of mangroves and salt marshes were created by spatially overlaying layers using ArcGIS software, resulting in a blue carbon storage map of my country's coastal wetlands.

[0060] S4. Based on the species spatial dataset and the pixel set of the target area, each pixel of the target area is regarded as a spatial unit, and the species richness and rarity range of the target area are obtained. By overlaying the species richness and rarity range in the layer space, the biodiversity maintenance capacity of the target area is obtained. The species richness is the total number of species distributed in each spatial unit. The rarity range is obtained by calculating the weighted sum of the proportions of the species distributed in the spatial unit to the global area.

[0061] Preferably, the formula for calculating the rarity range is:

[0062]

[0063] Among them, Rarity iLet A be the rarity range index of pixel i, N be the total number of species in the target area, j be a single species, and A be the rarity range index of pixel i. j Let S be the total area of ​​the j-th species in the target region. ij Let S be the area of ​​the j-th species distributed in pixel i. i Let be the area of ​​pixel i.

[0064] This embodiment obtains spatial vector data of animal, plant, fungal, and algal species in the coastal wetland ecosystem based on IUCN Red List data, and uses the cropping tool of ArcGIS software to obtain the spatial distribution information of species in the study area.

[0065] Biodiversity maintenance maps are constructed by spatially overlaying species richness and rarity range layers. These two indicators of biodiversity—species richness and rarity range—characterize the biodiversity maintenance capacity of coastal wetlands. These two indicators are considered important and complementary, and are widely used to reflect biodiversity maintenance capacity, serving as direct and effective key indicators for decision-making in ecological conservation.

[0066] Species richness represents the total number of species distributed within each spatial cell. Rarity range is calculated by weighting the proportion of species distributed within a spatial cell to the global area, reflecting the size of the species distribution range and the importance of their clustering.

[0067] S5. Based on the climate and environmental dataset and the socio-economic dataset, the exposure index under different scenarios is estimated using the InVEST model. The coastal protection capability of the target area is evaluated by the difference between the relative exposure indices under the two different scenarios.

[0068] This embodiment simulates two scenarios: one with and without mangroves, and the other with salt marsh habitats. It uses the reduction in coastal exposure caused by floods and storm surges to characterize the strength of coastal protection functions. This embodiment utilizes the coastal vulnerability module of the InVEST model to estimate the exposure index under different scenarios. This module couples seven biophysical vectors or raster maps—topography, terrain, habitat, net sea-level change, wind exposure, wave exposure, and potential beundles of giant waves—with a coastal population density raster map. It assesses the vulnerability levels in terms of topography, surface relief, natural habitat, wave exposure, and wind exposure (as shown in Table 1), as well as the degree of population risk from natural disasters, and estimates the relative exposure index of coastal areas to erosion or inundation.

[0069] In this embodiment, the InVEST model, short for Integrated Valuation of Ecosystem Services and Trade-offs, aims to provide decision-makers with a scientific basis for weighing the benefits and impacts of human activities by simulating changes in the quantity and value of ecosystem services under different land cover scenarios.

[0070] Preferably, the formula for calculating the exposure index is:

[0071]

[0072] Where CVI is the exposure index, n is the total number of coastal vulnerability indicators in the InVEST model, and R k This represents the vulnerability level corresponding to the k-th indicator.

[0073] The input data required for the coastal vulnerability model and their sources include: DEM, slope, population density, coastal wetland and geomorphological data; wave energy, maximum tidal range, maximum wind speed and significant wave height data are all from the third-generation wave numerical prediction model (WAVEWATCH III); water depth measurement and continental shelf vector data are from the InVEST model dataset. After inputting the above data into the InVEST coastal vulnerability model, the relative exposure index of my country's coastal areas under scenarios with and without mangroves and salt marsh habitats is obtained. The difference between the relative exposure indices of the two scenarios is used to represent the coastal protection capacity of coastal wetlands in my country's coastal zone.

[0074] Table 1. Calculation Method of Indices in the InVEST Coastal Vulnerability Model

[0075]

[0076] S6. Based on the aforementioned blue carbon storage, biodiversity maintenance capacity, and coastal protection capacity, i.e., the three ecosystem services, the coupling coordination degree is obtained to assess and analyze the overall coordination level among ecosystem services.

[0077] To further explore the spatial trade-offs and synergies among the three services, this embodiment uses the bivariate spatial autocorrelation model of Geoda software to perform spatial autocorrelation analysis on the ecosystem services, identifying the spatial high and low clustering characteristics of each pair of services; combined with self-organizing map neural network to identify ecosystem service clusters, and based on the coupling coordination degree model to quantify the overall synergy level of the three ecosystem services, it explores the typical spatial clustering characteristics and synergy status of the three ecosystem services.

[0078] (1) Spatial autocorrelation analysis

[0079] In this embodiment, spatial autocorrelation analysis is applicable to analyzing whether variables exhibit clustering in spatial distribution, including global spatial autocorrelation and local spatial autocorrelation.

[0080] Global spatial autocorrelation analysis can reflect the degree of correlation and overall distribution characteristics between each spatial unit and its neighboring spatial units within the target study area. It is represented by the global autocorrelation index Moran's I index. When the value is greater than 0, it means that the two variables are positively correlated in terms of spatial distribution. When the value is less than 0, it means that the two variables are negatively correlated in terms of spatial distribution.

[0081] Local spatial autocorrelation analysis is used to identify spatial association patterns and significance levels between variables in local areas. The local autocorrelation index, Local Moran's I, is used to grasp the spatial clustering and spatial differentiation between local elements. Under the significance test (P<0.05), LiSA cluster maps (Local Indicators of Spatial Association) are plotted to reflect the clustering phenomena of "high-high clustering", "low-low clustering", "high-low clustering" or "low-high clustering" of two variables in local areas. This is used to explore the synergistic relationship or trade-off relationship between variables in spatial terms, where they mutually enhance or diminish each other.

[0082] The formulas for calculating the two key indices are as follows:

[0083]

[0084] Where Moran's I represents the global autocorrelation index, Moran's I∈[-1,1]; S 2 Let X be the variance, L be the total number of pixels, and X be the variance. i and X j These are the attribute values ​​for pixel i and pixel j in the target region, respectively. k is the average attribute of a pixel. ij The spatial weight matrix is ​​established for k-nearest neighbor spatial relationships; Local Moran's I is the local autocorrelation index.

[0085] Because the complex relationships of trade-offs and synergies are difficult to fully reveal at large spatial scales, this study analyzes the correlations between ecosystem services at the grid scale. Considering the spatial extent of the study area and the fragmented distribution of coastal wetlands, the study area was divided into a 2km*2km grid. Using ArcGIS's net tool, a grid layer of a specified size was generated and intersected with the my country coastal wetland map layer using the clipping tool, ultimately yielding 24,846 wetland pixels (2km*2km) for the study area. Next, based on ArcGIS's zoning statistical tools, the average values ​​of the three ecosystem services for each wetland pixel were calculated and standardized. Finally, Geoda software's local bivariate spatial autocorrelation analysis was used to spatially visualize which areas exhibit trade-offs, synergies, or insignificant spatial relationships between paired ecosystem services.

[0086] (2) Ecosystem service cluster

[0087] When assessing the overall coordination level among ecosystem services, ecosystem service clusters are used to analyze the synergistic or trade-off relationships among ecosystem services in the target area. An ecosystem service cluster refers to the co-occurrence and aggregation of multiple ecosystem services in space or time. The combination characteristics of multiple ecosystem services within a cluster, exhibiting a trade-off or mutual benefit, represent trade-offs and synergistic relationships. Self-organizing maps (SOM), as an unsupervised learning neural network, can be used to identify spatially formed ecosystem service clusters and visualize the results, facilitating the analysis of the trade-off and synergistic relationships among the three ecosystem services in the study area from the perspective of service clusters. First, the pixels of the target area are taken as the research unit and considered as a spatial combination unit (i.e., a "cluster"). The kohonen package in R (a module in self-organizing map neural networks) is used to identify the ecosystem service clusters in the target area. After multiple experiments, when the number of ecosystem service clusters reaches five preset values ​​(other preset values ​​can also be used according to actual needs), the spatial combination characteristics of ecosystem services are more reasonable and have higher interpretability. Therefore, they are divided into five categories of ecosystem service clusters and spatially displayed in ArcGIS software.

[0088] (3) Coupling coordination degree

[0089] The coupling coordination degree model is an effective method for examining the synergistic development among multiple interacting functions. It describes whether ecosystem services are mutually synergistic at a high level or mutually restrictive at a low level using the coupling coordination degree index (D). Therefore, this study introduces this model to identify the overall coordination level of each wetland pixel among blue carbon storage, biodiversity maintenance, and coastal protection. High coupling coordination degree values ​​represent high-value synergy among the three services, as detailed in Table 2.

[0090] The formula for the coupling coordination degree is as follows:

[0091]

[0092] T = aU1 + bU2 + cU3;

[0093] Where D represents the coupling coordination degree; C represents the coupling degree, indicating the degree of interaction between various ecosystem services; T represents the multifunctionality level of ecosystem services; U1, U2, and U3 represent blue carbon storage, biodiversity maintenance capacity, and coastal protection capacity, respectively, i.e., ecosystem services; and a, b, and c represent the corresponding weighting coefficients of U1, U2, and U3, respectively.

[0094] Table 2. Types and characteristics of the coupling coordination degree of three ecosystem services

[0095]

[0096] S7. Based on the ecological space database, exploratory variables affecting the ecosystem services are selected to obtain natural and social factors; then, the natural and social factors of the target area pixels are calculated using ArcGIS and zoning statistical tools.

[0097] Changes in the ecological functions of coastal wetlands are the result of multiple driving factors. In terms of natural factors, rising temperatures typically enhance the carbon storage capacity of coastal wetland soils and significantly impact species abundance, productivity, and vegetation growth. In terms of social factors, high-intensity human activities such as economic development, land reclamation, and aquaculture are key factors leading to the encroachment and degradation of coastal wetlands. This study identified exploratory variables influencing three ecosystem services, including seven natural factors (annual mean temperature, annual mean precipitation, annual mean minimum sea surface temperature, seawater salinity, DEM, slope, and tidal range) and five social factors (GDP, electricity consumption, distance from reclamation, distance from nearshore mariculture, and location within a protected area). A Bayesian network-based collaborative prediction model for coastal wetland ecosystem services was constructed, consisting of three layers: driving factors (natural factors + social factors) and two target variables: the three ecosystem services and their coupling coordination degree.

[0098] S8. A Bayesian network-based ecosystem service collaborative prediction model is constructed using Netica software. The natural factors, social factors, ecosystem services, and coupling coordination degree are discretized using the natural breakpoint method and used as input variable nodes for the collaborative prediction model. The probability distribution of the ecosystem services and coupling coordination degree nodes is output.

[0099] In this embodiment, Netica software is a professional Bayesian network analysis tool used to build, analyze, and optimize Bayesian network models.

[0100] Bayesian networks are graphical networks based on probabilistic reasoning. They consist of two parts: a causal network, which is a directed acyclic graph that connects the causal relationships between variables with arrows; and a conditional probability table that quantifies the relationships and strengths between variables in the network, providing the conditional probability of each possible state of the target node.

[0101] The original formula is: P(A|B)=P(A,B) / P(B)=P(B|A)P(A) / P(A);

[0102] In the formula: P(A|B) and P(B|A) are both conditional probabilities (or posterior probabilities), representing the probability of event A occurring when event B occurs or the probability of event B occurring when event A occurs, respectively; P(A,B) represents the probability of events A and B occurring simultaneously, i.e., the joint probability; P(A) and P(B) are prior probabilities (or marginal probabilities), representing the probability of event A or B occurring.

[0103] Applying a Bayesian network to this embodiment, the optimization formula is:

[0104]

[0105] Where X1, X2, ..., X m Let P(X1, X2, ..., X...) represent the variable nodes in the directed acyclic graph of the Bayesian network, where m is the total number of variable nodes. m X is the joint probability obtained by multiplying the local conditional probability distributions. i ∈[X1,X m ], where is the i-th variable node, parents(X i () is event X i The parent node.

[0106] Then, Bayesian network model parameter learning is performed. Given the diversity of spatial scales and ranges of each dataset source, the data needs to be preprocessed to create a unified geographic database. This invention uses ArcGIS 10.2 to resample all driving factor raster layers to 30m, and uses zoning statistical tools to calculate the driving factor values ​​of wetland pixels in the study area. The input data for the Bayesian network requires each node to be discrete data. The natural breakpoint method is used to discretize the data of each variable node, mainly dividing it into three states: low, medium, and high. The results of the state classification for each variable are shown in Table 3.

[0107] Due to missing social and natural factor data and resampling failures in some coastal areas, wetland pixel loss occurred. These NoData wetland pixels were removed, resulting in a final sample size of 20,586 inputs into the Bayesian network model. Using SPSS software, the total sample points were randomly divided into a training set (70%) and a test set (30%). The training set was imported into the Bayesian network for learning, thereby obtaining the probability distribution of the nodes.

[0108] Table 3 Node states in the Bayesian network model

[0109]

[0110]

[0111] S9. Based on the probability distribution of the ecosystem services and the coupling coordination nodes, the impact of natural and social driving factors on ecosystem services and their coordination is assessed through sensitivity analysis. Key driving factors for ecosystem service coordination and optimal combinations of environmental variables under different coordination scenarios are identified, and potential coordination areas of ecosystem services under different scenarios are simulated.

[0112] In this embodiment, sensitivity analysis is used to assess the impact of natural and social driving factors on ecosystem services and their synergies, clarifying the key driving factors of ecosystem service synergy. Then, four scenarios with different synergy objectives are set up. Using the sensitivity analysis results of a Bayesian network model and conditional probability tables, the optimal combination of environmental variables for different synergy scenarios is identified. By capturing the key state combinations of the top two key factors corresponding to the highest probability occurrence of each ecosystem service's high value, these combinations are spatially visualized to simulate potential synergy areas of ecosystem services under different scenarios. Finally, the results are overlaid with the coupling coordination degree results to classify the priority levels of protection priority areas.

[0113] In another preferred embodiment, four cases were selected: Guangxi Zhuang Autonomous Region, Guangdong Province, Hong Kong Special Administrative Region, Macao Special Administrative Region (hereinafter referred to as Guangdong-Hong Kong-Macao), Zhejiang Province, and Jiangsu Province to construct a Bayesian network model of coastal wetland ecosystem services at the regional scale.

[0114] Bayesian networks of national coastal wetland ecosystem services, such as Figure 3 As shown, it includes 16 nodes (12 child nodes and 4 parent nodes) and 34 connecting lines to represent the correlation between variables. Based on this, all node data from four case studies—Guangxi Zhuang Autonomous Region, Guangdong-Hong Kong-Macao Greater Bay Area, Zhejiang Province, and Jiangsu Province—were extracted again for parameter learning to generate a regional-scale Bayesian network for coastal wetlands.

[0115] In this embodiment, a confusion matrix is ​​used to validate the Bayesian network model. Test sets (30% of the data) at the national and regional scales were randomly sampled using SPSS software. The test sample data for the whole country, Guangxi Zhuang Autonomous Region, Guangdong-Hong Kong-Macao Greater Bay Area, Zhejiang Province, and Jiangsu Province were 6169, 569, 443, 1368, and 496 respectively, and were used to validate the accuracy of each model. The error matrix was used to evaluate the model's prediction accuracy. Blue carbon storage, biodiversity maintenance, coastal protection, and coupling coordination were used as target nodes. As shown in Table 4, the accuracy rates for the national Bayesian model were 82%, 81%, 82%, and 72% respectively. The overall accuracy rates for the Guangxi Zhuang Autonomous Region, Guangdong-Hong Kong-Macao Greater Bay Area, Zhejiang Province, and Jiangsu Province were 88%, 80%, 85%, and 92% respectively. This indicates that the model has good predictive performance, the simulation results are relatively reliable, and it can meet the prediction requirements for the probability of target node states under different scenarios.

[0116] Table 4. Prediction accuracy of each target node in Bayesian networks nationwide and in case study regions.

[0117]

[0118]

[0119] Considering that stakeholders in different regions may have different public preferences for different ecosystem services of coastal wetlands, and that trade-offs between ecosystem services are unavoidable, this study set up four synergy scenarios that prioritize high-value synergy between two services and overall synergy among three services, as shown in Table 5. These scenarios are considered equally important and are used to simulate the spatial distribution of potential synergy zones of ecosystem services.

[0120] Table 5 Scenario settings for different collaborative objectives

[0121] scene describe BC_CS scenario Prioritize the highest probability of maintaining high values ​​for blue carbon reserves and biodiversity. BC_CP scenario Prioritize the highest probability of biodiversity maintenance and coastal protection CS_CP scenario Prioritize blue carbon reserves and the highest probability of coastal protection BC_CS_CP scenario Simultaneously considering the high probability of high values ​​occurring in blue carbon storage, biodiversity maintenance, and coastal protection.

[0122] Sensitivity analysis measures the influence of driving factors on the target variable and quantifies it through variance reduction percentage. A larger percentage reduction indicates a stronger influence of the driving factor on the target variable. This invention identifies important driving factors (i.e., key variables) for each ecosystem service based on variance reduction results. Simultaneously, the conditional probability table of a Bayesian network can capture all possible combinations of driving factor states under different probabilities of occurrence (low, medium, high) of ecosystem services. The state of the key variable corresponding to the highest probability of high-value occurrence of an ecosystem service is defined as the key state of the key variable. This invention captures the key state combination of the first two key factors corresponding to the highest probability of high-value occurrence of each ecosystem service, visualizes them spatially, and simulates potential synergistic regions of ecosystem services under different scenarios.

[0123] Based on the spatial distribution map of key ecosystem services in my country's coastal wetlands, this study quantifies the spatial distribution of blue carbon storage, biodiversity maintenance, and coastal protection ecosystem services in mangrove and salt marsh coastal wetlands across the country by integrating existing blue carbon storage databases, biodiversity models, and the InVEST model. To facilitate comparison of the three services and subsequent spatial cluster analysis, the services are standardized and divided into six levels from low to high using the natural breakpoint method.

[0124] The results show that the three ecosystem services of blue carbon storage, biodiversity maintenance, and coastal protection in my country's coastal wetlands exhibit similar spatial distribution patterns, namely a differentiation pattern of gradually decreasing from south to north. Spatially, high values ​​for blue carbon storage services are mainly concentrated in Guangxi Zhuang Autonomous Region, Hainan Province, and the Guangdong-Hong Kong-Macao Greater Bay Area (Zhanjiang City), while Shandong Province and Jiangsu Province have relatively lower values. Regarding biodiversity maintenance services, Hainan Province, the Guangdong-Hong Kong-Macao Greater Bay Area, Guangxi Zhuang Autonomous Region, and Fujian Province show high values, as these regions have abundant rainfall, which to some extent promotes species composition and richness. Low values ​​are mainly distributed in Liaoning Province, Hebei Province, and Shandong Province. High values ​​for coastal protection services are more significant in Guangxi Zhuang Autonomous Region, while the Guangdong-Hong Kong-Macao Greater Bay Area, Hainan Province, Jiangsu Province, and Shanghai Municipality have relatively higher values.

[0125] Bivariate spatial autocorrelation analysis revealed the pairwise spatial correlations among the three ecosystem services. The global spatial autocorrelation indices, Moran's I, were 0.241, 0.305, and 0.234, respectively, indicating positive spatial correlations between all services. The highest positive spatial correlation was found between biodiversity maintenance and coastal protection services. A bivariate local spatial autocorrelation (LISA) cluster map was constructed using Geoda (e.g., [image of a cluster map]). Figure 4 As shown in the figure, four spatial clustering patterns are characterized at significant levels (P<0.05) between pairs of ecosystem services, including spatial synergistic relationships of high-high clustering and low-low clustering, or spatial trade-off relationships of low-high clustering and high-low clustering.

[0126] The coupling coordination degree model is used to classify and analyze the coupling coordination level of the three services, and the clustering characteristics of ecosystem services under different coordination states are explored in combination with the dominant service cluster type in the region.

[0127] The key variables and key state subsets of different collaborative scenarios (i.e., Table 5) are spatially visualized to represent the spatial distribution of potential collaborative zones for coastal wetland ecosystem services in various regions.

[0128] The current level of coupling and coordination of coastal wetland ecosystem services and their future synergistic potential should be comprehensively considered to determine priority areas for ecological protection and ecological restoration, as well as their priority levels.

[0129] In summary, compared with the prior art, the present invention has the following beneficial effects:

[0130] The method of this invention can reveal the spatial distribution of key ecosystem services such as those in my country's coastal wetlands, clarify the trade-offs and synergies among ecosystem services, delineate the impact mechanisms of ecosystem service synergy, and simulate the distribution of potential synergy zones, thereby achieving spatial optimization of ecosystem service synergy. This will help provide policymakers with spatial management information for ecosystem construction and restoration, offer scientific references for the spatial management of my country's coastal wetlands, and promote the fulfillment of national and local commitments to global climate change adaptation and biodiversity.

[0131] Example 2

[0132] like Figure 5 As shown, the second embodiment of the present invention also provides a Bayesian-based ecosystem service spatial trade-off collaborative optimization device, comprising:

[0133] The database acquisition unit is used to acquire the ecological spatial database of the target area, which includes a basic geographic dataset, a blue carbon storage dataset, a species spatial dataset, a climate and environment dataset, and a socio-economic dataset.

[0134] The pixel set generation unit is used to generate a grid layer of a specified size for the target area based on the ecological space database using the fishing net tool and clipping tool of ArcGIS software, thereby obtaining the pixel set of the target area;

[0135] The blue carbon storage calculation unit is used to calculate the total biomass of the target area, i.e., the blue carbon storage, based on the blue carbon storage dataset and the pixel set of the target area, combined with the aboveground biomass and the underground biomass.

[0136] A biodiversity maintenance capacity calculation unit is used to obtain the biodiversity maintenance capacity of a target area. Based on the species spatial dataset and the pixel set of the target area, each pixel of the target area is considered as a spatial unit. The species richness and rarity range of the target area are obtained. The biodiversity maintenance capacity of the target area is obtained by overlaying the species richness and rarity range in a layer space. The species richness is the total number of species distributed in each spatial unit. The rarity range is obtained by calculating the weighted sum of the proportions of the species distributed in the spatial unit to the global area.

[0137] The coastal protection capacity calculation unit is used to estimate the exposure index under different scenarios based on the climate environment dataset and socio-economic dataset using the InVEST model, and evaluate the coastal protection capacity of the target area by the difference between the relative exposure indices under two different scenarios.

[0138] The coupling coordination degree calculation unit is used to obtain the coupling coordination degree based on the blue carbon storage, biodiversity maintenance capacity and coastal protection capacity, i.e., the three ecosystem services, in order to assess and analyze the overall coordination level among ecosystem services.

[0139] The driving factor acquisition unit is used to screen out exploratory variables that affect the ecosystem services based on the ecological space database, and obtain natural and social factors; then, it uses ArcGIS and zoning statistics tools to calculate the natural and social factors of the target area pixels.

[0140] The Bayesian model building unit is used to construct a Bayesian-based ecosystem service co-prediction model. The model is constructed using Netica software based on a Bayesian network. The natural factors, social factors, ecosystem services, and coupling coordination degree are discretized using the natural breakpoint method and used as input variable nodes for the co-prediction model. The output is the probability distribution of the ecosystem service and coupling coordination degree nodes.

[0141] The analysis unit is used to assess the impact of natural and social driving factors on ecosystem services and their synergy through sensitivity analysis based on the probability distribution of the ecosystem services and the coupling coordination nodes, identify key driving factors of ecosystem service synergy and the optimal combination of environmental variables under different synergy scenarios, and simulate potential synergy areas of ecosystem services under different scenarios.

[0142] Example 3

[0143] The third embodiment of the present invention also provides a Bayesian-based ecosystem service spatial trade-off collaborative optimization device, which includes a memory and a processor. The memory stores a computer program, which can be executed by the processor to realize the Bayesian-based ecosystem service spatial trade-off collaborative optimization method as described above.

[0144] Example 4

[0145] The fourth embodiment of the present invention also provides a computer-readable storage medium storing computer-readable instructions. When the computer-readable instructions are executed by the processor of the device where the computer-readable storage medium is located, they implement the above-described Bayesian-based ecosystem service space trade-off collaborative optimization method.

[0146] In the several embodiments provided in this invention, it should be understood that the disclosed apparatus and methods can also be implemented in other ways.

[0147] In the various embodiments of the present invention, the functional modules can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0148] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0149] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A Bayesian-based method for spatial trade-off and collaborative optimization of ecosystem services, characterized in that, include: S1, Obtain the ecological spatial database of the target area, which includes a basic geographic dataset, a blue carbon storage dataset, a species spatial dataset, a climate and environment dataset, and a socio-economic dataset. S2, Based on the ecological space database, the fishing net tool and clipping tool of ArcGIS software are used to generate a grid layer of a specified size for the target area, thereby obtaining the cell set of the target area; S3. Based on the blue carbon storage dataset and the pixel set of the target area, the total biomass of the target area, i.e., the blue carbon storage, is calculated by combining the aboveground biomass and the underground biomass. S4. Based on the species spatial dataset and the pixel set of the target area, each pixel of the target area is regarded as a spatial unit, and the species richness and rarity range of the target area are obtained. By overlaying the species richness and rarity range in the layer space, the biodiversity maintenance capacity of the target area is obtained. Wherein, the species richness is the total number of species distributed in each spatial unit; the rarity range is obtained by calculating the weighted sum of the proportions of the species distributed in the spatial unit to the global area. S5. Based on the climate and environmental dataset and the socio-economic dataset, the exposure index under different scenarios is estimated using the InVEST model. The coastal protection capability of the target area is evaluated by the difference between the relative exposure indices under the two different scenarios. S6. Based on the blue carbon storage, biodiversity maintenance capacity, and coastal protection capacity, i.e., the three ecosystem services, the coupling coordination degree is obtained to assess and analyze the overall coordination level among ecosystem services. S7. Based on the ecological space database, exploratory variables affecting the ecosystem services are selected to obtain natural and social factors; then, the natural and social factors of the target area pixels are calculated using ArcGIS and zoning statistical tools. S8. A Bayesian network-based ecosystem service collaborative prediction model is constructed using Netica software. The natural factors, social factors, ecosystem services, and coupling coordination degree are discretized using the natural breakpoint method and used as input variable nodes for the collaborative prediction model. The probability distribution of the ecosystem service and coupling coordination degree nodes is output. S9. Based on the probability distribution of the ecosystem services and the coupling coordination nodes, the impact of natural and social driving factors on ecosystem services and their coordination is assessed through sensitivity analysis. Key driving factors for ecosystem service coordination and optimal combinations of environmental variables under different coordination scenarios are identified, and potential coordination areas of ecosystem services under different scenarios are simulated.

2. The Bayesian-based collaborative optimization method for spatial trade-offs in ecosystem services according to claim 1, characterized in that... The formula for calculating the total biomass carbon storage is as follows: ； ； in, For pixels Total biomass carbon storage, For pixels Aboveground biomass; For pixels The underground biomass; , These are preset weighting coefficients; based on experience, It is 0.

5. It is 0.

38.

3. The Bayesian-based collaborative optimization method for spatial trade-offs in ecosystem services according to claim 2, characterized in that... The formula for calculating the rarity range is: ; in, For pixels Rarity range index, The total number of species in the target area. As a single species, For the first The total area of ​​each species in the target region For the first Each species in pixels The area of ​​distribution For pixels The area.

4. The Bayesian-based collaborative optimization method for spatial trade-offs in ecosystem services according to claim 3, characterized in that, The formula for calculating the exposure index is as follows: ； in, Here, n represents the exposure index, and n is the total number of coastal vulnerability indices in the InVEST model. For the first The vulnerability level corresponds to each indicator.

5. The Bayesian-based method for spatial trade-off optimization of ecosystem services according to claim 4, characterized in that, The formula for the coupling coordination degree is as follows: ； ； ； in, For coupling coordination degree; Coupling degree represents the degree of interaction between various ecosystem services; Multifunctionality of ecosystem services; , and These are blue carbon storage, biodiversity maintenance capacity, and coastal protection capacity, i.e., ecosystem services; , , They are respectively , and The corresponding weighting coefficients.

6. The Bayesian-based collaborative optimization method for spatial trade-offs in ecosystem services according to claim 5, characterized in that... The probability distribution calculation formula for the Bayesian network-based ecosystem service collaborative prediction model is as follows: ； in, represents each variable node in the directed acyclic graph of the Bayesian network, where m is the total number of variable nodes. The joint probability is obtained by multiplying the local conditional probability distributions. For the i-th variable node, For the event The parent node.

7. A Bayesian-based collaborative optimization method for spatial trade-offs in ecosystem services according to claim 6, characterized in that... When assessing the overall coordination level among ecosystem services, Geoda software is used to perform spatial autocorrelation analysis on the ecosystem services. The spatial autocorrelation analysis includes global spatial autocorrelation analysis and local spatial autocorrelation analysis to analyze whether the variables exhibit clustering in spatial distribution.

8. The Bayesian-based collaborative optimization method for spatial trade-offs in ecosystem services according to claim 7, characterized in that... The global spatial autocorrelation analysis reflects the degree of correlation and overall distribution characteristics between each spatial unit and its neighboring spatial units within the target area. It is represented by a global autocorrelation index. When the global autocorrelation index is greater than 0, it indicates that the two variables are generally positively correlated in spatial distribution; when the global autocorrelation index is less than 0, it indicates that the two variables are generally negatively correlated in spatial distribution. The calculation formula is as follows: ； ； in, Represents the global autocorrelation index. ; Let L be the variance and L be the total number of pixels. and Target area pixels and pixels The attribute value, The average attribute value of a pixel. The spatial weight matrix is ​​established for k-nearest neighbor spatial relationships; The local autocorrelation index represents the spatial clustering and spatial differentiation among elements in a local area. A LiSA clustering map is plotted under significance testing to analyze the spatial synergistic trade-offs among variables in the target region. The formula for calculating the local autocorrelation index is as follows: ； in, This is a local autocorrelation index.

9. A Bayesian-based method for spatial trade-off optimization of ecosystem services according to claim 1, characterized in that... When assessing and analyzing the overall coordination level among ecosystem services, ecosystem service cluster analysis is used to analyze the synergistic or trade-off relationships among ecosystem services in the target area. An ecosystem service cluster is defined as the spatial or temporal aggregation of multiple ecosystem services. The combination characteristics of the inverse or mutually beneficial relationships among the multiple ecosystem services within a cluster represent trade-offs and synergistic relationships, respectively. Specifically: The pixels of the target area are taken as the research unit and regarded as a spatial combination unit; The self-organizing map neural network was used to identify ecosystem service clusters in the target area; When the number of identified ecosystem service clusters reaches a preset number, the identified ecosystem service clusters will be spatially displayed in ArcGIS software to analyze the synergistic or trade-off relationships presented in space.

10. A Bayesian-based collaborative optimization device for spatial trade-offs in ecosystem services, characterized in that, include: The database acquisition unit is used to acquire the ecological spatial database of the target area, which includes a basic geographic dataset, a blue carbon storage dataset, a species spatial dataset, a climate and environment dataset, and a socio-economic dataset. The pixel set generation unit is used to generate a grid layer of a specified size for the target area based on the ecological space database using the fishing net tool and clipping tool of ArcGIS software, thereby obtaining the pixel set of the target area; The blue carbon storage calculation unit is used to calculate the total biomass of the target area, i.e., the blue carbon storage, based on the blue carbon storage dataset and the pixel set of the target area, combined with the aboveground biomass and the underground biomass. A biodiversity maintenance capacity calculation unit is used to obtain the biodiversity maintenance capacity of a target area. Based on the species spatial dataset and the pixel set of the target area, each pixel of the target area is regarded as a spatial unit, and the species richness and rarity range of the target area are obtained. The biodiversity maintenance capacity of the target area is obtained by overlaying the species richness and rarity range in a layer space. The species richness is the total number of species distributed in each spatial unit; the rarity range is obtained by calculating the weighted sum of the proportions of the species distributed in the spatial unit to the global area. The coastal protection capacity calculation unit is used to estimate the exposure index under different scenarios based on the climate environment dataset and socio-economic dataset using the InVEST model, and evaluate the coastal protection capacity of the target area by the difference between the relative exposure indices under two different scenarios. The coupling coordination degree calculation unit is used to obtain the coupling coordination degree based on the blue carbon storage, biodiversity maintenance capacity and coastal protection capacity, i.e., the three ecosystem services, in order to assess and analyze the overall coordination level among ecosystem services. The driving factor acquisition unit is used to screen out exploratory variables that affect the ecosystem services based on the ecological space database, and obtain natural and social factors; then, it uses ArcGIS and zoning statistics tools to calculate the natural and social factors of the target area pixels. The Bayesian model building unit is used to construct a Bayesian-based ecosystem service co-prediction model. The model is constructed using Netica software based on a Bayesian network. The natural factors, social factors, ecosystem services, and coupling coordination degree are discretized using the natural breakpoint method and used as input variable nodes for the co-prediction model. The output is the probability distribution of the ecosystem service and coupling coordination degree nodes. The analysis unit is used to assess the impact of natural and social driving factors on ecosystem services and their synergy through sensitivity analysis based on the probability distribution of the ecosystem services and the coupling coordination nodes, identify key driving factors of ecosystem service synergy and the optimal combination of environmental variables under different synergy scenarios, and simulate potential synergy areas of ecosystem services under different scenarios.

Images