A regional ecosystem health assessment method based on remote sensing data

Abstract

The present application relates to the technical field of ecosystem health assessment, and particularly relates to a regional ecosystem health assessment method based on remote sensing data.The method comprises the following steps: acquiring multi-temporal remote sensing images of a target ecological region; performing band reflectivity processing on the multi-temporal remote sensing images to extract lake attribute parameters; identifying a hub lake set based on the lake attribute parameters, evaluating changes in lake area and vegetation coverage to quantify regional ecological degradation, and comprehensively evaluating a regional ecological health index in a current state; and regulating the ecological water level of the hub lake according to the regional ecological health index, and formulating an ecological water replenishment scheme for atrophied lakes to achieve water resource management of the lake group regional ecosystem.The present application realizes systematic assessment of the ecological health of the lake group by coupling lake group network topology analysis and water quantity spatial redistribution, so as to accurately identify key hub lakes and support differentiated water resource management.

Description

Technical Field

[0001] This invention relates to the field of ecosystem health assessment technology, and in particular to a method for regional ecosystem health assessment based on remote sensing data. Background Technology

[0002] Current methods for lake ecological assessment using remote sensing technology mainly focus on monitoring the physical, chemical, and biological parameters of individual lakes. However, lakes in nature, especially in plains or wetlands, are often not isolated geographical units, but rather form a complex network system—a lake cluster network—through various hydraulic connections such as surface runoff. In this network, changes in water quantity or quality at any lake node will be transmitted upstream and downstream through interconnected channels, triggering a cascading response throughout the network. However, traditional methods tend to treat each lake as an independent assessment object, calculating its health index separately and then performing simple spatial aggregation. This approach severs the inherent hydraulic and ecological connections between lakes. This isolated assessment method cannot identify the root causes of changes in the network structure, leading to serious misjudgments of the causes of lake changes and hindering the coordinated management of regional water resources. Summary of the Invention

[0003] Based on this, the present invention provides a method for regional ecosystem health assessment based on remote sensing data to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a method for regional ecosystem health assessment based on remote sensing data includes the following steps:

[0005] Step S1: Acquire multi-temporal remote sensing images of the target ecological area; perform band reflectance processing on the multi-temporal remote sensing images, extract the boundaries of each lake and calculate its geometric and ecological attributes to generate lake attribute parameters;

[0006] Step S2: Based on lake attribute parameters, determine the topographic and hydrological connectivity between lakes, construct a lake connectivity matrix, and identify the hub lake set based on the topological structure of the lake connectivity matrix; construct a lake network topology graph based on the lake connectivity matrix, which includes lake nodes and connecting edges;

[0007] Step S3: Based on the hub lake set and lake attribute parameters, assess the changes in lake area and vegetation cover to quantify regional ecological degradation, and comprehensively evaluate the current state of the regional ecological health index;

[0008] Step S4: Regulate the ecological water level of the key lakes according to the regional ecological health index, and formulate ecological water replenishment plans for shrinking lakes in order to achieve water resource management of the lake group regional ecosystem.

[0009] The beneficial effects of this invention are as follows:

[0010] First, based on lake attribute parameters, a lake connectivity matrix reflecting the actual hydraulic connections between lakes was accurately constructed through topographic profile analysis and overlay of river network vector data. Building upon this, a lake network topology map was further constructed, and network analysis algorithms such as betweenness centrality were used to identify pivotal lakes that play a decisive role in the overall network's hydrological connectivity and ecological stability. This method integrates previously discrete lake units into an organic network system for holistic evaluation, clearly tracing how changes in one lake affect other lakes through connectivity channels, thereby identifying the cascading effects caused by changes in network structure. This completely solves the serious misjudgment of the causes of lake changes caused by the traditional isolated assessment method, which severs the intrinsic connections between lakes.

[0011] This invention combines multi-temporal lake area change analysis to identify "shrinking lake clusters" (sources) and "expanding lake clusters" (sinks) within a region. Furthermore, it determines the direction of water flow based on the difference in water level between lakes and traces the water transfer path from shrinking lakes to expanding lakes on a network topology map, establishing a spatial pairing relationship between the two. This method not only statistically analyzes the total increase or decrease in water volume but, more importantly, reveals the spatial transfer pattern of water volume among lakes, clearly answering the core question of "which shrinking lake's water flowed to which expanding lake," thereby quantifying the balance of water redistribution. This solves the problem that existing technologies can only perform macro-level total statistics but cannot reveal the dynamic allocation mechanism of water resources within lake groups, providing precise data support for formulating cross-lake and cross-regional ecological water replenishment and water resource collaborative management schemes.

[0012] This invention efficiently integrates multi-source information such as network topology, hub node functions, spatiotemporal dynamics of water volume, and the degree of ecological degradation. It innovatively proposes a comprehensive assessment model encompassing four dimensions: network connectivity, hub ecosystem services, water redistribution balance, and ecological stability. Finally, a weighted summation yields a "Regional Ecological Health Index" that intuitively reflects the overall health of the lake cluster. This efficient data fusion and index construction method transforms complex remote sensing monitoring data and network analysis results into intuitive and quantitative assessments of the regional ecosystem's health status. Whether for the vast Jianghan Plain lake cluster or the ecologically fragile plateau lake region, this method can quickly and accurately output assessment conclusions, greatly improving the efficiency of ecological health management and providing timely and powerful scientific support for dynamically regulating the ecological water levels of hub lakes and optimizing ecological water replenishment schemes. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of the steps in the regional ecosystem health assessment method based on remote sensing data of the present invention;

[0014] Figure 2 Example diagram of vector labels in remote sensing images;

[0015] Figure 3 This is a comparison chart of lake area changes in this invention;

[0016] Figure 4 This is a schematic diagram of the lake network topology in this invention;

[0017] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present 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.

[0019] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0020] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0021] To achieve the above objectives, please refer to Figures 1 to 4 This invention provides a method for assessing the health of a regional ecosystem based on remote sensing data, comprising the following steps:

[0022] Step S1: Acquire multi-temporal remote sensing images of the target ecological area; perform band reflectance processing on the multi-temporal remote sensing images, extract the boundaries of each lake and calculate its geometric and ecological attributes to generate lake attribute parameters;

[0023] In this embodiment of the invention, a large plain area with a high density of lakes is selected as the target ecological region. High-quality satellite multispectral remote sensing images from autumn (the transition period between the wet and dry seasons) covering this region in two different years (e.g., the baseline year and the assessment year) are selected as the multi-temporal remote sensing image data source. First, using specialized remote sensing data processing software, all images are standardized for radiometric calibration and atmospheric correction, converting the original recorded values ​​into dimensionless surface reflectance image data. Based on this data, specific spectral bands are extracted, and exponential calculation and image segmentation techniques are applied to accurately extract the boundaries of each lake, and its key geometric and ecological attributes are further calculated. All attributes are ultimately integrated into a structured dataset, namely, lake attribute parameters.

[0024] In one implementation of this invention, to assess the ecological health of a large freshwater lake group, two periods of multispectral satellite imagery (such as Landsat or Sentinel series) covering the region during the wet and dry seasons are first acquired from a public data platform. The acquired images are radiometrically calibrated, converting the raw DN values ​​into top-level atmospheric radiance. Subsequently, an atmospheric correction model is used, inputting atmospheric parameters from the image metadata to eliminate the effects of atmospheric scattering and absorption, resulting in surface reflectance imagery data. Reflectance in the green and mid-infrared bands is extracted from the surface reflectance imagery data, and the improved normalized water index (MNDWI) image of the region is obtained through normalized difference calculation between bands. Then, an adaptive threshold segmentation algorithm is used to process the MNDWI imagery, marking pixels with index values ​​higher than a dynamic threshold (e.g., automatically calculated to be 0.25 based on the image histogram) as water bodies, forming a binarized water body mask. After morphological denoising and connected component analysis of this mask, the main connected component boundary corresponding to "Lake A" is raster vectorized to generate lake boundary vector data. Based on this vector data, GIS spatial analysis tools are used to calculate attributes, including lake area, lake centroid coordinates, lake water level elevation, and vegetation cover, which are used as lake attribute parameters.

[0025] Step S2: Based on lake attribute parameters, determine the topographic and hydrological connectivity between lakes, construct a lake connectivity matrix, and identify the hub lake set based on the topological structure of the lake connectivity matrix; construct a lake network topology graph based on the lake connectivity matrix, which includes lake nodes and connecting edges;

[0026] In the embodiments of this invention, please refer to Figure 2The image shows a specific remote sensing image of a lake, with white lines representing vector lines. For example, the system imports lake attribute parameters for all lakes within the target area (e.g., 20 major lakes in total), iterates through all lake pairs, and performs topographic and hydrological connectivity assessment on each pair. The first step is topographic connectivity assessment, analyzing whether there are any high grounds between the two lakes that could obstruct water flow based on DEM data. The second step is hydrological connectivity assessment; if the topography is connected, regional river network vector data is used to confirm whether there are physical waterways connecting them. The assessment results are used to fill an N×N adjacency matrix (N being the total number of lakes), i.e., the lake connectivity matrix. This matrix is ​​then treated as a network graph, calculating the network centrality index (such as betweenness centrality) for each lake node, and selecting nodes with centrality values ​​significantly higher than the average as hub lakes. Finally, a lake network topology graph containing all lake nodes and connected edges is generated.

[0027] Step S3: Based on the hub lake set and lake attribute parameters, assess the changes in lake area and vegetation cover to quantify regional ecological degradation, and comprehensively evaluate the current state of the regional ecological health index;

[0028] In this embodiment of the invention, lake attribute parameters from the baseline year and the assessment year are compared to quantify the area change of each lake, and the degree of ecological degradation is calculated in conjunction with the surrounding vegetation conditions. Then, sub-indices are calculated from four dimensions: network connectivity, hub ecosystem services, water redistribution balance, and ecological stability. Finally, these four sub-indices are weighted and summed to obtain a regional ecological health index between 0 and 1, which intuitively reflects the overall health status of the current lake group.

[0029] Step S4: Regulate the ecological water level of the key lakes according to the regional ecological health index, and formulate ecological water replenishment plans for shrinking lakes in order to achieve water resource management of the lake group regional ecosystem.

[0030] In this embodiment of the invention, the regional ecological health index calculated in the previous step is compared with a preset management level threshold to trigger a corresponding management plan. Management measures are mainly divided into two categories: macro-control of key lakes, aimed at maintaining the stability and connectivity of the entire network; and precise restoration of specific shrinking lakes, aimed at improving local ecological problems.

[0031] In one implementation of this invention, it is assumed that the calculated regional ecological health index is 0.455, and the pre-set health threshold for the regional ecological health index is 0.6. Since the current calculated value of 0.455 is lower than this threshold, the lake cluster ecosystem is determined to be in a sub-healthy state, requiring the initiation of water resource management and ecological intervention procedures. Based on the identified set of key lakes, the primary task of the management plan is to ensure the stability of this core node. For example, the specific measures formulated are: "to carry out refined scheduling of the upstream sluice gate controlling the inflow of water into Lake L002, ensuring that during the dry season, the water level of Lake L002 is not lower than the minimum ecological water level line of 17.5 meters, so as to maintain its core water conveyance function in the network." Analysis of the water transfer path of the shrinking lake L001 reveals that it has no direct or indirect hydraulic connection with any expanding lake in the network, indicating that its shrinkage is isolated. Therefore, the proposed ecological water replenishment plan is: "Initiate dredging and exploration of the water diversion channel connecting Lake L001 and its upstream water source river, and assess the feasibility of restoring its water replenishment channel through engineering means." Regarding another shrinking lake, L008, path analysis shows that it is connected to the expanding lake L004 via a connecting river, and its water level is higher than the latter. This indicates that its water loss is a result of internal network redistribution. The proposed plan is: "At the river mouth from Lake L008, study the feasibility of adding a seasonal ecological weir to appropriately raise the water level of L008 during the non-flood season, reducing its excessively rapid downstream flow into the network."

[0032] Preferably, step S1 involves performing band reflectance processing on the multi-temporal remote sensing images, extracting the boundaries of each lake, and calculating its geometric and ecological attributes, including:

[0033] Radiometric calibration and atmospheric correction are performed on multi-temporal remote sensing images to generate surface reflectance image data;

[0034] Extract near-infrared and green reflectance from surface reflectance image data;

[0035] The normalized remote sensing index image is obtained by dividing the difference between the reflectance of the green band and the reflectance of the near-infrared band by the sum of the two.

[0036] The normalized remote sensing index image is vectorized to obtain the lake boundary vector data.

[0037] In one implementation of this invention, radiometric calibration is performed on the acquired multispectral image. The purpose of this process is to convert the dimensionless raw values ​​(i.e., DN values) recorded by the sensor into spectral radiance at the top of the atmosphere with practical physical meaning. Specifically, radiometric calibration is first performed on a multispectral remote sensing image covering the target area. This process aims to eliminate systematic errors inherent in the sensor itself. By reading the gain and offset parameters from the image metadata file, a linear radiometric calibration formula is applied to convert the DN value of each pixel into its spectral radiance value at the top of the atmosphere. For example, a pixel with an original DN value of 18500, after radiometric calibration, corresponds to a spectral radiance value of 65.5 watts / (square meter·steradian·micrometer).

[0038] It should be noted that the radiance value obtained at this point still includes the effects of atmospheric scattering and absorption, and does not represent the true reflectance of the ground object. Therefore, atmospheric correction processing is required immediately. In this embodiment of the invention, a FLAASH atmospheric correction module based on a radiative transfer model is used. Input parameters include the solar altitude angle, azimuth angle, sensor height, and atmospheric mode (e.g., mid-latitude summer). This model simulates the process of electromagnetic waves passing through the atmosphere and derives the true reflectance of the ground object. After this processing, the pixel with the aforementioned spectral radiance value of 65.5 has its final surface reflectance corrected to 0.12, a dimensionless physical quantity between 0 and 1, representing that the ground object reflects 12% of the incident solar radiation.

[0039] In one implementation of this invention, for a clear water pixel in an image, its surface reflectance values ​​are: green band reflectance equal to 0.10, and near-infrared band reflectance equal to 0.04. According to the normalized remote sensing index calculation method, that is, based on the difference between the green band reflectance and the near-infrared band reflectance divided by their sum, the index value calculated for this water pixel is (0.10-0.04) / (0.10+0.04), which is approximately equal to 0.43.

[0040] In another implementation of this invention, for a healthy vegetation pixel in an image, its surface reflectance values ​​are: green light band reflectance equal to 0.08, and near-infrared band reflectance equal to 0.50. The calculated index value for this vegetation pixel is (0.08-0.50) / (0.08+0.50), which is approximately -0.72.

[0041] It is important to note that these two bands were chosen because clear water has high reflectivity in the green light band and extremely strong absorption characteristics in the near-infrared band, resulting in very low reflectivity. This significant difference in spectral characteristics allows the index calculated through normalized difference to greatly enhance the information about water bodies. Water pixels typically exhibit high positive values ​​(e.g., 0.43), while vegetation, soil, and other land features exhibit negative or low positive values.

[0042] In one implementation of this invention, a threshold segmentation is performed on the normalized remote sensing index image. Based on histogram statistical analysis of the index image, a water body identification threshold is set, which is a critical value used to distinguish water bodies from non-water bodies. In this invention, the threshold is set to 0.2. All pixels in the image are traversed, and pixels with an index value greater than 0.2 are marked as 1 (representing water bodies), while pixels with an index value less than or equal to 0.2 are marked as 0 (representing non-water bodies), thereby generating a binary water body mask data containing only pixel values ​​of 0 and 1.

[0043] Preferably, the vectorization process of the normalized remote sensing index image includes:

[0044] Pixels in the normalized remote sensing index image whose pixel value is greater than the preset water body identification threshold are marked as water body pixels, forming water body mask data.

[0045] Connectivity analysis was performed on the water body mask data to identify each connected component;

[0046] Remove connected components whose area is smaller than the preset minimum lake area, and mark the remaining connected components as independent lakes;

[0047] The boundaries of each independent lake are vectorized into lake boundary vector data.

[0048] In one implementation of this invention, a reasonable water body identification threshold is determined, which is obtained by performing image histogram analysis on a normalized remote sensing index (NRA) image. By observing the histogram, two distinct peaks are typically observed, corresponding to water and non-water pixels respectively, while the location of the troughs represents the ideal segmentation threshold. In this invention, the preset water body identification threshold obtained by this method is 0.2. Subsequently, each pixel in the NRA image is traversed, and its pixel value is compared with this threshold. For example, if a pixel has an NRA value of 0.45, since it is greater than 0.2, this pixel will be assigned a value of 1 in the newly generated image; while another pixel has an NRA value of -0.15, since it is less than 0.2, it will be assigned a value of 0.

[0049] Specifically, an eight-neighborhood connected component analysis algorithm is performed on the water mask data. The image is scanned line by line. When a cell with a value of 1 is encountered, its eight neighboring cells (top, bottom, left, right, and four diagonal directions) are checked. If neighboring cells also have a value of 1, they are grouped into the same connected component. This process is performed recursively or iteratively until all cells with a value of 1 within a connected component are found and marked. For example, a region in the upper left corner of the image consisting of thousands of cells with a value of 1 will be marked as connected component 1 after analysis; while a water region in the lower right corner of the image that is not connected to the former will be marked as connected component 2.

[0050] In one implementation of this invention, a minimum lake area threshold is set. This threshold is determined based on the geographical features of the study area and the research objective. For example, in this invention, considering that the spatial resolution of satellite imagery is 30 meters and the pixel area is 900 square meters, the preset minimum lake area is 0.1 square kilometers, which is approximately equal to 111 pixels. All connected components identified in the previous step are traversed, and the number of pixels contained in each connected component is counted. For example, connected component 1 contains 50,000 pixels, and its area is much larger than the threshold, so it is retained. However, connected component 3 contains only 50 pixels, and its area is smaller than the threshold, so this connected component will be removed from the labeled image, and its pixel value will be reset to 0 (non-water background).

[0051] Specifically, a boundary tracing algorithm is applied to the water body mask data after area filtering. This algorithm tracks the boundary line between pixels with a value of 1 and pixels with a value of 0 along the periphery of each independent lake patch and records the geographic coordinate sequence of the center of all boundary pixels.

[0052] In another implementation of this invention, the obtained discrete coordinate point sequence is smoothed to eliminate the inherent jaggedness of the raster data, and then constructed into a closed polygon. This closed polygon, defined by a series of geographic coordinate points, is the final generated lake boundary vector data.

[0053] Preferably, step S1, which involves processing the multi-temporal remote sensing images for band reflectance, extracting the boundaries of each lake, and calculating its geometric and ecological attributes, further includes:

[0054] The area of ​​each lake is obtained by calculating the area of ​​the closed region enclosed by the boundary vectors of each lake based on the lake boundary vector data.

[0055] Calculate the coordinates of the geometric center point in the lake boundary vector data, and use them as the centroid coordinates of each lake;

[0056] Obtain digital elevation model data of the target ecological area;

[0057] Based on the lake centroid coordinates, the shoreline elevation values ​​at the corresponding locations are extracted from the digital elevation model data and used as the lake water level elevations for each lake.

[0058] Based on the lake boundary vector data, an ecological buffer zone is constructed by extending a preset buffer width outward from the boundary.

[0059] Vegetation pixel values ​​were extracted from the normalized remote sensing index image within the ecological buffer zone, and the arithmetic mean of all vegetation pixel values ​​was calculated as the vegetation cover.

[0060] The lake area, lake centroid coordinates, vegetation coverage, and lake water level elevation of each lake are associated as lake attribute parameters.

[0061] In one implementation of this invention, the spatial geometry calculation engine built into a Geographic Information System (GIS) is invoked to calculate the area of ​​each polygonal object in the lake boundary vector data. This engine typically uses the shoelace formula or trapezoidal rule, employing a sequence of polygon vertex coordinates to calculate the area of ​​the enclosed two-dimensional space. For example, for a polygon representing "Lake A," its boundary consists of 1500 vertex coordinates; these coordinates are substituted into the algorithm for iterative calculation.

[0062] Specifically, a centroid calculation algorithm is performed on each lake boundary vector polygon. This algorithm calculates the average position of the polygon's geometry. For an irregular polygon, its centroid coordinates (centroid x-coordinate, centroid y-coordinate) are obtained by triangulating the polygon, calculating the centroid of each triangle, and then taking a weighted average based on their areas. For example, for the vector polygon of "Lake A", the calculated projected coordinates of its geometric center point are (x-coordinate: 458210, y-coordinate: 3385450). This coordinate point is recorded as the centroid coordinates of "Lake A".

[0063] In one implementation of this invention, digital elevation model (DEM) data covering the entire target ecological area is obtained. In this invention, the data comes from a publicly available dataset with a spatial resolution of 30 meters released by the Space Shuttle Radar Topography Mission (SRTM). The elevation is not extracted directly using the lake centroid coordinates because the centroid is located inside the water body, and its DEM value is inaccurate or empty.

[0064] It should be noted that the technical solution here is optimized to extract the average elevation along the shoreline, extract the coordinates of all vertices on the boundary vector line of "Lake A", and extract the corresponding surface elevation values ​​from the digital elevation model data by batch interpolation based on these coordinates.

[0065] Specifically, a buffer analysis operation is performed on each lake boundary vector polygon. First, a preset buffer width is set, determined based on the ecological characteristics and management needs of the study area; in this invention, it is set to 500 meters. Using the lake boundary line as a reference, it is shifted 500 meters outward (towards the land side), forming a new closed boundary line. The annular area enclosed by the original lake boundary and this new boundary line is constructed as the ecological buffer zone of that lake. For example, performing this operation on "Lake A" will generate a 500-meter-wide annular polygon, which is the ecological buffer zone of "Lake A".

[0066] In one implementation of this invention, a Normalized Difference Vegetation Index (NDVI) image reflecting vegetation density is generated. This image is obtained by extracting the near-infrared and red bands from surface reflectance data and calculating the difference between them using normalized values. It is important to note that "Normalized Difference Vegetation Index image" here specifically refers to an NDVI image. Subsequently, the polygon of the "Jiahu" ecological buffer zone constructed in the previous step is used as a spatial mask, and all pixels completely falling within the buffer zone are extracted from the NDVI image. For example, NDVI values ​​of 12,500 pixels are extracted, and the arithmetic mean of these 12,500 NDVI values ​​is calculated, yielding 0.72. This average value directly reflects the average vegetation density of the lakeshore zone and is therefore used as the vegetation cover of "Jiahu".

[0067] Preferably, in step S2, based on lake attribute parameters, the topographic and hydrological connectivity relationships between lakes are determined, a lake connectivity matrix is ​​constructed, and the set of hub lakes is identified based on the topological structure of the lake connectivity matrix, including:

[0068] Extract the centroid coordinates of any two lakes from the lake attribute parameters, and calculate the Euclidean distance between the two lake centroid coordinates as the lake spacing.

[0069] Based on the distance between lakes, terrain connectivity is determined. When two lakes are determined to be terrain-connected, they are marked as the first lake and the second lake.

[0070] Construct a topographic profile along the line connecting the centroid coordinates of the first and second lakes;

[0071] Extract the lake water level elevations corresponding to the first and second lakes from the lake attribute parameters, and take the maximum value of the two as the water level threshold.

[0072] Determine whether there are topographic points in the topographic profile with elevation values ​​higher than the water level threshold. If there are no topographic points with elevation values ​​higher than the water level threshold, determine that the two lakes have a topographic connectivity relationship and use them as candidate connected lake pairs. Obtain the river network vector data of the area between the candidate connected lake pairs.

[0073] River channels are identified based on river network vector data, and a lake connectivity matrix is ​​constructed based on the marked river channels for all candidate connected lake pairs.

[0074] In one implementation of this invention, records of two lakes are arbitrarily extracted iteratively. For example, "Lake A" and "Lake B" are extracted, and their centroid coordinates are read from the records respectively. Assuming the centroid coordinates of "Lake A" are (x-coordinate A: 458210, y-coordinate A: 3385450) and the centroid coordinates of "Lake B" are (x-coordinate B: 465330, y-coordinate B: 3390100), the Euclidean distance formula is then applied to calculate the straight-line distance between these two coordinate points on the two-dimensional plane.

[0075] In one specific implementation of this invention, a maximum search radius is preset, for example, 10 kilometers. Only when the calculated distance between lakes is less than this maximum search radius is a more complex subsequent terrain connectivity determination process initiated. For example, if the distance between Lake A and Lake B is 8950 meters (approximately 8.95 kilometers), which is less than the preset threshold of 10 kilometers, then this pair of lakes is determined to require further terrain profile analysis. Lake A and Lake B are temporarily labeled as the first and second lakes for subsequent connectivity determination. If the distance between another pair of lakes, Lake C and Lake D, is calculated to be 15 kilometers, exceeding the search radius, then they are directly determined to be not directly connected, and subsequent analysis is skipped.

[0076] In one implementation of this invention, the centroid coordinates of the first lake (“Lake A”) and the second lake (“Lake B”) are connected into a straight line segment in a geographic information system. Equidistant sampling is performed along this straight line segment at a preset sampling interval (e.g., equal to the spatial resolution of the digital elevation model data, i.e., 30 meters). At each sampling point, the surface elevation value of that point is extracted by interpolation from the digital elevation model (DEM) data.

[0077] In one implementation of this invention, the elevation values ​​of all 298 sampling points along the topographic profile are traversed, and each value is compared with a water level threshold of 25.8 meters. Upon inspection, it is found that the elevation values ​​of all points along the profile are between 24.5 meters and 25.6 meters, with no point exceeding 25.8 meters. Based on this result, it is determined that there are no topographic obstacles between the first and second lakes sufficient to block surface runoff, thus indicating a topographic connectivity between the two lakes. The pair "Lake A" and "Lake B" is marked as a candidate connected lake pair, triggering the next step: a rectangular search box is defined centered on the line connecting the centroids of the two lakes, and all river network vector data falling within this search box are extracted from a pre-set regional water system database.

[0078] Specifically, for the candidate connected lake pair "Lake A" and "Lake B", the river network vector data obtained in the previous step is analyzed. Through spatial topological relationship query, a river vector segment is found, whose starting point is inside or intersects with the boundary polygon of "Lake A", and whose ending point is inside or intersects with the boundary polygon of "Lake B". This identified river segment is marked as the river channel connecting "Lake A" and "Lake B".

[0079] It should be noted that all lake pairs will be traversed, and all combinations that are determined to be candidate connected lake pairs and successfully identified by river channels will be marked in the lake connectivity matrix. This matrix is ​​an N×N square matrix, where N is the total number of lakes. For "Lake A" (assuming its index is i) and "Lake B" (assuming its index is j), since there is a river channel between them, the element values ​​in the i-th row and j-th column and the j-th row and i-th column of the matrix are both set to 1. For lake pairs that are not connected by river channels, the corresponding matrix element values ​​are 0.

[0080] Preferably, identifying river channels based on river network vector data includes:

[0081] Based on the ecological buffer zone, candidate connected lake pairs are divided into a first lake buffer zone and a second lake buffer zone, respectively.

[0082] Extract each river vector segment from the river network vector data and determine whether each river vector segment intersects with both the first and second lake buffer zones simultaneously.

[0083] When there is a river vector segment that intersects both the first lake buffer zone and the second lake buffer zone, it is determined that the two lakes have a hydrological connectivity relationship, and the river vector segment is marked as a river channel.

[0084] Specifically, for a previously identified candidate connected lake pair, such as "Lake A" and "Lake B," their respective lake boundary vector data are retrieved from the lake attribute parameters. Then, buffer analysis is performed on these two polygons respectively. In this invention, the width of the ecological buffer zone is preset to 500 meters, a value determined based on a comprehensive consideration of regional hydrological characteristics and remote sensing image interpretation errors. A ring-shaped buffer zone for "Lake A" is generated by shifting 500 meters outward from the boundary line of "Lake A," and this area is designated as the first lake buffer zone. Similarly, the same operation is performed on "Lake B" to generate the second lake buffer zone.

[0085] In one implementation of this invention, all river network vector data located between "Lake A" and "Lake B" are extracted from a pre-set regional water system database. This data typically consists of a series of line segments (vector segments) representing the river centerlines. Spatial intersection is determined iteratively for each river vector segment. For example, a river vector segment numbered "River Segment 007" is extracted, and two independent geometric intersection tests are performed: first, it is determined whether the geometry of "River Segment 007" spatially overlaps or contacts the polygon of the first lake buffer zone (the buffer zone of "Lake A"); second, it is determined whether "River Segment 007" also spatially intersects with the second lake buffer zone (the buffer zone of "Lake B").

[0086] Specifically, the discrimination results of all river vector segments are summarized. Since the two intersection test results of "river segment 007" are both "yes", which satisfies the condition of "intersecting with both the first lake buffer zone and the second lake buffer zone", it is finally determined that there is a clear waterway connection between "lake A" and "lake B", that is, a hydrological connectivity relationship. A special field is added to the attribute table of "river segment 007" and its value is marked as "river channel". At the same time, the unique identifiers of the two lakes it connects to are recorded (for example, L-001 and L-002).

[0087] It is important to note that if, after analyzing all river vector segments between a candidate connected lake pair, no river is found that simultaneously satisfies the condition of intersecting with both buffer zones, then even if the lake pair is topographically connected, they are determined not to have a direct hydrological connectivity.

[0088] Preferably, step S2, which involves determining the topographic and hydrological connectivity between lakes based on lake attribute parameters, constructing a lake connectivity matrix, and identifying the hub lake set based on the topological structure of the lake connectivity matrix, further includes:

[0089] Calculate the shortest path between any two lake nodes based on the lake connectivity matrix, and count the number of nodes that pass through the lake nodes.

[0090] Calculate the total number of shortest paths between all pairs of lake nodes, and divide the node path count by the total number of shortest paths to obtain the betweenness centrality value of the lake nodes.

[0091] Arrange the betweenness centrality values ​​in descending order to form a centrality sorting sequence;

[0092] The upper quartile of the centrality sort sequence is calculated as the hub threshold, and lake nodes with betweenness centrality values ​​greater than the hub threshold are selected to obtain a preliminary hub lake set.

[0093] Prioritize the initial set of hub lakes and then filter the hub lake sets.

[0094] In one implementation of this invention, a breadth-first search (BFS) algorithm is applied to calculate the shortest path between any two lake nodes in the network. Here, a "path" refers to the minimum number of lake nodes required to travel from one lake to another. The algorithm iteratively calculates all shortest paths between all pairs of lake nodes in the network. During the calculation, a path count is maintained for each lake node. For example, when calculating a shortest path from "Lake A" to "Lake E" as "Lake A -> Lake B -> Lake E", Lake B, as an intermediate node, will have its path count incremented by 1.

[0095] In one implementation of this invention, the total number of all unique lake node pairs in the network is calculated. In a network containing N lake nodes, this total number can be calculated using a combination formula: N multiplied by (N-1) divided by 2. For example, in a network containing 20 lakes, the total number of shortest paths between all lake node pairs is 20 multiplied by 19 divided by 2, which equals 190. The path count of each lake node is then divided by this total number. For example, if the path count of "Lake A" is 90, its betweenness centrality value is 90 divided by 190, approximately 0.474. The path count of "Lake B" is 10, and its betweenness centrality value is 10 divided by 190, approximately 0.053.

[0096] In one implementation of this invention, the upper quartile (also called the 75th percentile) of the centrality-sorted sequence is calculated. The upper quartile refers to the value that is located at the 75th percentile when the dataset is arranged in ascending order. For example, for a sequence containing 20 centrality values, its upper quartile is approximately at position 15, and we assume that the value at this position is 0.180. This value of 0.180 is determined as the hub threshold. The betweenness centrality values ​​of all lake nodes are traversed, and all nodes greater than 0.180 are selected. In this example, the centrality values ​​of "Lake A" (0.474) and "Lake C" (0.211) are both greater than this threshold, so they are selected to form the initial hub lake set: {"Lake A", "Lake C"}.

[0097] Specifically, for each lake in the initial hub lake set, two ecosystem service function-related indicators—lake area and surrounding vegetation cover—will be extracted from the lake attribute parameters. Then, they will be prioritized by calculating a comprehensive score (e.g., the standardized value of area multiplied by a weight of 0.6, plus the standardized value of vegetation cover multiplied by a weight of 0.4). For example, Lake A, with both high area and high vegetation cover, will have a comprehensive score of 0.85; while Lake C, although strategically located, has a smaller area and a comprehensive score of 0.60.

[0098] Of particular importance is the prioritization of the initial set of hub lakes, and the selection of the hub lake set includes:

[0099] For each lake node in the initial hub lake set, temporarily remove the node and its connected edges from the lake network topology graph;

[0100] Perform connectivity analysis on the network after removing nodes, and count the number of independent connected subgraphs formed, which is taken as the subgraph connection count;

[0101] When the number of connections in the subgraph is greater than one, the lake node is marked as a critical node and assigned a high priority identifier.

[0102] When the number of connections in the subgraph is equal to one, the lake node is assigned a normal priority identifier.

[0103] The lake nodes in the initial hub lake set are sorted by priority identifier and betweenness centrality value to form a hub lake set.

[0104] Please see Figure 4 Node representation: Circular nodes (L1-L5): White circles represent process nodes or functional nodes; Black arrows: Indicate the direction of information flow or process advancement between nodes; Hub representation: Circular nodes (hub H1, hub H2): Black filled circles represent data hubs or functional hubs, used for centralized processing and forwarding of information; Connection relationships: Solid arrows: Indicate direct connections between nodes, between nodes and hubs, and between hubs, used for transmitting data or control commands; Assuming the set is {"Lake A", "Lake C"}, a simulated removal operation will be performed iteratively on each lake node in the set. First, "Lake A" is selected as the test object, and a temporary copy of the original lake network topology is created in memory. Then, in this copy, the node representing "Lake A" is found and deleted. Simultaneously, all connected edges directly connected to the "Lake A" node (e.g., the edge connecting "Lake A" to "Lake B", and the edge connecting "Lake A" to "Lake D") are queried and removed from the temporary copy.

[0105] It should be noted that this removal operation is temporary and non-destructive, used only in the current simulation analysis, and the original lake network topology remains unchanged. After completing the simulated removal of "Lake A", the same operation is performed on "Lake C", creating another independent temporary network copy and removing "Lake C" and its associated connected edges from it.

[0106] In one implementation of this invention, a connectivity analysis algorithm, such as Depth-First Search (DFS) or Breadth-First Search (BFS), is performed on a temporary network copy after the "Lake A" node has been removed. This algorithm starts traversing from any remaining node in the network, finding all nodes reachable from each other via connected edges, and marking them as the first connected subgraph. Then, the algorithm continues searching for new starting points among the unmarked nodes, repeating this process until all nodes are assigned to a connected subgraph. Finally, the total number of independent, unconnected subgraphs is counted. Assuming that after removing "Lake A," the originally complete network split into two independent regions, the counted number of independent connected subgraphs is 2. This value of 2 represents the number of connections in the subgraph recorded as "Lake A."

[0107] In another implementation of this invention, when analyzing the temporary network after removing "Binghu", it was found that all the remaining 19 lake nodes could still be connected to each other via paths, and the entire network did not split. In this case, the number of independent connected subgraphs is counted as 1, and this value of 1 is the subgraph connection number of "Binghu".

[0108] In one implementation of this invention, for "Lake Bing", its subgraph connection number is 1, which is equal to 1. This result shows that although "Lake Bing" plays an important role in many shortest paths (its betweenness centrality value is very high), its removal will not cause a break in the network structure, as there are other redundant paths in the network that can maintain overall connectivity. Therefore, "Lake Bing" is given a normal priority identifier, such as "priority: normal".

[0109] Preferably, step S3, which assesses changes in lake area and vegetation cover to quantify regional ecological degradation based on the set of hub lakes and lake attribute parameters, includes:

[0110] Obtain the lake area from multiple time phases in the lake attribute parameters and calculate the area change rate;

[0111] Lakes are classified into shrinking lakes and expanding lakes based on the sign of their area change rate.

[0112] Calculate the difference in vegetation cover at different time phases based on the vegetation cover in the lake attribute parameters;

[0113] For lakes with concentrated shrinkage, the absolute value of the area change is multiplied by the corresponding vegetation cover to obtain the ecologically degraded area.

[0114] The total ecological degradation area of ​​the region is obtained by summing up the ecological degradation areas of all the shrunken lakes.

[0115] In one implementation of this invention, the changing characteristics of shrinking and expanding lakes are displayed through time-relative comparison, accompanied by an explanation of the quantitative calculation formula for ecological degradation. Lake area data for each lake in two different years (i.e., the baseline year and the assessment year) is extracted from a previously generated lake attribute parameter data table. For example, for "Lake A," its lake area in the baseline year is extracted to be 50 square kilometers, and its lake area in the assessment year is 40 square kilometers. Its area change rate is calculated by subtracting the baseline year area from the assessment year area, and then dividing by the baseline year area. For "Lake A," its area change rate is (40-50) / 50 = -0.2, or -20%.

[0116] In one implementation of this invention, the area change rate of all lakes is iterated. When the area change rate of a lake is significantly negative (e.g., less than -1%), the unique identifier of the lake is added to a list called "shrinking lake set". When the change rate is significantly positive (e.g., greater than 1%), it is added to a list called "expanding lake set". For example, "Lake A" has a change rate of -20%, so it is classified into the shrinking lake set; while "Lake C" has an area change rate calculated to be positive 15%, so it is classified into the expanding lake set.

[0117] In one implementation of this invention, the vegetation cover of each lake in the base year and the evaluation year is extracted from the lake attribute parameters. For example, if the vegetation cover of "Lake A" is found to be 0.65 in the base year and 0.60 in the evaluation year, the difference between the two is calculated to obtain -0.05.

[0118] Specifically, each lake in the cluster of shrinking lakes will be traversed. For "Lake A," its area change is 40-50=-10 square kilometers, and the absolute value is taken as 10 square kilometers. The vegetation cover for the assessment year is extracted from the lake's attribute parameters, which is 0.60. Then, these two values ​​are multiplied together to obtain 10. 0.60 = 6.0. This result was recorded as the ecologically degraded area of ​​"Lake A".

[0119] Please see Figure 3 This is a comparison chart of lake area changes in this invention, used to show the area change characteristics of different lakes in two time periods (base period T1 and current period T2). Lake representation: Elliptical shape: The solid black ellipse represents the spatial extent of the lake in the base period (T1); the dashed black ellipse represents the spatial extent of the lake in the current period (T2); Labeling information: Lake name (Lake A, Lake B), area identifier (…). , ), Change state (shrinkage, expansion); Time evolution: Black arrows: indicate the direction of time progression from the baseline period T1 to the current period T2, reflecting the evolution of lake area over time; Area change: Numerical relationship: This indicates that a lake has shrunk (e.g., lake A). Indicates lake expansion (e.g., lake B); rate of change calculation: The result is <0 for shrinkage and >0 for expansion; Lake A and Lake B are typical lake samples in the study area, representing the area evolution patterns of shrinking and expanding lakes, respectively. In this invention, these lakes are used as ecological element monitoring objects. By extracting their area data at time phases T1 and T2, the dynamic changes of the lakes can be quantitatively analyzed, providing data support for regional water resource management and ecological environmental protection. For example, the area of ​​Lake A at T1 is... The area of ​​T2 is The calculated rate of change is negative, classifying it as a shrinking lake; the area of ​​lake B at T1 is... The area of ​​T2 is The rate of change is positive, and it is classified as an expanding lake.

[0120] Preferably, the comprehensive assessment of the regional ecological health index in step S3 includes:

[0121] Calculate the difference in water level between the two lakes at the two ends of a connected edge in the lake network topology graph. When the difference in water level is greater than zero, assign a water flow direction from the high water level to the low water level to the connected edge.

[0122] Starting from the shrinking lakes where shrinking lakes are concentrated, trace the flow direction to the expanding lakes where expanding lakes are concentrated, and record the transmission path;

[0123] Establish pairing relationships between each shrinking lake and each expanding lake based on the transmission path.

[0124] In this embodiment of the invention, each connected edge in the lake network topology graph is traversed. For each edge, the two lake nodes it connects are identified. For example, for a connected edge connecting "Lake A" and "Lake B", the water level elevation of these two lakes in the evaluation year is queried from the lake attribute parameters.

[0125] In one implementation of this invention, two lists of lakes are obtained: a shrinking set of lakes (e.g., {"Lake B", "Lake C"}) and an expanding set of lakes (e.g., {"Lake E", "Lake F"}). Then, starting from each lake in the shrinking set, a path tracing algorithm, such as depth-first search (DFS), is performed on the directed network topology graph.

[0126] In one implementation of this invention, the tracing begins with the shrinking lake "Lake C". Assume the network topology shows that "Lake C" has a directed edge pointing to "Lake D", and "Lake D" has a directed edge pointing to "Lake E". When the tracing reaches "Lake E", the expanded lake set is queried, and "Lake E" is found to be a member. At this point, a complete propagation path has been successfully found. It is important to note that the tracing process proceeds along all water flow directions until an expanded lake or a path endpoint (i.e., a lake without an outlet) is encountered. If the endpoint of a path is not an expanded lake, the path is considered an invalid propagation path and is not recorded.

[0127] In one implementation of this invention, all successfully recorded transmission paths are summarized. For each path, its starting point and ending point are extracted. For example, for the path ["Binghu"->"Dinghu"->"Wuhu"], the starting point is the shrinking lake "Binghu" and the ending point is the expanding lake "Wuhu", thus establishing a pairing relationship from "Binghu" to "Wuhu".

[0128] Of particular importance is establishing pairing relationships between shrinking and expanding lakes based on transmission paths, including:

[0129] The average water level gradient is obtained by calculating the ratio of the cumulative water level drop along the transmission path to the path length.

[0130] The average water level gradient from the same shrinking lake to all expanding lakes is normalized and used as the probability of water transfer.

[0131] The absolute value of the water volume change of shrinking lakes is taken as the water loss value, and the water volume change value of expanding lakes is taken as the water increase value.

[0132] The water loss value is allocated to the expanding lake according to the transmission probability to obtain the theoretical transmission volume;

[0133] When the proportion of theoretically transferred water volume to the increase in water volume is greater than the preset contribution threshold, a pairing relationship is established.

[0134] In one implementation of this invention, for example, for a transfer path from the shrinking lake "Binghu" to the expanding lake "Wuhu" ["Binghu"->"Dinghu"->"Wuhu"], the water level elevation of the starting point "Binghu" (e.g., 24.5 meters) and the water level elevation of the ending point "Wuhu" (e.g., 22.0 meters) are extracted from the lake attribute parameters, and the cumulative water level drop along the path is calculated to be 24.5-22.0=2.5 ​​meters. At the same time, the length of the path is calculated, which is obtained in advance by summing the geometric lengths of the river channel vector segments constituting the path using GIS tools.

[0135] Specifically, suppose that the shrinking lake "Lake C" has another transmission path to the expanding lake "Lake F," with a calculated average water level gradient of 0.00015. In this case, "Lake C" has two potential flow directions: "Lake E" and "Lake F." Adding the average water level gradient values ​​of these two flow directions yields a total gradient of 0.00005 + 0.00015 = 0.00020. Then, the average water level gradient of each flow direction is divided by this total gradient and normalized to obtain the water transmission probability for each flow direction.

[0136] Specifically, it is necessary to calculate the water volume change value for each lake. In this invention, this value is estimated by multiplying the area change of the lake between the baseline year and the assessment year by an average effective water depth (e.g., 0.5 meters) determined based on regional station data. Assuming that the calculated water volume change value for the shrinking lake "Lake C" is negative 2.5 million cubic meters, the absolute value yields a water loss of 2.5 million cubic meters.

[0137] Specifically, the water loss of 2.5 million cubic meters in Lake Bing will be allocated according to the previously calculated water transfer probability. The theoretical water transfer volume allocated to Lake Wu is also 2.5 million cubic meters. 0.25 = 625,000 cubic meters. The theoretical water transfer volume allocated to "Ji Lake" is 2.5 million cubic meters. 0.75 = 1.875 million cubic meters. It should be noted that this theoretical water transfer volume represents, under ideal conditions, how much of the water lost from "Lake Bing" contributes to the expansion of "Lake Wu" and "Lake Ji" respectively.

[0138] In one implementation of this invention, a contribution threshold is preset. This threshold is set based on regional water resource management needs and ecological significance, for example, 10% (i.e., 0.1), to determine whether an upstream shrinking lake is one of the "main" water sources for a downstream expanding lake. The contribution ratio of each allocation path is calculated. For the path from "Lake C" to "Lake E": its theoretical water transfer volume is 625,000 cubic meters, accounting for 625,000 / 4,000 = 0.156 of the 4,000,000 cubic meter increase in the water volume of "Lake E". For the path from "Lake C" to "Lake F": its theoretical water transfer volume is 1,875,000 cubic meters, accounting for 1,875,000 / 6,000 = 0.313 of the 6,000,000 cubic meter increase in the water volume of "Lake F". Since both the ratio values ​​of 0.156 and 0.313 are greater than the preset contribution threshold of 0.1, both paths are determined to be significant, and two clear pairing relationships are finally established: ("Binghu", "Wuhu") and ("Binghu", "Jihu").

[0139] Preferably, step S3, which comprehensively assesses the regional ecological health index of the current state, further includes:

[0140] The ratio of the actual number of connected edges to the theoretical maximum number of connected edges in the network topology graph is used as the network connectivity index.

[0141] Extract the area and vegetation cover of each lake in the hub lake set, calculate the sum of their products and divide it by the total area of ​​the lake group to obtain the hub ecosystem service index;

[0142] Calculate the average proportion of successfully paired shrinking and expanding lakes in the pairing relationships to obtain the water redistribution balance.

[0143] The ecological stability index is obtained by calculating the proportion of the total area of ​​ecological degradation in the region to the total area of ​​the lake group, and then subtracting this proportion from one.

[0144] The regional ecological health index is obtained by weighting and summing the network connectivity index, hub ecosystem service index, water redistribution balance index, and ecological stability index.

[0145] In one implementation of this invention, for example, for a transmission path from the shrinking lake "Binghu" to the expanding lake "Wuhu" ["Binghu" -> "Dinghu" -> "Wuhu"], the water level elevation of the starting point "Binghu" (e.g., 24.5 meters) and the water level elevation of the ending point "Wuhu" (e.g., 22.0 meters) are extracted from the lake attribute parameters, and the cumulative water level drop along the path is calculated to be 24.5 - 22.0 = 2.5 meters. Simultaneously, the length of the path is calculated, i.e., the sum of the lengths of all connected edges (river channels) along the path, assumed to be 50,000 meters (50 kilometers).

[0146] Specifically, assume that the shrinking lake "Lake C" has another transmission path to the expanding lake "Lake F," with a calculated average water level gradient of 0.00015. In this case, "Lake C" has two potential flow directions: "Lake E" and "Lake F." Adding the average water level gradients of these two directions yields a total gradient of 0.00005 + 0.00015 = 0.00020. Then, the average water level gradient of each flow direction is divided by this total gradient for normalization. The probability of water transmission to "Lake E" is 0.00005 / 0.00020 = 0.25; the probability of water transmission to "Lake F" is 0.00015 / 0.00020 = 0.75.

[0147] In one implementation of this invention, the water volume change value for each lake is calculated. This can be obtained by multiplying its area change value by an average regional water depth or by estimating based on the water level-area relationship. For simplicity, this example directly uses the area change value as a proxy for water volume. Querying the lake attribute parameters, we find that the area of ​​the shrinking lake "Lake C" decreased from 35 square kilometers to 30 square kilometers, with a water volume change value of -5 square kilometers. Taking the absolute value, the water volume loss is 5 square kilometers. The area of ​​the expanding lake "Lake E" increased from 40 square kilometers to 48 square kilometers, with a water volume increase value of 8 square kilometers. Similarly, the water volume increase value of the expanding lake "Lake F" is 12 square kilometers.

[0148] In one implementation of this invention, the water loss value of 5 square kilometers in "Binghu" is allocated according to the previously calculated water transfer probability, and the theoretical water transfer volume allocated to "Wuhu" is 5. 0.25 = 1.25 square kilometers, the theoretical water transfer volume allocated to "Ji Lake" is 5. 0.75 = 3.75 square kilometers. It should be noted that this theoretical water transfer represents, under ideal conditions, how much of the water lost from "Lake Bing" contributes to the expansion of "Lake Wu" and "Lake Ji" respectively.

[0149] Specifically, a contribution threshold is preset, for example, 10% (i.e., 0.1). This threshold is used to determine whether an upstream shrinking lake is one of the "major" water sources for a downstream expanding lake. The contribution ratio of each allocation path will be calculated. For the path from "Lake C" to "Lake E": its theoretical water transfer volume is 1.25 square kilometers, accounting for 1.25 / 8 = 0.156 of the 8 square kilometer increase in the water volume of "Lake E"; for the path from "Lake C" to "Lake F": its theoretical water transfer volume is 3.75 square kilometers, accounting for 3.75 / 12 = 0.313 of the 12 square kilometer increase in the water volume of "Lake F".

[0150] Specifically, two basic data points are needed: the total number of lake nodes in the network and the actual number of connected edges. These two data points can be directly obtained from the previously constructed lake network topology polish graph. Assume there are 20 lake nodes in the study area, and the actual number of connected edges in the network topology graph is 25. Next, we calculate the theoretical maximum number of connected edges in a network containing 20 nodes. This number is calculated by multiplying the number of nodes by (number of nodes minus 1) and then dividing by 2, which is 20. (20-1) / 2=190 edges. Finally, divide the actual number of connected edges by the theoretical maximum number of connected edges to obtain the network connectivity index, which is 25 / 190≈0.132.

[0151] In one implementation of this invention, all members are extracted from a previously identified set of hub lakes. Assume the set of hub lakes is {"Lake A", "Lake C"}. Then, the area and vegetation cover of these hub lakes, as well as the total area of ​​the entire lake group, are queried from the lake attribute parameters. Assume that "Lake A" has an area of ​​62.5 square kilometers and a vegetation cover of 0.72; "Lake C" has an area of ​​45.0 square kilometers and a vegetation cover of 0.65; and the total area of ​​the entire lake group is 420 square kilometers. The product of the area and vegetation cover of each hub lake is calculated and summed to obtain (62.5... 0.72)+(45.0 0.65) = 45 + 29.25 = 74.25. Then, divide this sum by the total area of ​​the lake group to get the hub ecosystem service index, which is 74.25 / 420 ≈ 0.177.

[0152] Specifically, the calculations are based on the previously analyzed sets of shrinking and expanding lakes and their pairing relationships. Assume there are 5 shrinking lakes and 6 expanding lakes in the region. After water transfer path analysis and pairing relationship establishment, it was found that 4 shrinking lakes successfully established significant pairing relationships with expanding lakes, while 5 expanding lakes received significant water replenishment from shrinking lakes. The proportion of successfully paired shrinking lakes to the total is calculated as 4 / 5 = 0.8. Then, the proportion of successfully paired (i.e., receiving replenishment) expanding lakes to the total is calculated as 5 / 6 ≈ 0.833. Finally, the arithmetic mean of these two proportions is taken to obtain the water redistribution balance, i.e., (0.8 + 0.833) / 2 ≈ 0.817.

[0153] Specifically, two key data points are needed: the previously calculated total area of ​​regional ecological degradation and the total area of ​​the lake system. Let's assume the total area of ​​regional ecological degradation (combining shrinkage area and surrounding vegetation conditions) is 25.2 equivalent square kilometers, and the total area of ​​the lake system is 420 square kilometers. Calculate the proportion of the degraded area: 25.2 / 420 = 0.06. Then, subtract this proportion from the number 1 to obtain the ecological stability index: 1 - 0.06 = 0.94.

[0154] Specifically, weights need to be assigned to four sub-indices. These weights are pre-determined by ecological experts based on the specific conditions and management objectives of the study area, and the sum of the weights is 1. Assume the assigned weights are: network connectivity (weight w1 = 0.2), hub ecosystem services (weight w2 = 0.3), water redistribution balance (weight w3 = 0.3), and ecological stability (weight w4 = 0.2). Each sub-indice is multiplied by its corresponding weight, and the results are then summed. Regional ecological health index = (0.132) 0.2)+(0.177 0.3)+(0.817 0.3)+(0.94 0.2) = 0.5126.

[0155] Of particular importance is obtaining the degree of water redistribution balance, including:

[0156] The success rate of pairing is calculated by taking the average of the proportion of successfully paired shrinking lakes to the total number of shrinking lakes and the proportion of successfully paired expanding lakes to the total number of expanding lakes.

[0157] Extract the transmission path node sequence of each pairing record from the pairing relationship, calculate the number of nodes of each path minus one as the path length, and calculate the average of all path lengths as the average transmission path length of the network.

[0158] Multiplying the pairing success rate by the reciprocal of the average network transmission path length yields the water redistribution balance.

[0159] In one implementation of this invention, assuming there are 5 shrinking lakes and 6 expanding lakes in the study area, the pairing relationship set is traversed, and the number of non-repeating shrinking lakes involved is counted, assuming the result is 4. Then, the proportion of successfully paired shrinking lakes to the total number is calculated, i.e., 4 / 5 = 0.8. Next, in the same way, the number of non-repeating expanding lakes involved in the pairing relationships is counted, assuming the result is 5. The proportion of successfully paired expanding lakes to the total number is calculated, i.e., 5 / 6 ≈ 0.833.

[0160] Specifically, all established pairing records will be traversed, and the corresponding transmission path node sequence will be extracted from each record. Assume four valid transmission paths are extracted, with node sequences as follows: ["Lake A" -> "Lake B" -> "Lake C"], ["Lake D" -> "Lake E"], ["Lake A" -> "Lake F"], ["Lake G" -> "Lake H" -> "Lake I" -> "Lake G"]. Then, the length of each path will be calculated, defined as the number of nodes on the path minus 1.

[0161] Specifically, the results calculated in the previous two steps will be used: the pairing success rate (0.817) and the average network transmission path length (1.75). First, the reciprocal of the average network transmission path length is calculated, i.e., 1 / 1.75 ≈ 0.571. Then, the pairing success rate is multiplied by this reciprocal value, resulting in a water redistribution balance of 0.817. 0.571≈0.466.

[0162] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0163] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for assessing the health of a regional ecosystem based on remote sensing data, characterized in that, Includes the following steps: Step S1: Acquire multi-temporal remote sensing images of the target ecological area; Band reflectance processing is performed on multi-temporal remote sensing images to extract the boundaries of each lake and calculate its geometric and ecological attributes, generating lake attribute parameters. Step S1, which involves processing the band reflectance of multi-temporal remote sensing images to extract the boundaries of each lake and calculate its geometric and ecological attributes, includes: Radiometric calibration and atmospheric correction are performed on multi-temporal remote sensing images to generate surface reflectance image data; Extract near-infrared and green reflectance from surface reflectance image data; The normalized remote sensing index image is obtained by dividing the difference between the reflectance of the green band and the reflectance of the near-infrared band by the sum of the two. The normalized remote sensing index image is vectorized to obtain the lake boundary vector data; Step S2: Based on lake attribute parameters, determine the topographic and hydrological connectivity between lakes, construct a lake connectivity matrix, and identify the hub lake set based on the topological structure of the lake connectivity matrix; construct a lake network topology graph based on the lake connectivity matrix, which includes lake nodes and connecting edges; Step S3: Based on the set of key lakes and lake attribute parameters, assess the changes in lake area and vegetation cover to quantify regional ecological degradation, and comprehensively evaluate the current state of the regional ecological health index; Step S3, based on the set of key lakes and lake attribute parameters, assesses the changes in lake area and vegetation cover to quantify regional ecological degradation, including: Obtain the lake area from multiple time phases in the lake attribute parameters and calculate the area change rate; Lakes are classified into shrinking lakes and expanding lakes based on the sign of their area change rate. Calculate the difference in vegetation cover at different time phases based on the vegetation cover in the lake attribute parameters; For lakes with concentrated shrinkage, the absolute value of the area change is multiplied by the corresponding vegetation cover to obtain the ecologically degraded area. The total ecological degradation area of ​​all shrinking lakes is calculated by summing up the ecological degradation area of ​​the region; step S3 includes a comprehensive assessment of the regional ecological health index of the current state, which includes: Calculate the difference in water level between the two lakes at the two ends of a connected edge in the lake network topology graph. When the difference in water level is greater than zero, assign a water flow direction from the high water level to the low water level to the connected edge. Starting from the shrinking lakes where shrinking lakes are concentrated, trace the flow direction to the expanding lakes where expanding lakes are concentrated, and record the transmission path; Establish pairing relationships between each shrinking lake and each expanding lake based on the transmission path; Step S4: Regulate the ecological water level of the key lakes according to the regional ecological health index, and formulate ecological water replenishment plans for shrinking lakes in order to achieve water resource management of the lake group regional ecosystem.

2. The method for regional ecosystem health assessment based on remote sensing data according to claim 1, characterized in that, Vectorization of normalized remote sensing index images includes: Pixels in the normalized remote sensing index image whose pixel value is greater than the preset water body identification threshold are marked as water body pixels, forming water body mask data. Connectivity analysis was performed on the water body mask data to identify each connected component; Remove connected components whose area is smaller than the preset minimum lake area, and mark the remaining connected components as independent lakes; The boundaries of each independent lake are vectorized into lake boundary vector data.

3. The method for regional ecosystem health assessment based on remote sensing data according to claim 1, characterized in that, Step S1, which involves processing the band reflectance of multi-temporal remote sensing images, extracting the boundaries of each lake, and calculating its geometric and ecological attributes, also includes: The area of ​​each lake is obtained by calculating the area of ​​the closed region enclosed by the boundary vectors of each lake based on the lake boundary vector data. Calculate the coordinates of the geometric center point in the lake boundary vector data, and use them as the centroid coordinates of each lake; Obtain digital elevation model data of the target ecological area; Based on the lake centroid coordinates, the shoreline elevation values ​​at the corresponding locations are extracted from the digital elevation model data and used as the lake water level elevations for each lake. Based on the lake boundary vector data, an ecological buffer zone is constructed by extending a preset buffer width outward from the boundary. Vegetation pixel values ​​were extracted from the normalized remote sensing index image within the ecological buffer zone, and the arithmetic mean of all vegetation pixel values ​​was calculated as the vegetation cover. The lake area, lake centroid coordinates, vegetation coverage, and lake water level elevation of each lake are associated as lake attribute parameters.

4. The method for regional ecosystem health assessment based on remote sensing data according to claim 1, characterized in that, In step S2, based on lake attribute parameters, the topographic and hydrological connectivity relationships between lakes are determined, a lake connectivity matrix is ​​constructed, and the set of hub lakes is identified based on the topological structure of the lake connectivity matrix, including: Extract the centroid coordinates of any two lakes from the lake attribute parameters, and calculate the Euclidean distance between the two lake centroid coordinates as the lake spacing. Based on the distance between lakes, terrain connectivity is determined. When two lakes are determined to be terrain-connected, they are marked as the first lake and the second lake. Construct a topographic profile along the line connecting the centroid coordinates of the first and second lakes; Extract the lake water level elevations corresponding to the first and second lakes from the lake attribute parameters, and take the maximum value of the two as the water level threshold. Determine whether there are topographic points in the topographic profile with elevation values ​​higher than the water level threshold. If there are no topographic points with elevation values ​​higher than the water level threshold, determine that the two lakes have a topographic connectivity relationship and use them as candidate connected lake pairs. Obtain the river network vector data of the area between the candidate connected lake pairs. River channels are identified based on river network vector data, and a lake connectivity matrix is ​​constructed based on the marked river channels for all candidate connected lake pairs.

5. The method for regional ecosystem health assessment based on remote sensing data according to claim 4, characterized in that, River channels are identified based on river network vector data, including: Based on the ecological buffer zone, candidate connected lake pairs are divided into a first lake buffer zone and a second lake buffer zone, respectively. Extract each river vector segment from the river network vector data and determine whether each river vector segment intersects with both the first and second lake buffer zones simultaneously. When there is a river vector segment that intersects both the first lake buffer zone and the second lake buffer zone, it is determined that the two lakes have a hydrological connectivity relationship, and the river vector segment is marked as a river channel.

6. The method for regional ecosystem health assessment based on remote sensing data according to claim 1, characterized in that, Step S2, based on lake attribute parameters, determines the topographic and hydrological connectivity relationships between lakes, constructs a lake connectivity matrix, and identifies the hub lake set based on the topological structure of the lake connectivity matrix. This step also includes: Calculate the shortest path between any two lake nodes based on the lake connectivity matrix, and count the number of node paths passing through the lake nodes. Calculate the total number of shortest paths between all pairs of lake nodes, and divide the node path count by the total number of shortest paths to obtain the betweenness centrality value of the lake nodes. Arrange the betweenness centrality values ​​in descending order to form a centrality sorting sequence; The upper quartile of the centrality sort sequence is calculated as the hub threshold, and lake nodes with betweenness centrality values ​​greater than the hub threshold are selected to obtain a preliminary hub lake set. Prioritize the initial set of hub lakes and then filter the hub lake sets.

7. The method for regional ecosystem health assessment based on remote sensing data according to claim 1, characterized in that, Step S3, which comprehensively assesses the regional ecological health index of the current state, also includes: The ratio of the actual number of connected edges to the theoretical maximum number of connected edges in the network topology graph is used as the network connectivity index. Extract the area and vegetation cover of each lake in the hub lake set, calculate the sum of their products and divide it by the total area of ​​the lake group to obtain the hub ecosystem service index; Calculate the average proportion of successfully paired shrinking and expanding lakes in the pairing relationships to obtain the water redistribution balance. The ecological stability index is obtained by calculating the proportion of the total area of ​​ecological degradation in the region to the total area of ​​the lake group, and then subtracting this proportion from one. The regional ecological health index is obtained by weighting and summing the network connectivity index, hub ecosystem service index, water redistribution balance index, and ecological stability index.

Images