A method and system for identifying and zoning control of hot and cold spots of ecological carrying capacity of a hydropower project

Abstract

This invention provides a method and system for identifying hotspots and cold spots in the ecological carrying capacity of hydropower projects and for zoning and managing them, applied in the fields of ecological environment monitoring and spatial statistics. This application first acquires multi-source basic data of hydropower projects, completes standardized preprocessing, and constructs a spatial weight matrix adapted to different disturbance patterns; then, through global and local spatial autocorrelation and hotspot statistics, spatial risk types are generated; subsequently, engineering disturbance correlation analysis is conducted to quantify risk contribution and identify key units; the evolution of hotspots and cold spots is dynamically assessed by combining multi-period data, and time-series indicators are quantified; finally, multi-dimensional information is integrated to delineate control zones, generate control recommendations, and integrate the entire chain of results to form a standardized control system, supporting precise management and control of ecological risks in hydropower projects.

Description

Technical Field

[0001] This invention relates to the field of ecological environment monitoring and spatial statistics technology, and in particular to a method and system for identifying hot and cold spots and zoning control of ecological carrying capacity in hydropower projects. Background Technology

[0002] In existing technologies, spatial diagnosis of ecological carrying capacity mainly relies on multi-source remote sensing, topographic and ground monitoring data. It identifies high / low value distribution patterns by constructing an evaluation index system, calculating the carrying capacity index, and conducting spatial statistical analysis. Spatial statistical methods primarily utilize global / local spatial autocorrelation and hot / cold spot analysis: global Moran's I is used to determine the overall clustering or dispersion characteristics of carrying capacity; local Moran's I (LISA) identifies four types of local clusters: high-high, low-low, high-low, and low-high; Getis-OrdGi... Statistical analysis identifies significantly cold and hot areas. The methods described above can reveal spatial clustering trends in carrying capacity, providing a fundamental support for risk identification.

[0003] However, existing technologies still have significant shortcomings in practical applications of hydropower projects: First, existing spatial analysis methods mostly remain at the level of pattern description and are difficult to directly translate into executable control zones for engineering projects. Global / local spatial autocorrelation and hotspot analysis outputs are mostly statistical quantities, significance, and thematic maps, lacking standardized discrimination rules for transforming spatial statistical results into engineering zones for strict control, ecological restoration, risk prevention, key protection, and general monitoring. This makes it difficult for analysis results to directly guide construction boundary control, spoil disposal, vegetation restoration, and dynamic inspections. Second, global spatial autocorrelation can only reflect the overall pattern and cannot accurately locate local high-risk patches. Ecological risks in hydropower projects are often locally concentrated, such as low-value clusters easily forming at spoil disposal sites, tunnel entrances, and along construction roads. Global statistical results cannot pinpoint specific risk locations, resulting in insufficient refined control capabilities. Third, single local clustering or hotspot analysis cannot simultaneously consider cluster type and significance intensity. LISA focuses on local similarity / anomalies and cannot quantify the significance of cold and hot spots; cold and hot spot analysis focuses on significance and has difficulty distinguishing high and low anomaly transition zones, which can easily lead to rough risk classification and insufficient targeted control.

[0004] Fourth, existing methods lack spatial correlation with engineering disturbance units. After identifying cold and hot spots, it is difficult to directly trace the source of disturbance to specific disturbance sources such as construction roads, spoil heaps, and tunnel entrances. This makes it impossible to quantify the risk contribution of different disturbances, hindering the clarification of control responsibilities and remediation priorities. Fifth, there is a lack of temporal dynamic identification capabilities. Existing technologies mostly conduct static analysis on single-period data, making it difficult to identify the evolution of cold and hot spots during construction expansion, operation phase decline, and new construction phase additions. This fails to support dynamic control and remediation effectiveness evaluation throughout the entire process. Sixth, multi-scale adaptability is insufficient. Hydropower engineering disturbances include point-like (tunnel entrances), linear (construction roads), patchy (spoil heaps), and corridor-like (riverbank disturbances). Existing methods often use a uniform neighborhood scale, making it difficult to adapt to the spatial impact characteristics of different disturbances, easily leading to identification bias. Seventh, there is a lack of a standardized output system. Existing analysis results have heterogeneous formats and inconsistent fields, making it difficult to directly interface with engineering management platforms and monitoring systems, resulting in insufficient reusability and operability of the results.

[0005] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this disclosure, and therefore includes information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] According to one aspect of this application, a method for identifying and zoning control of hot and cold spots in the ecological carrying capacity of hydropower projects is provided, comprising: acquiring ecological carrying capacity evaluation results, spatial data of engineering disturbances, ecological restoration data, and auxiliary geographic environment data of the impact area of ​​hydropower projects; completing data preprocessing based on spatial registration, scale unification, effective pixel screening, and standardization; constructing a multi-scale adaptive spatial weight matrix according to different engineering disturbance forms of construction roads, spoil heaps, and tunnel entrances; calculating the global spatial autocorrelation index, outputting the global spatial autocorrelation results, diagnosing the overall clustering degree of ecological carrying capacity, calculating local spatial autocorrelation indicators, outputting local clustering maps, identifying strong / medium / weak hot and cold spots through Getis-OrdGi statistics, integrating local clustering types and the significance of hot and cold spots, outputting a spatial risk map, and constructing spatial risk types of ecological carrying capacity; performing spatial overlay and buffer analysis on spatial risk types and engineering disturbance units, calculating the correlation ratio of cold spots, etc. Based on the low-to-low clustering correlation ratio, risk intensity, and risk area proportion, a risk contribution index for engineering disturbance units is constructed to identify key risk contribution units. Dynamic changes in multi-period ecological carrying capacity results are identified to determine the change types of newly added / continuous / receding cold spots and newly added / continuous / degraded hot spots, quantifying the expansion and regression rates of cold spots and outputting time-series change maps. Based on spatial risk types, the significance of cold and hot spots, the correlation intensity of engineering disturbances, and multi-period change trends, strict control zones, ecological restoration priority zones, risk diffusion prevention and control zones, key protection zones, and general monitoring zones are delineated according to preset rules, outputting zoning boundaries, hierarchical distribution maps, and risk tracing results, generating differentiated control recommendations. Finally, by integrating global spatial autocorrelation results, local cluster maps, cold and hot spot distribution maps, spatial risk maps, engineering disturbance risk contribution lists, time-series change maps, zoning boundaries, hierarchical attributes, risk tracing, and differentiated control recommendations, the results of cold and hot spot identification and zoning control for hydropower engineering ecological carrying capacity are formed.

[0007] Another aspect of this application discloses a system for identifying hot and cold spots and implementing zoned management of ecological carrying capacity in hydropower projects, comprising: a data acquisition module for acquiring ecological carrying capacity evaluation results, spatial data of engineering disturbances, ecological restoration data, and auxiliary geographic environment data of the hydropower project's impact area, and generating basic dataset information; a data preprocessing and weight construction module for performing spatial registration, scale unification, effective pixel screening, and standardization processing, constructing a multi-scale adaptive spatial weight matrix based on the differentiated disturbance patterns of construction roads, spoil heaps, and tunnel entrances, and generating adaptive weight information; and a spatial statistics and risk construction module for calculating the global spatial autocorrelation index, local spatial autocorrelation index, and Getis-OrdGi statistics, identifying spatial clustering characteristics, clustering types, and hot and cold spot levels, fusing them to generate spatial risk types of ecological carrying capacity, and outputting global statistics and local clustering data. The system comprises the following modules: a category, hotspot and coldspot, and spatial risk information; a disturbance correlation and contribution assessment module, used to conduct spatial overlay and buffer analysis, calculate the coldspot correlation ratio, low-low clustering correlation, risk intensity and area proportion, construct the risk contribution index of engineering disturbance units and identify key risk units, and generate disturbance correlation assessment results; a dynamic change identification module, used to conduct time series analysis on multi-period data, determine the evolution type of hotspots and coldspots and quantify the expansion rate and recession rate, and output time series change maps and dynamic evolution information; a zoning and suggestion generation module, used to integrate risk type, significance, correlation intensity and time series trend, delineate five types of control zones and output boundary, hierarchical attributes, and risk source tracing results, and generate differentiated control suggestion information; and a results integration and output module, used to summarize the full-chain analysis results to form standardized results information on the identification and zoning control of hotspots and coldspots in the ecological carrying capacity of hydropower projects.

[0008] According to another aspect of this application, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a second processor, implements the above-described method for identifying hot and cold spots and zoning control of ecological carrying capacity in hydropower projects.

[0009] The beneficial effects of this application are as follows: First, multi-source basic data is acquired, and after standardized preprocessing, a multi-scale adaptive spatial weight matrix is ​​constructed. Then, spatial risk types are generated by fusing global and local spatial autocorrelation and hotspot statistics. Next, engineering disturbance correlation analysis is carried out to quantify risk contribution and identify key units. The evolution of hotspots is dynamically assessed by combining multi-period data, and time-series indicators are quantified. Finally, five-level control zones are delineated by comprehensively considering multi-dimensional information, and differentiated control recommendations are generated, forming standardized results. This invention solves the problems of existing technologies, such as difficulty in implementation, difficulty in tracing the source of risks, difficulty in assessing time series, and poor adaptability, realizing the full-chain transformation of ecological risks from statistical identification to engineering control.

[0010] This application aims to integrate spatial autocorrelation and hotspot / colds statistics to accurately identify localized risk patches, enhance refined management capabilities, quantify the risk contribution of different disturbance sources, clarify management responsibilities, and improve the targeting of management. Furthermore, it identifies temporal evolution patterns to support dynamic decision-making throughout the construction and operation cycle, constructs multi-scale weights to adapt to point-like, linear, and patchy disturbances, and achieves more accurate identification; ultimately, it outputs standardized results that can be directly integrated into engineering management, improving the standardization and intelligence of ecological management.

[0011] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0012] Figure 1 This invention provides a flowchart of a method for identifying hot and cold spots and zoning control of ecological carrying capacity in hydropower projects, according to an embodiment of this application. Figure 2 This paper presents a schematic diagram of the structure of a hydropower project ecological carrying capacity hotspot identification and zoning control system according to an embodiment of this application. Detailed Implementation

[0013] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0014] The following is combined Figure 1 This application describes a method for identifying hot and cold spots and implementing zoning control of ecological carrying capacity in hydropower projects according to exemplary embodiments: S101, obtain the evaluation results of the ecological carrying capacity of the hydropower project impact area, spatial data of project disturbance, ecological restoration data and auxiliary geographical environment data.

[0015] In one implementation, ecological carrying capacity raster data, vector evaluation unit data, and classification results are collected for 3-5 consecutive periods (including pre-construction, mid-construction, and initial operation) covering the hydropower project construction area, reservoir area, drawdown zone, and a surrounding 5-10 km influence range. The raster resolution is uniformly 30m, with an ecological carrying capacity index of 0-1 pixels, classified into three levels: 0-0.3 (low), 0.3-0.7 (medium), and 0.7-1 (high). Vector units are divided by small watershed, administrative village, and construction area, with attributes including index value, level, area, and boundary coordinates. For example, a construction area raster index of 0.22 corresponds to a low carrying capacity level; a reservoir area vector unit with an area of ​​2.8 km² and an index of 0.78 corresponds to a high carrying capacity level. This data is used for subsequent global Moran's I, local LISA, and Getis-OrdGi analyses. Spatial heterogeneity and clustering analysis.

[0016] The system collects vector boundaries and attributes for all types of disturbances, including construction roads, spoil heaps, tunnel entrances, mixing systems, construction platforms, temporary storage yards, riverbank disturbance zones, construction camps, and quarrying areas. Construction roads are linear, with widths of 5–12 m and lengths of 1.2–8.5 km, and attributes include grade, construction stage, and disturbance intensity (high / medium / low). Spoil heaps are patchy, with areas of 0.08–1.5 km², and attributes include stockpile volume, activation time, and slope gradient. Tunnel entrances are point-like, with coordinate accuracy ≤1 m, and attributes include diameter, excavation depth, and disturbance grade. Riverbank disturbance zones are wide, with widths of 20–50 m and lengths of 3–12 km, and attributes include shoreline type and disturbance duration. Spoil heap No. 1 has an area of ​​0.62 km² and high disturbance intensity; construction road R3 is 4.8 km long and has medium disturbance intensity. The system accurately characterizes the four types of disturbance morphologies and intensities: point-like, linear, patchy, and corridor-like.

[0017] Vector boundaries and records were collected for the revegetation area, slope protection area, topsoil backfilling area, vegetation reconstruction area, drainage engineering area, and ecological restoration monitoring area of ​​the spoil heap. The revegetation area ranged from 0.1 to 1.2 km², with attributes including tree species, planting density, and restoration years. The slope protection area ranged from 25° to 45° and length from 0.8 to 3.6 km, with attributes including protection type and construction time. The monitoring area included 10m × 10m quadrats, containing vegetation cover, biomass, and annual monitoring records. The revegetation area of ​​spoil heap No. 2 was 0.35 km², restored for 3 years, with a vegetation cover of 78%. The slope protection area S1 was 2.1 km long with a slope of 32°. This provides quantitative data on area, years, and effectiveness for the delineation of priority ecological restoration areas.

[0018] Data was collected for a digital elevation model (DEM), including slope, aspect, drainage system, watershed boundaries, land cover, nature reserve boundaries, administrative boundaries, and meteorological station data. The DEM resolution was 30m, with elevations ranging from 500 to 4500m; slopes from 0° to 60°, and aspects from 0° to 360°; drainage system accuracy was 1:50,000, with a watershed area ≥5km²; land cover included forest, grassland, cultivated land, and construction land, with a classification accuracy ≥90%; meteorological station data included monthly precipitation, temperature, and wind speed data for the past 5 years, with accuracy of 0.1℃, 0.1mm, and 0.1m / s, respectively. All data were collected using the CGCS2000 coordinate system and UTM projection to ensure spatial registration, scale consistency, and neighborhood correction accuracy, supporting the construction of distance attenuation weights, slope confluence neighborhoods, and wind direction correction neighborhoods.

[0019] S102 completes data preprocessing based on spatial registration, scale unification, effective pixel screening and standardization, and constructs a multi-scale adaptive spatial weight matrix according to the different engineering disturbance forms of construction roads, spoil heaps and tunnel entrances.

[0020] In one implementation, the system first integrates three types of basic data on the impact area of ​​hydropower projects to ensure coverage of the ecological baseline, engineering disturbances, and geographical background, providing a complete and reliable data foundation for subsequent spatial weighting and spatial statistical analysis. The first type is multi-form ecological carrying capacity data, encompassing raster-format continuous index data, vector patch zoning data, small watershed unit statistical data, and construction management unit division data. The raster data uses a 30m standard resolution, with numerical ranges corresponding to ecological carrying capacity from low to high; the vector patch area is controlled between 0.5 and 5 km², recording carrying capacity level, boundary coordinates, and unit area; small watershed units are divided according to natural catchment boundaries, with areas ranging from 5 to 50 km²; construction management units are divided by work area and operational area, adapting to different spatial scale evaluation and analysis needs. The second type is comprehensive engineering disturbance data, including linear construction roads, patchy spoil heaps, point-like tunnel entrances, mixing systems, construction platforms, construction camps, quarrying areas, and corridor-like riverbank disturbance zones. The construction roads are 5–12m wide and 1–10km long; the spoil heap area is 0.1–2km²; the tunnel entrance diameter is 3–8m; the riverbank disturbance bandwidth is 20–50m wide and 2–15km long. All disturbance units are recorded with location, area, type, construction stage, and disturbance intensity (high / medium / low), fully covering four disturbance types: point-like, linear, patchy, and corridor-like. The third category is geographic auxiliary data, including digital elevation models, slope, aspect, water systems, watershed boundaries, land cover, nature reserve boundaries, administrative boundaries, and meteorological station observation data. The digital elevation model has a resolution of 30m and an elevation range of 500–4500m; slopes range from 0° to 60° and aspects from 0° to 360°; the water system accuracy is 1:50,000; and the meteorological data accuracy is 0.1℃, 0.1mm, and 0.1m / s, providing a geographic benchmark for spatial registration, scale unification, neighborhood correction, and topographic impact analysis.

[0021] A unified coordinate system was implemented, converting all data to the CGCS2000 national geodetic coordinate system and UTM projection. The original WGS84 coordinate data underwent parameter transformation, with coordinate errors controlled within 0.1m, eliminating benchmark deviations. Spatial resolution was matched, with raster data uniformly using a 30m resolution and vector data aligned to a 1:10,000 spatial accuracy. Existing 10m resolution elevation data was resampled to 30m to ensure consistent spatial scale between raster and vector data. The study area was cropped, using the hydropower project construction area, reservoir area, drawdown zone, and a surrounding 5km buffer zone as boundaries. Irrelevant areas were removed. The total study area is approximately 800km²; after cropping, the effective analysis area was retained, while peripheral mountains, water bodies, and other irrelevant areas were eliminated.

[0022] Invalid data was removed, including missing pixels, abnormal pixels, and invalid vector patches with an area less than 0.01 km². Approximately 23,000 missing or abnormal pixels were removed, along with 12 small, invalid disturbance patches, ensuring data quality. Time-series registration was performed, selecting data from three phases: pre-construction, mid-construction, and initial operation, with a unified timeframe to ensure time-series comparability. Data from all three phases was collected annually in October, with time errors controlled within 15 days to align the time-series dimensions. Ecological carrying capacity was classified into three levels—high, medium, and low—based on numerical ranges: 0–0.3 for low carrying capacity, 0.3–0.7 for medium carrying capacity, and 0.7–1 for high carrying capacity. For example, a raster value of 0.22 in a construction area corresponds to a low level, while a unit value of 0.78 in the reservoir area corresponds to a high level, eliminating dimensional differences and facilitating spatial comparison. Inter-phase unit comparability verification involved checking the consistency of values ​​and levels for each unit across the three phases, ensuring direct comparison between data from different periods and units. Fifty spatial units were randomly selected, and the consistency of the three phases reached over 95%, ensuring accurate identification of temporal changes.

[0023] After preprocessing including unified coordinates, scale matching, invalid removal, temporal registration, hierarchical classification, and comparability verification, a multi-mode, multi-scale adaptive spatial weight matrix is ​​constructed based on four differentiated disturbance morphologies of hydropower projects: linear, corridor-like, patchy, and point-like. This matrix accurately adapts to the spatial impact characteristics of different disturbances, providing a reliable neighborhood relationship foundation for subsequent global / local spatial autocorrelation and hotspot statistics. For linear construction roads and corridor-like riverbank disturbance zones, the construction roads are linear disturbances, typically 5–12m wide and 1–10km long; the riverbank disturbance zones are corridor-like, with a bandwidth of 20–50m and a length of 2–15km, extending in a strip-like pattern. Spatial weights are constructed using a distance attenuation model and a river network adjacency model. Specifically, the distance attenuation weight formula... ,in, spatial unit With spatial units The distance between them The neighborhood distance threshold, This represents the distance attenuation coefficient. For river network adjacency, adjacency relationships are established along the river flow direction. Units within the same river segment have a weight of 1, while cross-segment units are assigned values ​​based on distance attenuation. Neighborhood correction is as follows: wind direction (with a 20% increase in weight downstream of the prevailing wind direction), annual average wind speed, and dust diffusion distance are used to correct the neighborhood range, matching the zonal diffusion effect. For a certain construction road, unit weights within 300m on both sides are 0.3–0.8, while weights beyond 800m are 0, consistent with the zonal influence characteristics of linear disturbances.

[0024] For patchy spoil heaps and construction platforms, with spoil heap areas ranging from 0.1 to 2 km² and construction platforms from 0.05 to 0.8 km², exhibiting a planar radiation effect, a planar buffer weighting method is constructed using an adjacency matrix + K-nearest neighbor model. Specifically, for the adjacency matrix, directly adjacent units have a weight of 1, while indirect adjacent units decrease in weight based on distance. For the K-nearest neighbor model, the 8 to 12 nearest neighbor units are used in the weighting calculation to match the patch radiation range. For neighborhood correction, slope (increased weight for <25°, decreased weight for >45°) and slope confluence direction are considered, with downstream units of the confluence having a 30% increased weight to match the slope material migration effect. The core weighting zone is defined as 500m around a 0.6 km² spoil heap, with a weight ranging from 0.4 to 1.0, gradually decreasing outwards to match the planar radiation characteristics of the patch.

[0025] For point-like tunnel entrances and mixing systems, with tunnel entrance diameters of 3–8 m and mixing system footprints of 0.02–0.2 km², the impact is concentrated in concentric rings. A ring-shaped buffer weight is constructed using a distance attenuation + K-nearest neighbor model. Specifically, for the ring-shaped layering, the weight of the core area (0–200 m) is 0.6–1.0, the weight of the transition area (200–500 m) is 0.2–0.5, and the weight outside 500 m is 0. For the K-nearest neighbor model, 6–10 surrounding units are selected to highlight the point-like core impact. For neighborhood correction, the slope confluence and dust diffusion range are integrated, and the weight of the downstream of the confluence and the main dust diffusion direction is increased by 25% to match the point-like concentric ring impact. The weight of units within 150 m of the tunnel entrance is 0.7–0.9, and the weight outside 300 m approaches 0, which fits the characteristics of the point-like concentric ring impact.

[0026] The linear, patch, and point weight matrices are fused unit by unit to form a multi-scale adaptive spatial weight matrix. Specifically, the linear, patch, and point weights are weighted averaged, with linear perturbation weights accounting for 40%, patch weights for 35%, and point weights for 25%. The weight matrix has a non-zero row sum, no null values, and smooth transitions between adjacent unit weights. Outliers are removed and reassigned. The final output is a weight matrix in a unified format, compatible with subsequent global Moran's I, local LISA, and Getis-OrdGi algorithms. Calculations ensure that spatial statistical results accurately match the actual impact range of disturbances.

[0027] In another implementation, spatial autocorrelation and hot / cold spot identification require first determining the adjacency relationships between spatial units. This invention constructs a spatial weight matrix based on the disturbance characteristics of hydropower projects to quantify the degree of mutual influence between different spatial units, improving the adaptability of subsequent spatial statistical analysis to different engineering disturbance patterns and avoiding bias in hot / cold spot identification results caused by a fixed neighborhood scale. Let the spatial weight matrix be... ,in, Representing spatial units With spatial units Spatial relationship weights between them.

[0028] In one embodiment, the spatial weights can be determined using a distance attenuation method, where the weights decrease as the distance between cells increases. The specific formula is as follows: ,in, spatial unit With spatial units The distance between them The neighborhood distance threshold, This is the distance attenuation coefficient. Set the neighborhood distance threshold. =500m, distance attenuation coefficient λ=1.5. When the distance between units is 100m, the weight is 1 / 1001.5=0.001; when the distance exceeds 500m, the weight is directly set to 0, so as to reflect the law of the attenuation of the disturbance effect with distance.

[0029] In another embodiment, the adjacency matrix method can also be used to determine spatial weights, assigning weights only to directly adjacent units. The eight-adjacency rule is adopted. If cell i and cell j have a common edge or vertex in the grid, they are determined to be adjacent and the weight is set to 1; otherwise, it is set to 0, which directly reflects the adjacency relationship of the spatial cells.

[0030] For the impact zone of hydropower projects, a multi-scale adaptive neighborhood approach is preferred, constructing differentiated neighborhood rules for different disturbance source morphologies: For linear or corridor-shaped disturbance sources such as construction roads and riverbank disturbance zones, a buffer neighborhood along the route is adopted. Using the centerline of the disturbance source as a reference, buffer zones are generated on both sides along the direction. The buffer distance can be set according to the disturbance impact range; for example, the buffer distance on one side of the construction road is set to 500m, and the buffer distance on one side of the riverbank disturbance zone is set to 800m. The neighborhood weight is calculated according to the distance attenuation formula to reflect the extended impact of the strip-shaped disturbance.

[0031] For patchy disturbance sources such as spoil heaps and construction platforms, a planar buffer neighborhood is used. Multiple planar buffer zones are generated outwards from the disturbance source boundary. For example, a 0-300m core buffer zone and a 300-800m transition buffer zone are set around a spoil heap. The core buffer zone has a higher weight, while the transition zone weight decreases with distance, reflecting the radiating effect of planar disturbances. For point disturbance sources such as tunnel entrances and mixing systems, a ring-shaped buffer neighborhood is used. Multiple ring-shaped buffer zones are generated outwards from the disturbance source point. For example, a 0-200m core ring and a 200-500m transition ring are set around a tunnel entrance. The weight of the core ring is calculated using a distance-attenuation formula, while the weight of the transition ring gradually decreases, reflecting the layered effect of point disturbances.

[0032] For areas with significant slope disturbances, the neighborhood correction can be applied using slope aspect, slope gradient, and confluence path. For example, in areas with a slope greater than 25°, the weight of downhill units should be appropriately increased, while the weight of uphill units should be appropriately decreased. Simultaneously, units along the slope confluence path should have a higher weight than those in non-confluence paths, reflecting the guiding role of slope material migration on the disturbance's impact. For areas with significant dust or gas diffusion, the neighborhood correction can be applied using wind direction and wind speed. For example, downstream of the prevailing wind direction, the unit weight should be increased by 20%–30%, while the weight upstream should be appropriately decreased, reflecting the wind field's guiding role in the diffusion of the disturbance's impact.

[0033] S103 calculates the global spatial autocorrelation index, outputs the global spatial autocorrelation results, diagnoses the overall clustering degree of ecological carrying capacity, calculates local spatial autocorrelation indicators, outputs local clustering maps, identifies strong / medium / weak hot and cold areas through Getis-OrdGi statistics, integrates local clustering types and the significance of hot and cold areas, outputs spatial risk maps, and constructs spatial risk types for ecological carrying capacity.

[0034] In one implementation, the preprocessed ecological carrying capacity data is analyzed using the global Moran's I index algorithm to determine overall spatial clustering. The significance of Z-values ​​and P-values ​​is used to identify clustering, discrete, or random distribution characteristics, thus revealing global spatial heterogeneity patterns. After preprocessing steps including unified coordinates, scale matching, invalid data removal, temporal registration, grading, and comparability verification, global spatial autocorrelation analysis is performed on the rasterized ecological carrying capacity data. The global Moran's I index algorithm is then used to diagnose the spatial distribution pattern of ecological carrying capacity in the study area at a holistic level. The input data consists of standardized raster ecological carrying capacity values ​​with a unified spatial benchmark and 30m resolution. Each raster pixel corresponds to a spatial unit, with values ​​ranging from 0 to 1, representing carrying capacity from low to high.

[0035] First, the global spatial autocorrelation index of ecological carrying capacity is calculated to determine whether there is significant clustering in the overall spatial distribution of ecological carrying capacity in the study area. Let the mean ecological carrying capacity be... Where m is the total image element. Let be the carrying capacity value of the i-th cell. The global Moran's I can be represented as... ,in, ,when Furthermore, passing the significance test indicates that high or low values ​​of ecological carrying capacity exhibit spatial clustering; when When the value is close to 0, it indicates that the spatial distribution is approximately random; when... When, it indicates that the high and low values ​​are spatially discrete or interleaved.

[0036] Simultaneously, standardized statistics can be calculated. and according to and The value is used to determine whether the global spatial autocorrelation is significant. It assesses the clustering of the spatial pattern of ecological carrying capacity at the overall scale and provides a basis for subsequent identification of local hotspots and cold spots.

[0037] A local Moran's I calculation model is constructed, and combined with a neighborhood association analysis mechanism, to accurately identify five spatial clustering types: high-high, low-low, high-low, low-high, and insignificant, thereby locating local clusters and anomalous regions. Based on global spatial autocorrelation diagnosis, this invention further calculates local spatial autocorrelation indices to identify the local clustering type of each spatial unit, locating local high-value clusters, low-value clusters, and spatial anomalous regions in the spatial pattern of ecological carrying capacity. The local Moran's I calculation model is constructed as follows: the local Moran's I of the i-th spatial unit can be expressed as... ,in, In the formula, Let be the ecological carrying capacity value of the i-th spatial unit; This represents the average ecological carrying capacity of the study area. The variance of the ecological carrying capacity of the study area; The spatial weight matrix elements are determined by the aforementioned multi-scale adaptive neighborhood rule; m is the total number of spatial units in the study area. The model input consists of the ecological carrying capacity value and spatial weight matrix of a single spatial unit, and the output is the local Moran's I index value, significance p-value, and corresponding clustering type for each unit.

[0038] according to Based on the sign relationship of the deviation from the neighborhood mean, spatial units can be divided into five categories, as follows: The high-high agglomeration zone (HH) has high ecological carrying capacity in both the unit itself and its neighboring areas. The forest unit surrounding the reservoir has a carrying capacity value of 0.82 and a neighboring average of 0.78. The calculated local Moran's I index is positive with P < 0.01, thus it is identified as a high-high agglomeration zone and can be considered as a candidate area for the core protection zone of ecological carrying capacity.

[0039] The low-low agglomeration zone (LL) has low ecological carrying capacity in both the local and neighboring areas. The surrounding units of the spoil disposal site have a carrying capacity value of 0.21 and a neighboring average of 0.24. The calculated local Moran's I index is positive with P < 0.01, classifying it as a low-low agglomeration zone and a candidate area for strict ecological risk control.

[0040] The High-Low Anomaly Zone (HL) has high ecological carrying capacity within its own unit but low ecological carrying capacity in its neighboring areas. A small, remaining forest unit within the construction area has a carrying capacity value of 0.76 and a neighboring average of 0.32. The calculated local Moran's I index is negative with P < 0.05, classifying it as a High-Low Anomaly Zone. This area can be considered a candidate for a risk diffusion control zone or an ecological transition zone.

[0041] The Low-High Anomaly Zone (LH) has low ecological carrying capacity within the unit itself, but high ecological carrying capacity in its neighboring areas. Units at the edge of forest land disturbed by construction activities have a carrying capacity value of 0.28 and a neighboring mean of 0.75. The calculated local Moran's I index is negative with P < 0.05, thus it is identified as a Low-High Anomaly Zone and can be considered as a candidate area for risk diffusion prevention and control or ecological restoration potential.

[0042] The non-significant area (NS) has no significant local spatial correlation. The flat grassland unit far from the construction area has a bearing capacity value of 0.51, which is close to the average value of the whole area. The neighboring values ​​are randomly distributed. The calculated local Moran's I index is close to 0 and P>0.10. It is judged as a non-significant area and can be used as a regular monitoring area.

[0043] Through the above calculations and classifications, the local spatial pattern of the impact zone of hydropower projects can be accurately identified: the HH zone is the core protection area for ecological carrying capacity and will be used as the basis for delineating key protection areas in the future; the LL zone is the area with concentrated ecological risks and will be used as the basis for delineating strict control areas in the future; the HL and LH zones are ecological transition or risk diffusion frontiers and will be used as the basis for delineating risk diffusion prevention and control areas or priority areas for ecological restoration in the future.

[0044] Introducing Getis-OrdGi The statistical calculation model uses three significance thresholds—P<0.01, 0.01≤P<0.05, and 0.05≤P<0.10—to classify strong / medium / weak cold spots and strong / medium / weak hot spots, quantifying the clustering intensity of cold and hot spots and generating local cluster maps. Building upon local spatial clustering identification, this application further utilizes cold and hot spot statistics to identify significant hot and cold spots related to ecological carrying capacity, quantifying the significance intensity of high-value and low-value spatial clustering, and providing a risk level basis for subsequent zoning and hierarchical management. (This is related to Getis-OrdGi...) The statistical calculation model, the Getis-OrdGi of the i-th spatial unit. The statistic can be expressed as ,in, These are the elements of the spatial weight matrix; This represents the ecological carrying capacity value of the neighborhood unit. denoted as the mean ecological carrying capacity of the study area; S is the standard deviation of the ecological carrying capacity values; and m is the total number of neighborhood units. The model input consists of the preprocessed ecological carrying capacity values ​​and a multi-scale adaptive spatial weight matrix, and the output is the ecological carrying capacity value for each spatial unit. Statistical value, standardized Z-score, and significance P-value. When When the value is significantly positive, it indicates that the spatial unit is in a high-value hotspot region; when A significantly negative value indicates that the spatial unit is in a low-value cold spot region.

[0045] Hotspots and coldspots are categorized into different levels based on their significance, as follows: Strong hotspots: A p-value greater than 0 and p < 0.01 indicates a high concentration of high values ​​for ecological carrying capacity; (Key points: ...) Values ​​>0 and 0.01≤p<0.05 indicate a significant clustering of high values ​​for ecological carrying capacity; weak hotspots: Values ​​greater than 0, and 0.05 ≤ p < 0.10, indicate high ecological carrying capacity with weak aggregation; strong cold point: <0 and p<0.01 indicates a high concentration of low values ​​of ecological carrying capacity; Mid-cold point: <0, and 0.01≤p<0.05, indicates a significant clustering of low values ​​for ecological carrying capacity; weak cold spots: <0, and 0.05≤p<0.10, indicates low ecological carrying capacity and weak clustering; not significant: in cases other than the above conditions, there is no significant clustering characteristics of hot and cold spots.

[0046] Ecological carrying capacity value of the grid unit surrounding the tunnel entrance of a hydropower project =0.18, substituting into the formula, we get: = A value of 2.85, corresponding to p=0.008<0.01, indicates a strong cold spot, signifying a high concentration of low values ​​in the area, directly related to strong disturbances such as tunnel excavation and spoil disposal, representing the highest ecological risk. Ecological carrying capacity value of forest land units in the reservoir area. =0.83, calculated as follows: =2.76, corresponding to p=0.015, indicating a medium-hotspot, signifying a significant clustering of high values ​​and stable, high-quality ecological conditions in the area. Ecological carrying capacity value of the edge unit of the construction area. =0.35, calculated as follows: = A value of 1.62, corresponding to p=0.078, indicates a weak cold spot, signifying low-value weak clustering and a low risk level in this area. Based on the above classification results, a distribution map of cold and hot spots was generated, marking the spatial locations of strong / medium / weak cold and hot spots, and quantifying the clustering intensity of different regions. This result provides a direct risk level basis for subsequent construction of spatial risk types for ecological carrying capacity, correlation analysis of engineering disturbances, and zonal management.

[0047] To overcome the limitations of single LISA or single hotspot / coldspot analysis, this invention integrates local spatial clustering type with the saliency of hotspots / coldspots to construct a refined spatial risk type for ecological carrying capacity, thereby improving the precision of hotspot / coldspot identification results. Let the LISA type of the i-th spatial unit be... Hot and cold types are The model takes as input the LISA clustering results and hot / cold spot significance levels for each spatial unit, and outputs the corresponding composite spatial risk type. The comprehensive discrimination rules are as follows: If the core cold spot area of ​​spatial risk is determined =LL (low-low agglomeration area), and If the area is classified as a strong or moderate cold spot, it is determined to be a core cold spot area for spatial risk. A certain unit surrounding a spoil disposal site has an LISA type of LL and a cold spot level of strong cold spot (P=0.007<0.01), and is determined to be a core cold spot area, belonging to the area with the highest ecological risk and the strictest control.

[0048] Generally, the determination of low-value cluster areas is as follows: =LL, but If the cold spot is weak or insignificant, it is classified as a general low-value cluster area. The edge unit of the construction road, with LISA type LL and cold spot level of weak cold spot (P=0.08), is classified as a general low-value cluster area with medium risk intensity and slightly lower control requirements than the core cold spot area.

[0049] Determination of core hotspot areas for ecological carrying capacity =HH (high-high agglomeration area), and If a hotspot is identified as a strong or medium hotspot, it is determined to be a core hotspot area for ecological carrying capacity. The native forest unit in the reservoir area, with a LISA type of HH and a cold hotspot level of medium hotspot (P=0.02), is determined to be a core hotspot area, representing the region with the best ecological conditions and requiring key protection.

[0050] Generally, the determination of high-value agglomeration areas is as follows: =HH, but If the hot spot is weak or insignificant, it is classified as a general high-value cluster area. Forest units slightly farther from the reservoir area, with a LISA type of HH and a cold / hot spot level of weak hot spot (P=0.09), are classified as general high-value cluster areas with relatively good ecological conditions, requiring only routine protection.

[0051] Low-value anomaly zone determination If the value is LH (Low-High Anomaly Zone), it is identified as a low-value anomaly zone, which can be considered as a risk diffusion front or a local disturbance patch. The boundary unit between forest land and spoil heap has low carrying capacity but is surrounded by high-value zones. The LISA type is LH, which is identified as a low-value anomaly zone. It is a key frontier area for the diffusion of ecological risks to high-quality areas.

[0052] High-value anomaly zone determination =HL (High-Low Anomaly Zone) indicates a high-value anomaly zone, which can be considered as an area with potential for ecological restoration or a high-carrying-capacity isolated area. The small remaining woodland units in the construction area have high carrying capacity but are surrounded by low-value zones. The LISA type is HL, which indicates a high-value anomaly zone and is a key potential patch for ecological restoration.

[0053] General monitoring area or background area determination if If the LISA type is NS (Not Significant), it is classified as a general monitoring area or background area. Flat grassland units far from the construction area, with a LISA type of "Not Significant," are classified as general monitoring areas, showing no obvious aggregation or abnormal characteristics, and indicating a low risk level.

[0054] Through the above fusion and discrimination, this step unifies spatial similarity, spatial anomaly, and the salience of hot and cold spots, providing a refined basis for spatial risk types for subsequent engineering disturbance correlation analysis, risk contribution assessment, and five-level zoning control.

[0055] S104 involves spatially overlaying and buffering spatial risk types with engineering disturbance units, calculating the cold spot correlation ratio, low-low clustering correlation ratio, risk intensity, and risk area ratio, constructing an engineering disturbance unit risk contribution index, and identifying key risk contribution units.

[0056] In one implementation, a triple analysis mechanism of spatial overlay, buffer generation, and neighborhood association strength is constructed based on a spatial risk type layer, engineering disturbance unit vector data, and distance threshold parameters. Using the disturbance unit's influence area as the analysis scope, four quantitative indicators are extracted: cold point association ratio, low-to-low cluster association ratio, average risk intensity of the disturbance area, and risk area proportion. Based on the spatial risk type layer, engineering disturbance unit vector data, and preset distance thresholds, a triple analysis mechanism of spatial overlay, buffer generation, and neighborhood association strength is established. First, spatial overlay matches the positional relationship between risk types and disturbance units. Then, based on the distance thresholds, a buffer zone surrounding the disturbance unit is generated as the influence range. Finally, the neighborhood association strength between the risk unit and the disturbance source within the buffer zone is calculated, forming a complete association analysis link. Taking a hydropower project spoil disposal site as an example, the vector boundary of the spoil disposal site is overlaid with the ecological carrying capacity spatial risk map. A reasonable influence distance is set to generate a buffer zone. The correlation between cold points, low-value clusters, and the spoil disposal site within the buffer zone is analyzed to ensure that the analysis scope closely matches the actual disturbance influence boundary.

[0057] Within the affected area of ​​a disturbance unit, the system extracts four core quantitative indicators: first, the cold point correlation ratio, which is the proportion of cold point areas within the buffer zone; second, the low-to-low clustering correlation ratio, which is the proportion of low-value clustered areas within the buffer zone; third, the average risk intensity of the disturbance zone, which is the average risk level of all risk units within the buffer zone; and fourth, the risk area ratio, which is the proportion of the total risk area within the buffer zone. For example, in the buffer zone of a construction road, the cold point correlation ratio reflects the proportion of low-value cold points around the road, the low-to-low clustering correlation ratio reflects the degree of contiguous low-value areas along the road, the average risk intensity characterizes the overall risk level, and the risk area ratio quantifies the size of the risk impact range.

[0058] In another implementation, to enable the hot and cold spot identification results to serve on-site engineering management, this application further correlates the spatial risk types of ecological carrying capacity with engineering disturbance units, quantifying the contribution of different disturbance sources to ecological risk. Let the set of engineering disturbance units be... ,in, It can represent construction roads, spoil heaps, tunnel entrances, mixing systems, construction platforms, riverbank disturbance zones, etc.

[0059] For each type of engineering disturbance unit To construct its area of ​​influence ,in, spatial unit To engineering disturbance unit distance, Distance threshold for disturbance impact. Point disturbances such as those at tunnel entrances and mixing systems: Take 300m; Linear disturbances such as construction roads: unilateral Take 500m; patchy disturbances such as spoil heaps and construction platforms: Take 800m; corridor-like disturbances such as riverbank disturbance zones: unilateral Take 1000m. (Referring to a certain spoil heap) For example, set If the distance is 800m, then all spatial units within 800m of the boundary of the spoil disposal site belong to its influence zone. The total area of ​​this region is approximately 3.2 km². Calculate the... The cold point correlation ratio corresponding to the engineering disturbance unit. ,in: The hot and cold spot types of the i-th spatial unit; The affected area The total number of spatial units within the affected area; the numerator is the number of units belonging to cold points within the affected area. A construction road's affected area. It contains 800 spatial units, of which 280 units are classified as cold points. The correlation ratio of these cold points is... =280 / 800=0.35, indicating that 35% of the units within the affected area of ​​the construction road are cold spots.

[0060] Calculate the first The low-to-low clustering correlation ratio corresponding to engineering disturbance units ,in, Let represent the local clustering type of the i-th spatial unit; LL represents a low-to-low clustering region; and the numerator is the number of units within the affected area belonging to the low-to-low clustering region. The aforementioned construction road affected area... In the example, 360 units have an LISA type of LL, indicating a low-to-low clustering association ratio: =360 / 800=0.45, indicating that 45% of the units within the construction road impact area are low-value contiguous clusters. The cold spot association ratio and the low-low cluster association ratio quantify the degree of correlation between engineering disturbance units and ecological risks from two dimensions: the significance of cold and hot spots and local clustering. This provides key quantitative basis for the subsequent construction of risk contribution indices and the identification of key disturbance units.

[0061] After extracting four types of indicators—cold spot correlation ratio, low-low clustering correlation, risk intensity, and risk area proportion—a single-level linear weighted model was established, following three criteria: risk correlation, disturbance contribution, and indicator balance. This model quantifies the comprehensive contribution of different engineering disturbance units to ecological risk, ensuring balanced explanatory power and reliable results. The model construction criteria are as follows: Prioritize indicators directly related to ecological risk and with clear physical meaning; weights are positively correlated with correlation strength, highlighting core risk drivers. Emphasize the indicator's ability to characterize the intensity and scope of engineering disturbances; weights increase with the significance of disturbance impact, strengthening the differentiation of risk contributions from disturbance sources. Avoid excessively high or low weights for any single indicator; allocate weights reasonably to ensure balanced explanatory power of each dimension of indicators for risk contribution and prevent result bias.

[0062] The model adopts a single-level linear combination form, with four types of quantitative indicators as inputs and the risk contribution index of engineering disturbance units as outputs. The calculation formula is: In the formula, The risk contribution index for the qth type of engineering disturbance unit ranges from 0 to 1, with a higher value indicating a higher risk contribution. The correlation ratio of cold spots; The proportion of low-to-low clustering associations; The average risk intensity within the affected area of ​​the engineering disturbance unit; The area of ​​the risk zone within the affected area of ​​the engineering disturbance unit; For weighting coefficients. When When the threshold is exceeded, the disturbance unit of the project is determined to be a key risk contribution unit.

[0063] Combining the three criteria, the weighting coefficients are assigned as follows: cold point correlation ratio. =0.4, percentage of risky area =0.3, low-low clustering correlation ratio =0.15, average risk intensity =0.15, balancing risk correlation, disturbance contribution, and indicator balance. A large spoil heap (patchy disturbance unit) is selected, and its influence area indicators are as follows: cold point correlation ratio. =0.75; Low-low clustering association ratio =0.68; average risk is high Percentage of risky area =0.82. Substitute into the model calculation, =0.4×0.75+0.15×0.68+0.15×3.2+0.3×0.82=0.3+0.102+0.48+0.246=1.128, normalized risk contribution index =0.94, classifying it as a high-contribution key risk unit, indicating that this spoil heap has a significant contribution to ecological risk and requires priority control. Next, ordinary construction roads (linear disturbance units) were selected, with the following indicators... ; calculated =0.4×0.35+0.15×0.45+0.15×2.1+0.3×0.50=0.14+0.0675+0.315+0.15=0.6725, after normalization =0.56, which is judged as a medium contribution general risk unit.

[0064] After the model is built, a triple verification is performed on indicator completeness, weight suitability, and index discrimination: For indicator completeness, the four types of indicators are verified to be without missing data and the data to be valid, and outliers are removed; for weight suitability, the weights are fine-tuned for point-like, linear, and patchy disturbances, such as point-like tunnel entrances. Increase to 0.45, linear construction road The index value was increased to 0.35. Regarding index discrimination, the model verified that the differences in indices between different disturbance units were significant (e.g., the difference between the indices of a spoil heap and a construction road is ≥0.3), enabling effective classification. After verification, the model adapts to the risk quantification needs of different disturbance types, ensuring that the results can be directly used for the identification and control decisions of key risk units. This linear weighted model considers the directness of risk correlation, the significance of disturbance impact, and the balance and rationality of indicators. It can quantify the differences in ecological risk contribution of different engineering disturbance units, accurately identify high-contribution key risk units, medium-contribution general risk units, and low-contribution low-risk units, providing a quantitative basis for subsequent risk tracing and control priority classification, and supporting differentiated engineering control.

[0065] After constructing the engineering disturbance risk contribution model, a systematic verification is required to ensure that the indicators are complete, the weights are reasonable, and the classification is reliable, providing a stable quantitative basis for subsequent management and control. Verification includes three aspects: indicator completeness, weight suitability, and index discrimination. Based on the verification results, parameters are calibrated to form a classification result that can be directly used for the project. The purpose of indicator completeness verification is to ensure that the four types of indicators—cold spot correlation ratio, low-low clustering correlation ratio, average risk intensity of the disturbance area, and risk area proportion—are complete, the data are valid, and there are no missing or abnormal values. All disturbance units have complete indicator records, the values ​​are within a reasonable range, the correlation ratio is between 0 and 1, the risk intensity is an integer from level 1 to 5, and there are no null values ​​or outliers. In the impact area of ​​a hydropower project construction road, the cold spot correlation ratio is 0.35, the low-low clustering correlation ratio is 0.45, the average risk intensity is 2.1, and the risk area proportion is 0.50. All four types of indicators are complete and the values ​​are reasonable, thus passing the completeness verification.

[0066] The purpose of weight suitability verification is to confirm whether the weight allocation accurately reflects the risk distribution characteristics of point-like, linear, and patchy disturbances, avoiding weight imbalances and deviations from actual impact patterns. Point-like disturbances emphasize the proportion of cold-point associations, linear disturbances emphasize the proportion of risky areas, and patchy disturbances balance association and area, with weight deviations not exceeding ±0.05. The tunnel entrance represents a point-like disturbance, with original weights of [weight value missing]. =0.4、 =0.3; fine-tuned after verification. =0.45、 =0.25, highlighting the impact of concentrated point risks. The construction road represents a linear disturbance, with the original weight... =0.4、 =0.3; fine-tuned after verification. =0.35、 =0.35, suitable for linear wide-area risk diffusion characteristics. The purpose of index discrimination verification is to ensure that the contribution indices of different disturbance units are significantly different, clearly graded, and without overlapping intervals. The difference in indices between high, medium, and low contribution units is not less than 0.2, and the grading accuracy is not less than 95%. The contribution index of large spoil heaps is 0.94, ordinary construction roads are 0.56, and small temporary camps are 0.21. The difference between adjacent grades is greater than 0.2, the discrimination is good, and the verification is passed.

[0067] Based on the verification results, the weighting coefficients and grading thresholds are adjusted for different disturbance types: Point disturbance weights: =0.45、 =0.15、 =0.15、 =0.20; Classification thresholds: High contribution ≥ 0.8, Medium contribution 0.4~0.8, Low contribution < 0.4. Linear perturbation weights: =0.35、 =0.15、 =0.15、 =0.35; Classification thresholds: High contribution ≥ 0.7, Medium contribution 0.3~0.7, Low contribution < 0.3. Patch perturbation weight: =0.40、 =0.15、 =0.15、 =0.30; Classification thresholds: high contribution ≥0.8, medium contribution 0.4~0.8, low contribution <0.4. The calculation format of the model remains consistent after calibration. .

[0068] Based on the calibrated contribution index thresholds, disturbance units are divided into three levels: High-contribution key risk units: the index reaches or exceeds the high threshold, the risk contribution is significant, and the control priority is the highest. Medium-contribution general risk units: the index is in the medium threshold range, the risk contribution is moderate, and routine control is implemented. Low-contribution low-risk units: the index is below the medium threshold, the risk contribution is weak, and only simple monitoring is required. A large spoil heap with an area of ​​0.62 km² has a calculated contribution index of 0.94, which is higher than the patchy high threshold of 0.8, and is therefore identified as a high-contribution key risk unit, associated with the core cold spot area, and given priority for strict control. A typical construction road with a length of 4.8 km has a contribution index of 0.56, which is in the linear medium threshold range, and is identified as a medium-contribution general risk unit, and routine inspections are carried out. A small temporary camp with an area of ​​0.05 km² has a contribution index of 0.21, which is lower than the point medium threshold, and is identified as a low-contribution low-risk unit, requiring only periodic monitoring.

[0069] After classification, a list of engineering disturbance risk contributions is generated. The list includes the name of the disturbance unit, disturbance type, spatial area, risk contribution index, risk level, associated risk types, and control priority. The list clearly presents the risk differences of each disturbance unit, supporting differentiated control decisions for the project. For example, the No. 1 spoil heap, a patchy area of ​​0.62 km², has a contribution index of 0.94, a high risk level, and is associated with a core cold spot area, thus having the highest control priority. The R3 construction road, linear in shape and 4.8 km long, has a contribution index of 0.56, a medium risk level, and is associated with a general low-value cluster area, thus having a medium control priority. The small temporary camp, a point-like area of ​​0.05 km², has a contribution index of 0.21, a low risk level, no significant risk association, and has the lowest control priority.

[0070] S105 identifies dynamic changes in multi-period ecological carrying capacity results, determines the change types of newly added / continuous / receding cold spots and newly added / continuous / degraded hot spots, quantifies the expansion rate and decline rate of cold spots, and outputs a time-series change map.

[0071] In one implementation, after completing the spatial risk identification and engineering disturbance correlation analysis of single-phase ecological carrying capacity, further identification of temporal dynamic changes is carried out. This requires first constructing a multi-phase analysis dataset that is temporally aligned, spatially unified, and clearly defined in stages. Based on the temporally aligned multi-phase ecological carrying capacity results, dynamic change identification is conducted, and standardized dynamic analysis data is generated. The input data includes ecological carrying capacity raster data for three or more consecutive phases, phase-by-phase hot and cold spot type maps, and spatial risk type layers. Simultaneously, it is associated with engineering stage identifier data from before construction, during construction, and the initial operation phase to ensure that the temporal data is continuous in the time dimension, unified in the spatial dimension, and clearly defined in stage attributes. Ecological carrying capacity data from before construction, during construction, and the initial operation phase are selected to cover the complete construction phase: Before construction: the period with the weakest engineering disturbance and stable ecological background; During construction: the period of concentrated disturbances such as spoil heaping, tunnel excavation, and road construction; Initial operation: the period of completion of the main project and gradual ecological recovery.

[0072] All three phases of data were generated using a 30m resolution raster, CGCS2000 coordinate system, and UTM projection. Coordinate unification, pixel matching, boundary clipping, and invalid data removal were completed. Coordinate unification: All three phases of data were converted to the CGCS2000 national geodetic coordinate system, with a projection error of ≤0.1m. Pixel row and column numbers were unified to ensure that each pixel in the three phases of data corresponds one-to-one without misalignment. The data was uniformly clipped to the hydropower project construction area, reservoir area, drawdown zone, and a surrounding 5km range, with a total area of ​​approximately 800km². Invalid data removal: Approximately 12,000 pixels that were missing or abnormal in all three phases were removed to ensure that the effective range of the three phases of data was consistent.

[0073] Time-series registration was completed, with data collection for all three phases uniformly scheduled for October each year, ensuring a time error of ≤15 days and eliminating the impact of seasonal differences. Simultaneously, engineering stage identifier fields were associated with the data for each of the three phases: before construction: the field was assigned a value of "1", representing the ecological background period; during construction: the field was assigned a value of "2", representing the peak disturbance period; and in the initial stage of operation: the field was assigned a value of "3", representing the ecological restoration period.

[0074] The dynamic analysis input data is shown below: Ecological carrying capacity raster data for each period: 30m resolution, pixel value 0-1. Before construction, the forest area pixel value is 0.82; during construction, the spoil heap pixel value is 0.21; and at the beginning of operation, the same location pixel value is 0.35. Period-by-period hot and cold spot type map: raster attributes include strong cold, moderate cold, weak cold, strong hot, moderate hot, weak hot, and insignificant. During construction, the raster around the tunnel entrance is a strong cold spot (P=0.008). Spatial risk type layer: vector / raster attributes include core cold spots, general low values, core hot spots, risk frontier, and general monitoring. The raster at the edge of the construction road is a risk frontier area. Engineering stage identification data: vector attributes include before construction, during construction, and at the beginning of operation. The vector label of the construction road is a disturbance unit during construction. After the above-mentioned coordinate unification, pixel matching, temporal alignment, stage association, and invalid removal, a dynamic analysis basic dataset with three unified phases, spatial consistency, clear stages, and valid data is formed. It can be directly used to identify dynamic changes such as newly added cold spots, persistent cold spots, fading cold spots, newly added hot spots, persistent hot spots, and hot spot degradation, supporting the dynamic assessment of ecological risks throughout the entire project process.

[0075] A typological evolution assessment was conducted on time-aligned multi-period data, generating seven change types: newly added cold spots, persistent cold spots, receding cold spots, newly added hot spots, persistent hot spots, hot spot degradation, and hot spot turning into cold spots. Four evolutionary characteristics of cold and hot spots—expansion, contraction, migration, and receding—were identified. Based on time-series alignment, spatial unification, and data standardization, a pixel-by-pixel typological evolution assessment was conducted on ecological carrying capacity cold and hot spot data from three phases: pre-construction, mid-construction, and early-operation. Dynamic change patterns of cold and hot spots were identified through before-and-after comparisons, clarifying the evolutionary patterns of ecological carrying capacity with engineering stages. The differences in cold and hot spot types and spatial risk types between adjacent periods (pre-construction to mid-construction, mid-construction to early-operation) were compared pixel-by-pixel to determine the evolution type. Let the first... Each spatial unit in time period and The hot and cold hot types are as follows: Based on the type of change, the following are identified: Newly added cold spot area: Not a cold spot in the previous period, but becomes a cold spot in the next period; Continuous cold spot area: Both the previous and next periods are cold spots; Fading cold spot area: A cold spot in the previous period, but no longer a cold spot in the next period; Newly added hot spot area: Not a hot spot in the previous period, but becomes a hot spot in the next period; Continuous hot spot area: Both the previous and next periods are hot spots; Hot spot deterioration area: A hot spot in the previous period, but becomes a non-hot spot or a cold spot in the next period.

[0076] At the same time, four types of overall evolution characteristics are summarized: expansion: new cold spots are added in clusters and their range is expanded; contraction: cold spots disappear in clusters and their range is reduced; migration: the overall position of cold and hot spots moves; and decline: the area of ​​hot spots decreases and cold spots weaken.

[0077] Examples of newly added cold spots are as follows: Before construction, the surrounding fence of the spoil disposal site had a bearing capacity of 0.72, making it a medium-hot spot (P=0.03); during construction, the same location had a bearing capacity of 0.21, making it a strong cold spot (P=0.008); these were determined to be newly added cold spots, corresponding to the overall evolution and expansion.

[0078] Examples of persistent cold spots are as follows: Before construction, the bearing capacity around the tunnel entrance was 0.19, indicating a strong cold spot (P=0.007); during the middle of construction, the bearing capacity at the same location was 0.22, indicating a strong cold spot (P=0.009); it was determined to be a persistent cold spot with stable evolution characteristics.

[0079] An example of a receding cold point is as follows: During the middle of construction, a slope has a bearing capacity of 0.28, which is a weak cold point (P=0.08); at the beginning of operation, the same location has a bearing capacity of 0.55, which is not significant (P=0.22); it is determined to be a receding cold point, and its evolution characteristic is shrinkage.

[0080] Examples of hotspot degradation are as follows: Before construction, the edge of the reservoir area had a bearing capacity of 0.81, indicating a strong hotspot (P=0.006); at the same location during the initial operation period, the bearing capacity was 0.63, which was not significant (P=0.18); this was determined to be hotspot degradation.

[0081] The following is an example of a hot spot turning into a cold spot: Before construction, the riverbank area had a bearing capacity of 0.78, making it a medium hot spot (P=0.02); during the middle of construction, the same location had a bearing capacity of 0.25, making it a medium cold spot (P=0.04); it was determined to be a hot spot turning into a cold spot, with the evolution characteristics of migration and expansion.

[0082] The formula for calculating the cold point expansion rate is: ,in, Indicates the newly added cold spot area. This represents the total area of ​​cold spots in the previous period. The area of ​​cold spots before construction was 12 km², and the area of ​​newly added cold spots during the construction period was 4.8 km². The expansion rate is 4.8 / 12 = 40%. The formula for calculating the cold spot decay rate is as follows: The cold spot area during the construction phase was 16.8 km², and it receded by 2.52 km² during the initial operation phase, with a receding rate of 2.52 / 16.8 = 15%. Through the above pixel-by-pixel comparison, type determination, feature summarization, and quantitative calculation, the evolution pattern of hot and cold spots in ecological carrying capacity during hydropower project construction disturbance and ecological restoration can be clearly identified, providing a direct basis for dynamic management and restoration sequence arrangement.

[0083] Simultaneously, a triple verification was conducted on the completeness of time-series data, the distinguishability of evolution types, and the accuracy of indicator calculations to ensure the reliability of the indicators and their applicability for dynamic monitoring during the engineering phase. Specifically, for the verification of time-series data completeness: the ecological carrying capacity data and hot and cold spot data for the three phases of pre-construction, mid-construction, and initial operation were checked to ensure they were complete, with no missing measurements or blank values, and the raster data completely covered the study area. All three phases of data had no missing pixels, and the hot and cold spot layers were complete, thus passing the completeness verification.

[0084] For the differentiation of evolution types, the seven evolution types—newly added cold spots, persistent cold spots, receding cold spots, newly added hot spots, persistent hot spots, and hot spot degradation—were verified to have clear boundaries and no overlap, with a differentiation accuracy of ≥95%. Newly added cold spots and receding cold spots had no overlapping areas, demonstrating good differentiation. The accuracy of indicator calculations was verified: the area statistics accuracy (error ≤5%), correct formula application, and reasonable numerical range (0–1) were checked. The cold spot area statistics error was 3%, and the expansion rate, receding rate, and degradation rate were all within a reasonable range, passing the accuracy verification. The three types of time-series indicators can be directly used for dynamic engineering analysis: a high cold spot expansion rate indicates increased construction disturbance, requiring strengthened control; a satisfactory cold spot receding rate indicates effective ecological restoration; and an abnormal hot spot degradation rate requires key protection of high-quality ecological areas, providing a quantitative basis for dynamic management throughout the entire process.

[0085] Based on temporal evolution types and spatial unit change characteristics, a quantitative calculation model is constructed to generate three core temporal indicators. The model is based on periodic spatial unit changes, combining area statistics and temporal difference calculations to output three indicators: cold spot expansion rate, cold spot regression rate, and hot spot degradation rate. Simultaneously, multi-dimensional verification is conducted, including verification of temporal data integrity, evolution type differentiation, and indicator calculation accuracy, to ensure that the indicators reliably match the dynamic monitoring needs of the engineering phase. The ratio of newly added cold spot area during construction to the total cold spot area before construction is calculated to determine the cold spot expansion rate. During verification, missing temporal data is removed, similar evolution types are distinguished, and the accuracy of area calculations is checked to ensure that the indicators reflect true dynamic changes.

[0086] After completing the determination of temporal evolution types, quantification of temporal indicators, and multi-dimensional verification, three core data categories—evolution types, temporal indicators, and engineering stage characteristics—are integrated to construct a standardized temporal change map. This map visually presents the spatiotemporal evolution patterns of hot and cold spots in ecological carrying capacity in a spatial visualization format, providing visual support for the dynamic management and control of the entire hydropower project process. The integrated data includes three categories, as follows: Evolution type data: seven pixel-by-pixel evolution results for newly added cold spots, persistent cold spots, receding cold spots, newly added hot spots, persistent hot spots, hot spot degradation, and hot spot to cold spot transformation; Temporal indicator data: three quantitative indicators for cold spot expansion rate, cold spot receding rate, and hot spot degradation rate; Engineering stage data: three stage identifiers for pre-construction, mid-construction, and initial operation. All data are unified at a 30m raster resolution and the CGCS2000 coordinate system, with each pixel associated with evolution type, temporal indicator, and stage attributes to ensure spatiotemporal consistency.

[0087] Based on three spatial units, an evolution type layer, a time-series index grading layer, and an engineering stage boundary layer are overlaid to generate three types of standardized visualization results: a time-series evolution thematic map, an expansion / regression rate grading map, and a stage comparison map. For the evolution thematic map: different colors are used to distinguish seven evolution types: new cold spots are red, persistent cold spots are dark red, regressing cold spots are orange, and hot spots are degrading, visually indicating the spatial distribution of each type. For the time-series index grading map: expansion and regression rates are graded as 0–20%, 20–40%, and above 40%, with light blue, blue, and dark blue gradients showing the intensity of change. For the stage comparison map: engineering boundaries such as construction areas, reservoir areas, and spoil heaps are overlaid, marking the construction and recovery periods and linking the evolution areas with the locations of engineering disturbances.

[0088] The following is an example of a time-series change diagram for a hydropower project. During the construction phase, newly added cold spots are concentrated around the No. 1 spoil heap, covering an area of ​​4.8 km², marked in red. The area around the tunnel entrance is a persistent cold spot, covering an area of ​​2.1 km², marked in dark red. The slope area is a receding cold spot, covering an area of ​​2.52 km², marked in orange. The cold spot expansion rate during construction is 40% (dark blue), and the receding rate during operation is 15% (light blue). The newly added cold spots completely coincide with the boundaries of the construction road and spoil heap, while the receding cold spots are consistent with the ecological restoration area, intuitively reflecting the spatiotemporal relationship between disturbance and restoration. A standardized time-series change diagram can clearly show, specifically: during the construction phase: the expansion range of cold spots and the location of newly added risk areas, supporting dynamic adjustments to construction boundaries; during the operation phase: the receding areas of cold spots and the effectiveness of ecological restoration, optimizing the restoration sequence; throughout the entire process: the migration and expansion / receding trends of cold and hot spots, supporting the prediction of risks in cascade development. The results can be directly connected to the engineering management platform, providing a visual basis for dynamic inspections, adjustment of restoration priorities, and selection of new project sites, so as to realize that ecological risks are "visible, assessable, and manageable".

[0089] S106, based on spatial risk type, the significance of hot and cold spots, the correlation strength of engineering disturbances and multi-period change trends, delineates strict control areas, ecological restoration priority areas, risk diffusion prevention and control areas, key protection areas and general monitoring areas according to preset rules, outputs zoning boundaries, hierarchical distribution, risk source tracing results, and generates differentiated control recommendations.

[0090] In one implementation, control zones are automatically delineated based on local spatial clustering type, hot and cold spot significance, engineering disturbance correlation strength, and multi-period change trends, directly converting spatial statistical results into executable engineering control zones. Four core judgment criteria are integrated to construct a basic dataset for zone delineation. Input data includes spatial risk types of ecological carrying capacity, hot and cold spot significance levels, engineering disturbance correlation strength thresholds, and multi-period hot and cold spot change trends. All four types of data use a unified 30m raster resolution and CGCS2000 coordinate system to ensure matching and overlay analysis requirements. Taking a hydropower project as an example, risk types such as strong / medium / weak hot and cold spots and core cold spot areas, strong / medium / weak significance levels, high / medium / low disturbance correlation strength, and temporal trends such as expansion during construction and decline during recovery are integrated to form unified basic data for zone determination.

[0091] Based on the pre-set five-level zoning rules, spatial characteristics and evolution trends are matched item by item to accurately delineate five types of control zones. Strict control zone determination criteria are applied (meeting any one of the following is sufficient): LISA type is LL, and the hot / cold spot type is strong cold spot or moderate cold spot; the cold spot area highly overlaps with engineering disturbance units such as construction roads, spoil heaps, tunnel entrances, and mixing systems; it shows as a persistent cold spot or a newly enhanced cold spot in multi-period analysis; the risk contribution index of engineering disturbance units exceeds the threshold. Control positioning: Areas with significant overlap of low-value ecological carrying capacity and construction disturbance are identified as the highest-risk areas. Areas with strong cold spot low-value aggregation around spoil heaps, and where the disturbance correlation intensity exceeds the standard, are designated as strictly controlled zones.

[0092] Criteria for determining priority ecological restoration areas (meeting any one of the following is sufficient): Located in low-to-low agglomeration areas or cold spot areas, but already in the recovery phase; located around repairable engineering units such as spoil heaps, construction platforms, and slope protection zones; transitioning from strong cold spot to weak cold spot, showing a recovery trend; located on the edge of a risk cold spot, possessing value in preventing the spread of risk. Management and positioning: Areas with ecological restoration potential are selected and prioritized for restoration. Areas on the edge of weak cold spots around tunnel entrances, where multiple data points show a receding trend of the cold spot, are designated as priority ecological restoration areas.

[0093] The criteria for determining a risk diffusion control zone (meeting any one of the following is sufficient): LISA type is LH or HL; located in the transition zone surrounding cold spots; located at the boundary between low-value and high-value clustering areas; cold spots show an expanding trend in multi-period analysis. Control positioning: To prevent the expansion of low-bearing-capacity patches to the surrounding areas, this is a key area for risk diffusion. The abnormally high / low transition zone along the construction road, where cold spots continue to expand during the construction period, is designated as a risk diffusion control zone.

[0094] Criteria for determining key protected areas (meeting any one of the following is sufficient): LISA type HH; hotspot type strong or medium hotspot; consistently a hotspot in multiple analysis periods; high ecological carrying capacity and proximity to areas affected by engineering disturbances. Control positioning: Identify ecologically sound and stable areas and restrict the spread of construction disturbances. A contiguous high-value cluster area surrounding the reservoir, consistently a strong hotspot in multiple analysis periods, is designated as a key protected area.

[0095] Criteria for determining a general monitoring area: Areas that do not meet the above criteria and whose spatial statistical results are not significant or have low risk. Control positioning: Routine control areas, primarily for routine monitoring. Woodlands far from construction areas, without hotspots or disturbance associations, are designated as general monitoring areas.

[0096] Based on the risk characteristics and control needs of the five zones, differentiated control measures are precisely matched, as detailed below: For strictly controlled areas: priority should be given to measures such as construction boundary control, temporary protection, spoil disposal, water and sediment control, and ecological restoration. For example, in strictly controlled areas surrounding spoil heaps, it is recommended to reduce the construction scope, temporarily cover exposed slopes, and strictly control further disturbance.

[0097] For priority ecological restoration areas: topsoil backfilling, vegetation restoration, slope protection, drainage works, and soil and water conservation measures should be prioritized. For example, in priority restoration areas around tunnel entrances, it is recommended to plant native vegetation, improve drainage systems, and accelerate ecological restoration.

[0098] For risk-prevention and control zones: Strengthen dynamic monitoring and boundary protection. It is recommended to set up buffer zones, cover bare land, protect riverbanks, and conduct dynamic patrols. For example, in control zones along construction roads, it is recommended to set up vegetated buffer zones and regularly patrol bare land to block the risk spread path.

[0099] For key protected areas: Limit the expansion of construction disturbance, maintain the original ecological structure and soil and water conservation functions, and it is recommended to add disturbance restrictions, protect native vegetation, and require construction detours. For example, in key protected areas of reservoir areas, it is recommended to prohibit new construction land occupation, require construction roads to detour, and maintain ecological stability.

[0100] For general monitoring areas: routine ecological monitoring should be the primary method, along with periodic remote sensing monitoring and regular ground patrols. For example, in forest areas far from construction sites, it is recommended to conduct remote sensing monitoring quarterly and ground patrols annually, implementing routine management and maintenance.

[0101] S107 integrates global spatial autocorrelation results, local clustering maps, hot and cold spot distribution maps, spatial risk maps, engineering disturbance risk contribution lists, time series change maps, zoning boundaries, hierarchical attributes, risk tracing and differentiated management suggestions to form the results of hot and cold spot identification and zoning management of ecological carrying capacity of hydropower projects.

[0102] In one implementation, based on nine core output data categories—global spatial autocorrelation results, local clustering diagrams, hot and cold spot distribution maps, spatial risk maps, engineering disturbance risk contribution lists, time-series variation maps, partition boundaries, hierarchical attributes, and risk tracing—layer standardization, field mapping, and format adaptation processing are performed. After completing global spatial autocorrelation diagnosis, local clustering identification, hot and cold spot statistics, risk fusion, disturbance correlation, and time-series dynamic analysis, layer standardization, field mapping, and format adaptation processing are carried out based on these nine core data categories. This unifies spatial representation, field definitions, and storage formats, eliminates matching barriers caused by data heterogeneity, and ensures that the results can be directly used for integration, statistical analysis, and visualization on engineering management platforms.

[0103] For raster, vector, and tabular result layers, a unified spatial reference, resolution, and representation rule were established. Specifically, all layers were uniformly converted to the CGCS2000 National Geodetic Coordinate System and UTM projection, with coordinate errors controlled within 0.1m to avoid spatial misalignment caused by projection deviations. Raster results (hot and cold spot distribution maps, local cluster maps, and temporal variation maps) uniformly adopted a 30m resolution with pixel row and column alignment to ensure accurate pixel-by-pixel temporal comparison and spatial overlay. Vector results (zoning boundaries and disturbance unit boundaries) were uniformly adapted to a 1:10,000 spatial accuracy to ensure accurate matching of area boundaries, linear corridors, and point disturbance locations. All layers were uniformly cropped to the hydropower project construction area, reservoir area, drawdown zone, and a surrounding 5km influence range, totaling approximately 800km², eliminating irrelevant areas to ensure that the results focused on the core area of ​​project management. The original hot and cold spot grids were in the WGS84 coordinate system with a resolution of 10m. After standardization, they were converted to CGCS2000 with a resolution of 30m. After row and column alignment, they were matched pixel by pixel with the local clustering map. The spatial misalignment error was less than 0.5 pixels, which met the accuracy requirements of the overlay analysis.

[0104] For attribute fields of various results, unified naming, data types, and value rules have been implemented to eliminate field ambiguity. Specifically, the following fields have been added to cold and hot spot distribution maps: "Cold and Hot Spot Level (Integer: 1-Strong Cold, 2-Medium Cold, 3-Weak Cold, 4-Weak Hot, 5-Medium Hot, 6-Strong Hot, 0-Not Significant)" and "Significance P-value (Floating Point: 0~1)". The "Clustering Type (Character: HH / LL / HL / LH / NS)" field has been added to local clustering maps. The "Zoning Type (Character: Strictly Controlled Area / Ecological Restoration Priority Area / Risk Diffusion Prevention and Control Area / Key Protected Area / General Monitoring Area)" and "Risk Level (Integer: 1-High / 2-Medium / 3-Low)" fields have been added to zoning boundary layers. The "Disturbance Type (Character: Point / Linear / Patch / Corridor)" and "Risk Contribution Index (Floating Point: 0~1)" fields have been added to engineering disturbance unit layers. The engineering disturbance risk contribution list has been standardized with the addition of "Disturbance Unit Name (character)" and "Associated Risk Type (character)" fields, and numerical fields are now uniformly retained to two decimal places. The original "Contribution Value" field (floating point, 0~1.0) in the risk contribution list has been standardized and mapped to "Risk Contribution Index (floating point, 0.00~1.00)" to unify precision and avoid value confusion; the "Control Level" field of the partition boundary has been standardized and mapped to "Risk Level (integer: 1-High, 2-Medium, 3-Low)" to correspond to the classification threshold.

[0105] Heterogeneous data types—raster, vector, and tabular—are uniformly converted into a standard engineering geographic data format. Specifically, hotspot and cold spot distribution maps, local cluster maps, and time-series variation maps are uniformly converted to GeoTIFF format, using LZW lossless compression, with pixel values ​​stored as integers for easy reading and visualization rendering by GIS software. Zoning boundaries and disturbance unit boundaries are uniformly converted to Shapefile format, including complete .shp / .shx / .dbf / .prj files, with projection files matching the CGCS2000 coordinate system, supporting topology verification and boundary editing. The engineering disturbance risk contribution list is uniformly converted to Excel (.xlsx) format, with fixed field order and consistent numerical precision, supporting data filtering, sorting, and statistical summarization. The original hotspot and cold spot raster data, originally in ENVI format, is converted to GeoTIFF and can be directly loaded and displayed in commonly used engineering software such as ArcGIS and QGIS. The risk contribution list, converted from CSV format to Excel, retains field mapping relationships, supports filtering of high-contribution disturbance units, and adapts to engineering statistical needs.

[0106] After format adaptation, compatibility checks were conducted, including spatial matching checks: raster and vector layers were overlaid with a boundary overlap error of ≤1 30m pixel, with no spatial misalignment; field integrity checks were performed: all required fields had no empty or outlier values, and values ​​conformed to standardization rules; and format usability checks were performed: the converted data could be read, edited, and exported normally by commonly used GIS software and management platforms. After passing these checks, the nine core data categories achieved full spatial, field, and format compatibility, eliminating heterogeneous data matching barriers and providing a standardized data foundation for subsequent results integration, dynamic updates, and engineering applications.

[0107] The output data is categorized, organized, and matched according to five dimensions: spatial statistics, risk identification, disturbance correlation, dynamic evolution, and zoning control. After completing layer standardization, field mapping, and format adaptation, all standardized data is systematically categorized, organized, and matched according to the five core dimensions of spatial statistics, risk identification, disturbance correlation, dynamic evolution, and zoning control. This achieves data aggregation within the same dimension and data linkage across dimensions, ensuring clear data logic, convenient access, and explicit supporting relationships. It provides a standardized data organization framework for subsequent results integration, query retrieval, and visualization.

[0108] The system aggregates global and local spatial analysis results, including global Moran's I statistical results (index value, Z-value, P-value); local Moran's I clustering maps (HH / LL / HL / LH / NS type rasters); and spatial weight matrix parameter files (weight coefficients, neighborhood thresholds). The global Moran's I index is 0.72, Z-value is 12.5, and P-value is <0.01. In the local clustering maps, LL pixels around the spoil heap account for 65%, and these are uniformly included in the spatial statistical dimensions for overall clustering and local clustering queries.

[0109] Data collected by combining hot and cold spot statistics with spatial risk data includes: Getis-OrdGi Distribution maps of hot and cold spots (strong / medium / weak cold, strong / medium / weak hot grids); spatial risk type map (core cold spots, general low values, core hot spots, risk frontier grids); statistical table of the significance of hot and cold spots (area and percentage of each level). The strong cold spot grid around the tunnel entrance has a value of P=0.008 and an area of ​​0.25 km²; the core cold spot area around the spoil heap has an area of ​​1.8 km², and is included in the risk identification dimension for risk level query and location.

[0110] The results of the correlation between engineering disturbances and risks were collected, including: vector data of engineering disturbance units (spoil disposal sites, construction roads, tunnel entrance boundaries); a disturbance correlation index table (cold spot correlation ratio, low-low clustering correlation ratio, average risk intensity, risk area percentage); and a list of engineering disturbance risk contributions (unit name, contribution index, risk level). Spoil disposal site No. 1 has a cold spot correlation ratio of 0.75, a risk contribution index of 0.94, and a high contribution level, and is included in the disturbance correlation dimension for risk tracing and key disturbance identification.

[0111] Data from multiple time-series periods was collected, including: a three-phase time-series comparison grid (hot and cold spots and risk types before / during / after construction); tables of indicators for cold spot expansion rate, recession rate, and hot spot degradation rate; and an evolution type distribution map (newly added cold spots, persistently added cold spots, and receding cold spots grids). The cold spot expansion rate during the construction period was 40%, and the recession rate during the operation period was 15%. Newly added cold spots were concentrated along the construction road and were included in the dynamic evolution dimension for trend analysis throughout the entire process.

[0112] The final zoning results and source tracing data are collected, including: five-level zoning boundary vectors (strict control, ecological restoration, risk prevention and control, key protection, and general monitoring); zoning attribute tables (zoning area, risk level, and associated disturbances); and risk source tracing reports (zoning causes and control basis). The total area of ​​the strict control zone is 2.1 km², associated with the No. 1 spoil disposal site and the tunnel entrance, and is included in the zoning control dimension to support control decision-making queries.

[0113] Centered on the risk-disturbance-time series-regional logical chain, a cross-dimensional data linkage relationship is established to ensure data traceability, correlation, and cross-analysis: spatial statistics. Risk identification: Locally clustered LL grids are spatially superimposed with cold and hot spot cold grids. Areas with consistent matching are identified as low-risk core areas. The LL grid area is 3.2 km², of which 2.5 km² is a medium-to-cold area, which is identified as a general low-risk area.

[0114] Risk Identification Disturbance Correlation: The distribution maps of hot and cold spots are overlaid with the engineering disturbance buffer zone. The area of ​​cold spots within the buffer zone is extracted, and the correlation ratio of cold spots is calculated. The cold spot area in the 500m buffer zone of the construction road is 0.84km², and the total buffer zone area is 2.4km². The correlation ratio of cold spots = 0.84 / 2.4 = 0.35, thus establishing the correlation between risk and disturbance.

[0115] Disturbance correlation Dynamic evolution: The superposition of high-contribution disturbance units with newly added cold spot areas identifies newly added risk areas due to disturbance. The No. 1 spoil disposal site (contribution index 0.94) overlaps with the newly added cold spot (4.8km²) during the construction period by 85%, and is identified as the main cause of the new risk.

[0116] Dynamic evolution Zoned Control: The area with the overlapping of a persistent cold spot and a high-contribution disturbance area is designated as a strictly controlled zone. The persistent cold spot (2.1km²) at the tunnel entrance, along with the high-contribution associated area, is designated as a strictly controlled zone.

[0117] By classifying data across five dimensions and linking it across dimensions, unified management of data within the same dimension is achieved, facilitating quick retrieval of similar results; cross-dimensional data linkage allows for tracing back from partitions to the root causes of risks, disturbances, and time series; the data logic is clear and adaptable to the needs of engineering management platforms for invocation, visualization, and report generation.

[0118] After completing the data classification and organization across five dimensions, five core parameters—spatial coordinate system I, projection transformation, topology verification, field integrity verification, and temporal alignment quality control—were further added. These parameters were used to conduct comprehensive quality checks on all standardized data, ensuring accuracy and reliability at the spatial, attribute, and temporal levels. Unqualified data was removed, and anomalies were corrected, providing high-quality data support for subsequent results integration, analysis, and application.

[0119] For spatial coordinate system and projection transformation verification, ensure the spatial reference of all raster and vector data is unified, and eliminate projection deviation and coordinate misalignment issues. All data is unified to the CGCS2000 National Geodetic Coordinate System and UTM projection, and the consistency of projection parameters is checked, controlling the coordinate transformation error to ≤0.1m. The original construction road vector was in the WGS84 coordinate system. After projection transformation, it is superimposed on the CGCS2000 reference zone boundary, with a misalignment error ≤0.08m, meeting the accuracy requirements. After reprojection of the original 10m resolution hot and cold spot raster, the row and column alignment deviation is <0.5 30m pixels, and the verification is qualified. Unqualified data (such as coordinate errors and missing projection parameters) are directly discarded, and the reference transformation is re-matched.

[0120] For vector data topology rationality verification, ensure that the topology of vector data such as zoning and disturbance boundaries is correct, without errors such as overlap, gaps, or suspended arcs. The focus is on checking three types of vectors: zoning boundaries, engineering disturbance unit boundaries, and ecological restoration zone boundaries, identifying overlap, gaps, suspended points, and self-intersection issues. The topology error rate must be ≤5%. After topology checking, only two gaps with a length <1m were found in the fifth-level zoning boundary vectors; after correction, there were no overlaps or suspended arcs. The spoil heap boundary vectors had no self-intersections or nesting errors, and the topology rationality met the standards. Overlapping zoning boundaries were directly trimmed and corrected, and suspended arcs were automatically cleaned up.

[0121] For field completeness and validity checks, ensure that all key data fields are complete, numerically compliant, and free of null or outlier values. Specifically, for raster data: hot / cold spot level, cluster type, and time-series evolution fields have no null values ​​and the values ​​conform to the coding rules; for vector data: partition type, risk level, and disturbance type fields are complete and without missing values; for tables: numerical fields such as risk contribution index and correlation ratio have no null values ​​and the values ​​are within a reasonable range of 0 to 1. The engineering disturbance risk contribution list has 28 records. The weights η1 to η4 and the contribution index RC fields have no null values. RC values ​​range from 0.21 to 0.94, with no negative values ​​and no outliers greater than 1. The HH / LL / HL / LH / NS codes in the local clustering diagram are complete and without missing fields. List records containing null values ​​are directly filled in, and outliers (such as RC=1.2) are removed and recalculated.

[0122] For time-series alignment consistency verification, it was ensured that the time nodes of the data for the three phases of pre-construction, mid-construction, and initial operation were matched and the time series could be compared. The data for the three phases were uniformly collected in October of each year, with a time error of ≤15 days. The bearing capacity values ​​and hot / cold spot types of the three phases were checked pixel by pixel, and no sudden changes or anomalies were found. The time error of the data for the three phases of pre-construction (October 2022), mid-construction (October 2023), and initial operation (October 2024) was ≤12 days. The bearing capacity values ​​of the reservoir area forest land in the three phases were 0.81, 0.79, and 0.82, with fluctuations of ≤0.03 and no sudden changes or anomalies. The tunnel entrance was a strong cold spot in all three phases, and the time series logic was consistent. Data with a time error of more than 15 days was removed, and pixels with sudden changes or anomalies were checked and corrected.

[0123] After full-dimensional verification, non-compliant data undergoes removal, correction, and completion processing: spatially misaligned data is reprojected; topologically incorrect data is trimmed and cleaned; missing field data is completed, and abnormal data is recalculated; time-series deviation data is removed and re-collected. Ultimately, this ensures that all data have a unified spatial benchmark, reasonable topology, complete fields, and comparable time series, with a quality compliance rate of ≥95%, meeting the accuracy requirements for subsequent results integration and engineering applications.

[0124] After completing the full-dimensional quality verification of data, classification and organization of five dimensions, and cross-dimensional correlation matching, the verified global spatial autocorrelation statistical results, local cluster type maps, hot and cold spot significance distribution maps, ecological carrying capacity spatial risk type maps, engineering disturbance risk contribution lists, time series change maps, control zone boundaries, hierarchical attributes, risk tracing results, and differentiated control suggestions are integrated to form a complete result of hot and cold spot identification and zoned control of ecological carrying capacity of hydropower projects.

[0125] The complete results include ten core categories, covering the entire chain of spatial statistics, risk identification, disturbance correlation, dynamic evolution, and zonal management: Global spatial autocorrelation statistics include global Moran's I index, Z-value, P-value, and cluster type determination. Global Moran's I = 0.72, Z = 12.5, P < 0.01, indicating significant positive clustering. Local cluster type maps include 30m raster data, labeling five cluster types: HH, LL, HL, LH, and NS, and their corresponding area percentages. The LL cluster area is 3.2 km², accounting for 4% of the study area. Cold / hot spot significance distribution maps are used to distinguish between strong / medium / weak cold, strong / medium / weak hot, and insignificant, labeling the area of ​​each level. Strong cold spots cover 1.8 km², and medium cold spots 2.5 km². Ecological carrying capacity spatial risk type maps are used to label the boundaries of core cold spot areas, general low-value areas, core hot spot areas, risk frontier areas, and general monitoring areas. The engineering disturbance risk contribution list includes the name, type, area, risk contribution index, risk level, and associated risk type of the disturbance unit. Waste disposal site No. 1, patchy, 0.62 km², contribution index 0.94, high contribution, associated with the core cold spot area. The multi-period cold and hot spot time-series change map is used to mark newly added / continuous / regressing cold spots, newly added / continuous / degraded hot spots and their expansion rate, and regression. The cold spot expansion rate during the construction period is 40%, and the regression rate during the operation period is 15%.

[0126] The five-level control zone boundary vector includes closed boundaries for strictly controlled zones, priority ecological restoration zones, risk diffusion prevention and control zones, key protected zones, and general monitoring zones. The zone classification attribute table includes zone name, area, risk level, associated disturbances, and control priority. The strictly controlled zone covers 2.1 km², has a high risk level, and is associated with spoil disposal sites and tunnel entrances. The risk source tracing results report explains the sources of risk in the zone, the proportion of disturbance contribution, and the causes of evolution trends. 85% of the risk in the strictly controlled zone originates from disturbances from spoil disposal sites. Differentiated control recommendations include matching construction, restoration, prevention, protection, and monitoring measures according to zone type.

[0127] The above ten types of deliverables are uniformly summarized into three standardized deliverable packages: visualization atlases, statistical documents, and control reports, adapted to the direct access needs of engineering management. The visualization atlases include global autocorrelation thematic maps, local cluster maps, hot and cold spot level maps, spatial risk maps, time-series evolution maps, and five-level zoning maps, with unified scales, legends, and coordinates for easy on-site viewing and presentation. The statistical documents include global / local statistical reports, hot and cold spot area statistics tables, disturbance risk contribution lists, time-series indicator summary tables, and zoning area statistics tables, with standardized fields and accurate values, supporting data traceability and verification. The control reports include deliverable descriptions, risk source tracing, zoning control recommendations, and implementation priorities, with plain language and clear conclusions, serving as a direct basis for engineering environmental management decisions.

[0128] The standardized deliverables package can be directly integrated with the construction phase management, operation phase restoration, and cascade development planning of hydropower projects, making ecological risks "visible, calculable, and manageable." During the construction phase, priority is given to managing high-contribution disturbance units and strictly controlling the boundaries of strictly controlled areas. During the operation phase, priority is given to implementing ecological restoration in priority areas and preventing the spread of risks. During the planning phase, hotspot degradation areas are avoided and the site selection of new projects is optimized.

[0129] In one implementation, such as Figure 2 As shown, this application also provides a system for identifying hot and cold spots and zoning control of ecological carrying capacity in hydropower projects, including: Data acquisition module 201 is used to acquire the ecological carrying capacity evaluation results of the hydropower project impact area, spatial data of project disturbance, ecological restoration data and auxiliary geographic environment data, and generate basic dataset information; The data preprocessing and weight construction module 202 is used to carry out spatial registration, scale unification, effective pixel screening and standardization processing, and construct a multi-scale adaptive spatial weight matrix based on the differentiated disturbance patterns of construction roads, spoil disposal sites and tunnel entrances, and generate adaptive weight information. The Spatial Statistics and Risk Construction Module 203 is used to calculate the global spatial autocorrelation index, local spatial autocorrelation index and Getis-OrdGi statistics, identify spatial clustering characteristics, clustering types and hot and cold spot levels, integrate and generate spatial risk types of ecological carrying capacity, and output global statistics, local clustering, hot and cold spot and spatial risk information. The disturbance correlation and contribution assessment module 204 is used to conduct spatial overlay and buffer analysis, calculate the cold point correlation ratio, low-low clustering correlation, risk intensity and area ratio, construct the risk contribution index of engineering disturbance units and identify key risk units, and generate disturbance correlation assessment results information. The dynamic change identification module 205 is used to conduct time series analysis on multi-period data, determine the evolution type of hot and cold spots and quantify the expansion rate and decline rate, and output time series change graph and dynamic evolution information. The zoning and recommendation generation module 206 is used to integrate risk type, significance, correlation strength and time series trend to delineate five types of control zones and output the boundaries, hierarchical attributes and risk tracing results, and generate differentiated control recommendation information; The results integration and output module 207 is used to summarize the full-chain analysis results and form standardized results information on the identification of hot and cold spots and zoning management of ecological carrying capacity of hydropower projects.

[0130] All embodiments in this application are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on its differences from other embodiments. In particular, the embodiments of the method for identifying and zoning control of hot and cold spots in hydropower projects, the electronic device, the electronic device, and the readable storage medium are basically similar to the above-described embodiments of the method for identifying and zoning control of hot and cold spots in hydropower projects. Therefore, the descriptions are relatively simple, and relevant parts can be referred to in the descriptions of the above-described embodiments of the method for identifying and zoning control of hot and cold spots in hydropower projects.

Claims

1. A method for identifying hot and cold spots and implementing zoned management of ecological carrying capacity in hydropower projects, characterized in that, include: Obtain the ecological carrying capacity assessment results, spatial data of engineering disturbance, ecological restoration data, and auxiliary geographic environment data of the impact area of ​​hydropower projects; Data preprocessing was completed based on spatial registration, scale unification, effective pixel screening and standardization. A multi-scale adaptive spatial weight matrix was constructed based on the different engineering disturbance patterns of construction roads, spoil disposal sites and tunnel entrances. Calculate the global spatial autocorrelation index, output the global spatial autocorrelation results, diagnose the overall clustering degree of ecological carrying capacity, calculate the local spatial autocorrelation index, output the local clustering map, identify strong / medium / weak hot and cold hot areas through the Getis-OrdGi statistic, integrate the local clustering type and the significance of hot and cold hot areas, output the spatial risk map, and construct the spatial risk type of ecological carrying capacity; Spatial risk types are spatially superimposed and buffered with engineering disturbance units to calculate the cold point correlation ratio, low-low clustering correlation ratio, risk intensity, and risk area ratio, thereby constructing an engineering disturbance unit risk contribution index and identifying key risk contribution units. Dynamic changes in multi-period ecological carrying capacity results are identified to determine the change types of newly added / continuous / receding cold spots and newly added / continuous / degraded hot spots, quantify the expansion rate and decline rate of cold spots, and output a time-series change map; Based on spatial risk types, the significance of hot and cold spots, the correlation strength of engineering disturbances, and multi-period change trends, strict control areas, ecological restoration priority areas, risk diffusion prevention and control areas, key protection areas, and general monitoring areas are delineated according to preset rules. The results of zoning boundaries, hierarchical classification, and risk source tracing are output, and differentiated control recommendations are generated. By integrating global spatial autocorrelation results, local clustering maps, hot and cold spot distribution maps, spatial risk maps, engineering disturbance risk contribution lists, time series change maps, zoning boundaries, hierarchical attributes, risk tracing and differentiated management recommendations, the results of hot and cold spot identification and zoning management of hydropower project ecological carrying capacity are formed.

2. The method as described in claim 1, characterized in that, Data preprocessing was completed based on spatial registration, scale unification, effective pixel selection, and standardization. A multi-scale adaptive spatial weight matrix was constructed according to the different engineering disturbance patterns of construction roads, spoil heaps, and tunnel entrances, including: Based on multi-form ecological carrying capacity data of hydropower project impact areas (grids / vector patches / small watersheds / construction management units), full-type engineering disturbance data of construction roads / spoil disposal sites / tunnel entrances / mixing systems / construction platforms / riverbank disturbance zones / construction camps / material quarries, and geographic auxiliary data of DEM / water system / protection zone boundaries, we performed preprocessing including unified coordinate system, resolution matching, study area cropping, invalid / missing pixel removal, multi-period time series registration, standardization / grading of ecological carrying capacity values, and inter-period unit comparability verification. To address the differentiated disturbance patterns of construction roads, spoil heaps, tunnel entrances / mixing systems, and riverbank disturbance zones, a multi-scale adaptive spatial weight matrix is ​​constructed using distance attenuation, adjacency matrix, K-nearest neighbor, and river network adjacency modes. This matrix includes buffers along the route, areal buffers, ring buffers, and corridor adaptation. It also integrates slope aspect and gradient, slope runoff, wind direction and speed, and dust diffusion correction neighborhoods to match the scale of linear, point, patchy, and corridor disturbance influences.

3. The method as described in claim 1, characterized in that, Calculate the global spatial autocorrelation index, output the global spatial autocorrelation results, diagnose the overall clustering degree of ecological carrying capacity, calculate local spatial autocorrelation indicators, output local clustering maps, identify strong / medium / weak hot and cold areas using the Getis-OrdGi statistic, integrate local clustering types and the significance of hot and cold areas, output spatial risk maps, and construct spatial risk types for ecological carrying capacity, including: For the preprocessed ecological carrying capacity data, the global Moran's I index algorithm is used to carry out overall spatial clustering diagnosis. The clustering, discrete or random distribution characteristics are determined by the significance test of Z value and P value, and the global spatial heterogeneity pattern is identified. A local Moran's I computational model was constructed, and combined with the neighborhood association analysis mechanism, to accurately identify five spatial clustering types: high-high, low-low, high-low, low-high, and insignificant, and to locate local clusters and anomalous regions. Introducing Getis-OrdGi The statistical calculation model uses three significance thresholds—P<0.01, 0.01≤P<0.05, and 0.05≤P<0.10—to classify strong / medium / weak cold spots and strong / medium / weak hot spots, quantify the clustering intensity of cold and hot spots, and generate local cluster maps. A fusion discrimination model of LISA clustering type and cold / hot spot salience is established. Based on the fusion rules of LL+ strong / medium cold spots and HH+ strong / medium hot spots, multiple types of ecological carrying capacity spatial risk are constructed, including core cold spot area, core hot spot area and risk frontier area, to achieve composite identification of spatial pattern.

4. The method as described in claim 1, characterized in that, Spatial risk types are spatially overlaid and buffered with engineering disturbance units to calculate the cold spot correlation ratio, low-to-low clustering correlation ratio, risk intensity, and risk area proportion. This constructs a risk contribution index for engineering disturbance units and identifies key risk contribution units, including: Based on spatial risk type layers, engineering disturbance unit vector data, and distance threshold parameters, a triple analysis mechanism of spatial overlay, buffer generation, and neighborhood association strength is constructed. Taking the influence area of ​​the disturbance unit as the analysis scope, four types of quantitative indicators are extracted: cold point association ratio, low-low clustering association ratio, average risk intensity of the disturbance area, and risk area ratio. A weighting criterion is constructed based on risk correlation, disturbance contribution, and indicator balance, and a linear weighting model is used to determine the weight coefficients. Establish a risk contribution index calculation model ;in, For the first Risk contribution index of engineering disturbance units; The correlation ratio of cold spots; The proportion of low-to-low clustering associations; The average risk intensity within the affected area of ​​the engineering disturbance unit; The area of ​​the risk zone within the affected area of ​​the engineering disturbance unit; Verify the completeness of indicators, the suitability of weights, and the distinguishability of the index; calibrate the weight coefficients and threshold ranges; and match the risk quantification requirements of different types of engineering disturbances. Based on the risk contribution index threshold, the risk contribution levels are divided into high-contribution key risk units, medium-contribution general risk units, and low-contribution low-risk units, and a list of risk contributions of disturbance units is output.

5. The method as described in claim 1, characterized in that, Dynamic change identification is performed on multi-period ecological carrying capacity results to determine the change types of newly added / persistent / receding cold spots and newly added / persistent / degraded hot spots, quantify the expansion rate and decline rate of cold spots, and output time-series change maps, including: Dynamic changes are identified in the time-series aligned multi-period ecological carrying capacity results, generating dynamic analysis data covering cold / hot spot evolution types, time-series change indicators, and stage characteristics. The time-series aligned multi-period data includes ecological carrying capacity grids, cold / hot spot type maps, and spatial risk type layers for each period, and the associated data are stage identifier data for the construction period, operation and recovery period, and new construction and start-up period. A typological evolution judgment was made on time-series aligned multi-period data, generating seven change types: newly added cold spots, persistent cold spots, fading cold spots, newly added hot spots, persistent hot spots, hot spot degradation, and hot spot turning into cold spots. Four evolution characteristics of cold and hot spots were clarified: expansion, contraction, migration, and fading. Based on the temporal evolution type and spatial unit change, a quantitative calculation model is constructed to generate three core time series indicators: cold point expansion rate, cold point decline rate, and hot point degradation. The integrity of time series data, the distinguishability of evolution type, and the accuracy of indicator calculation are verified simultaneously to match the dynamic monitoring needs of the engineering stage. By integrating evolution types, time-series indicators, and stage characteristic data, a time-series change diagram is generated to support dynamic management and control throughout the entire process.

6. The method as described in claim 5, characterized in that, By integrating global spatial autocorrelation results, local clustering maps, hot and cold spot distribution maps, spatial risk maps, engineering disturbance risk contribution lists, time-series variation maps, zoning boundaries, hierarchical attributes, risk tracing, and differentiated management recommendations, the results of hot and cold spot identification and zoning management of hydropower project ecological carrying capacity are formed, including: Based on nine core output data categories—global spatial autocorrelation results, local clustering diagrams, hot and cold spot distribution maps, spatial risk maps, engineering disturbance risk contribution lists, time-series variation diagrams, partition boundaries, hierarchical attributes, and risk tracing—layer standardization, field mapping, and format adaptation processing are performed. The output data is categorized, organized, and matched according to five dimensions: spatial statistics, risk identification, disturbance correlation, dynamic evolution, and zoning control. Add spatial coordinate system I, projection transformation, topology verification, field integrity verification, and time-series alignment quality control parameters; By integrating global spatial autocorrelation statistics, local cluster type maps, hot and cold spot significance distribution maps, ecological carrying capacity spatial risk type maps, engineering disturbance unit risk contribution lists, multi-period hot and cold spot change maps, control zone boundaries, zone level and zone attribute tables, risk source identification results, and differentiated control recommendations, the results of hot and cold spot identification and zoned control of hydropower project ecological carrying capacity are formed.

7. A system for identifying hot and cold spots and controlling zoning of ecological carrying capacity in hydropower projects, characterized in that, The system includes: The data acquisition module is used to acquire the ecological carrying capacity evaluation results of the hydropower project impact area, spatial data of project disturbance, ecological restoration data and auxiliary geographic environment data, and generate basic dataset information; The data preprocessing and weight construction module is used to carry out spatial registration, scale unification, effective pixel screening and standardization. Based on the differentiated disturbance patterns of construction roads, spoil heaps and tunnel entrances, a multi-scale adaptive spatial weight matrix is ​​constructed to generate adaptive weight information. The Spatial Statistics and Risk Construction Module is used to calculate the global spatial autocorrelation index, local spatial autocorrelation indicators and Getis-OrdGi statistics, identify spatial clustering characteristics, clustering types and hot and cold spot levels, integrate and generate spatial risk types of ecological carrying capacity, and output global statistics, local clustering, hot and cold spot and spatial risk information. The disturbance correlation and contribution assessment module is used to conduct spatial overlay and buffer analysis, calculate the cold point correlation ratio, low-low clustering correlation, risk intensity and area ratio, construct the risk contribution index of engineering disturbance units and identify key risk units, and generate disturbance correlation assessment results information. The dynamic change identification module is used to conduct time series analysis on multi-period data, determine the evolution type of hot and cold spots and quantify the expansion rate and decline rate, and output time series change graphs and dynamic evolution information; The zoning and recommendation generation module is used to integrate risk type, significance, correlation strength and time series trend to delineate five types of control zones and output the boundaries, hierarchical attributes and risk tracing results, and generate differentiated control recommendation information; The results integration and output module is used to summarize the full-chain analysis results and form standardized results information on the identification of hot and cold spots and zoning management of ecological carrying capacity of hydropower projects.

8. An electronic device, characterized in that, include: First processor; The processor includes a memory for storing executable instructions of the first processor; wherein the first processor is configured to execute the method for identifying hot and cold spots of ecological carrying capacity in hydropower projects according to any one of claims 1 to 6 by executing the executable instructions.

Images