Reservoir ecological regulation system for natural breeding of fish

Abstract

This invention discloses a reservoir ecological scheduling system for fish natural reproduction, comprising a scheduling control module, a data integration and storage module, a water ecology simulation and analysis module, and a scheduling decision evaluation module. The scheduling control module manages scheduling thresholds, formulates scheduling requirements, optimizes model calculation node allocation through resource scheduling algorithms, and monitors platform operation status. The data integration and storage module collects and stores multi-source data, performs standardized cleaning and spatiotemporal fusion on the multi-source data, forming a structured database system. The water ecology simulation and analysis module drives models such as hydrodynamic simulation, habitat suitability assessment, and reproduction scale prediction to couple data from the data integration and storage module. The scheduling decision evaluation module provides dynamic assessment of spawning ground suitability and reproduction scale prediction information, calls verification data from the data integration and storage module under the coordination of the scheduling control module, compares and analyzes monitoring data with simulation results, and evaluates the effectiveness of the actual scheduling process.

Description

Technical Field

[0001] This invention relates to the field of ecological environmental protection, specifically to a reservoir ecological regulation system for the natural reproduction of fish. Background Technology

[0002] While the water environment quality of the Yangtze River Basin continues to improve, the decline in biodiversity has not been fundamentally curbed, with flagship species numbers continuing to decrease and river and lake fishery resources plummeting to approximately 20% of historical levels. Simultaneously, the cumulative storage capacity of over 50,000 reservoirs has reached 37% of total surface runoff, significantly altering natural hydrological rhythms and highlighting the growing conflict between water conservancy project operation and ecosystem integrity. Against this backdrop, developing scientific analytical tools for quantitatively assessing the effectiveness of ecological regulation has become a crucial support for promoting systematic governance and ecological restoration in the Yangtze River Basin.

[0003] In recent years, research on watershed ecological scheduling has shown a trend towards multi-model fusion and multi-scale integration. In habitat simulation, physical habitat assessment methods based on hydrodynamic models, such as PHABSIM and River 2D, have been widely used to evaluate the impact of factors such as flow velocity and water depth on fish fitness. In biological response analysis, statistical models such as the IHA-RVA method and generalized additive models (GAMs) have revealed the quantitative relationship between flow variation and fish spawning scale, providing a scientific basis for setting ecological scheduling thresholds. The HEC-RAS and MIKE series models have also made significant progress in the joint scheduling of reservoir groups and the optimization of ecological release schemes. However, existing research mostly focuses on the simulation of single-stage hydrological processes or post-hoc statistical analysis, lacking dynamic coupling and feedback assessment of the entire process of "scheduling scheme - hydrological change - habitat response - reproductive effectiveness."

[0004] Therefore, it is urgent to build a comprehensive analysis platform that integrates multi-source monitoring data, couples hydrological and ecological processes, and has dynamic decision support functions, so as to achieve scientific and precise watershed ecological scheduling. Summary of the Invention

[0005] This invention proposes a reservoir ecological scheduling system for fish natural reproduction, which realizes unified management of multi-source heterogeneous data, integrates intelligent calling and coupled calculation of models such as hydrodynamic simulation, habitat suitability evaluation, and reproduction scale prediction, supports scheduling demand generation, scheme simulation and effect evaluation, and realizes the scientific and precise ecological scheduling of the watershed.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows: a reservoir ecological scheduling system for the natural reproduction of fish, comprising a scheduling control module, a data integration and storage module, a water ecological simulation and analysis module, and a scheduling decision evaluation module; The scheduling control module is communicatively connected to the data integration and storage module, the water ecology simulation and analysis module, and the scheduling decision evaluation module. It is configured to acquire hydrological monitoring data of the scheduling scheme. The hydrological monitoring data includes flow and water temperature monitoring records. The data integration and storage module constructs a spatial computing grid based on the physical boundary constraints and DEM elevation data of the river channel in the target study area. It acquires discrete monitoring data from multiple monitoring points distributed in the river channel of the study area in the scheduling scheme. Through spatial interpolation and temporal interpolation processing, the multi-source heterogeneous discrete monitoring data is transformed into continuous field data covering the spatial computing grid and eliminating temporal resolution differences. The monitoring data includes hydrological monitoring data, fish reproduction monitoring data, and water quality data. The continuous field data generated based on the hydrological monitoring data drives the hydrodynamic model to obtain dynamic hydraulic attribute data containing water depth and flow velocity vectors, and calculates the water level rise. Then, based on the dynamic hydraulic attribute data and environmental data including water level, water temperature, and water level rise, threshold matching and retrieval are performed through a preset suitability curve parameter table to quantify and output the single-factor suitability of each grid unit for the target fish. The aquatic ecosystem simulation and analysis module assesses habitat suitability based on the distribution of individual fitness scores for each target fish species under each scheduling condition in all grid cells, provided by the data integration and storage module. Finally, the random forest model predicts the fish reproduction scale. The scheduling decision evaluation module compares and evaluates scheduling schemes based on the prediction results output by the water ecology simulation analysis module.

[0007] Furthermore, the construction of a spatial computing grid based on the physical boundary constraints and DEM elevation data of the river channel in the target study area includes the following steps: generating a triangular unstructured grid with the irregular shoreline boundary of the river channel in the study area as a constraint, and mapping the DEM elevation data of the river channel in the study area to the grid cells to characterize the terrain features. Spatial and temporal interpolation processing includes the following steps: converting the discrete monitoring data into continuous spatial field data covering the spatial computing grid; performing time series standardization processing on the continuous spatial field data to eliminate the temporal resolution differences between heterogeneous discrete monitoring data sources and generate continuous field data with a unified spatiotemporal reference. The derived calculation of water level rise includes the step of converting the water depth stored in the grid cells into the difference between the water depth of the current day and the water depth of the previous day; Threshold matching and retrieval are performed using a pre-set suitability curve parameter table. The quantitative output of the single-factor suitability of each grid cell for the target fish includes the following steps: matching the environmental data stored in each grid cell with the environmental data threshold range corresponding to the target fish; checking whether the current environmental data value falls between the minimum and maximum values ​​defined by the suitability curve corresponding to the environmental data; if so, returning the suitability score corresponding to the environmental data value on the suitability curve to the grid cell; otherwise, returning 0.

[0008] Furthermore, the unstructured triangular mesh is generated using the Delaunay triangulation algorithm; the DEM elevation data is fused into the mesh cells through nearest neighbor search or bilinear interpolation, and for areas containing hydraulic structures such as groynes and revetments, the riverbed elevation of specific meshes is manually corrected.

[0009] Furthermore, spatial interpolation includes: using the inverse distance weighted IDW algorithm to interpolate discrete monitoring data sampled synchronously at the same sampling frequency. Spatial interpolation is performed to map the discrete observations of the monitoring points to the grid cells, thereby obtaining continuous spatial field data of the river channel covering the study area. ; in, The distance is determined by the weighted average of all monitoring points. The closer the weight The larger; For grid cells to monitoring point Euclidean distance, The exponent is set to 2, which controls the rate at which the weight decays with distance; Time interpolation includes: using linear interpolation to interpolate the target time. Previous continuous space field data and target time Subsequent continuous spatial field data Alignment to obtain continuous spatial field data at the target time. , ; .

[0010] Furthermore, the suitability curves include flow velocity suitability curves, water depth suitability curves, water temperature suitability curves, and water level rise suitability curves. The assessment of habitat suitability includes calculating a comprehensive suitability index using the geometric mean method. ; The suitability score is the value of a point on the flow velocity suitability curve. The suitability score is the value of a point on the water depth suitability curve. The suitability score is the value of a point on the water temperature suitability curve. The suitability score for a point on the water level rise suitability curve; Assessing habitat suitability also includes calculating the weighted usable area of ​​habitat. : In the formula: Let be the area of ​​the i-th grid cell, and n be the total number of grid cells.

[0011] Furthermore, the random forest model uses historical fish reproduction monitoring data and hydrological characteristic parameters stored in the data integration and storage module as training data, and generates the model through Bootstrap resampling. Individual training set For each subset Construct a regression tree Each tree is trained independently based on randomly selected samples and features. The final output is the predicted egg production value, which is the average of the predictions from all regression trees. This enables the prediction of fish spawning numbers under different scheduling schemes; The hydrological feature parameters input to the random forest model include: habitat-weighted available area. Water level rise.

[0012] Furthermore, the coefficient of determination of the trained random forest model is greater than a threshold. The coefficient of determination is calculated as follows: In the formula: As the coefficient of determination, These are actual observations. These are predicted values.

[0013] The beneficial effects are as follows: This system deploys a functionally coordinated scheduling and control module, a data integration and storage module, a water ecological simulation and analysis module, and a scheduling decision evaluation module. Through key technologies such as multi-source ecological data fusion and mechanism and data dual-drive model integration, it has constructed a decision support system covering the entire scheduling process. It has realized the unified management of multi-source heterogeneous data such as hydrology, topography, fish behavior and reproduction monitoring. It integrates the intelligent calling and coupled calculation of models such as hydrodynamic simulation, habitat suitability evaluation, and reproduction scale prediction. It supports functions such as scheduling demand generation, scheme simulation and effect evaluation. It provides quantitative, visualized and intelligent decision support for differentiated ecological scheduling of key sections of the Yangtze River, such as Jiangjin, Xiaojiang and downstream of Gezhouba, significantly improving the scientificity and accuracy of ecological scheduling of water conservancy projects in the basin, and strongly supporting the restoration and sustainable development of the water ecosystem under the background of the Yangtze River protection. Attached Figure Description

[0014] Figure 1 The architecture of the water ecological scheduling and analysis platform; Figure 2 For the integration of multi-source heterogeneous ecological and environmental big data in the Yangtze River Basin; Figure 3 Comparison of predicted and measured fish egg quantity results; Figure 4 The flow velocity suitability curve and suitability score; Figure 5 This refers to the water depth suitability curve and suitability score. Figure 6 This includes the water temperature suitability curve and suitability score. Figure 7 This refers to the water level rise suitability curve and suitability score. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Example 1

[0016] like Figure 1 As shown, the reservoir ecological scheduling system for the natural reproduction of fish includes a scheduling control module, a data integration and storage module, a water ecological simulation analysis module, and a scheduling decision evaluation module, integrating data management, simulation calculation, scheduling control, and decision evaluation functions.

[0017] The dispatch control module acquires hydrological monitoring data for the dispatch plan, including flow and water temperature monitoring records, through existing channels such as the National Flood Control and Drought Relief Command System. In this embodiment, the flow rate of the dispatch plan is used from... Every Increase sequentially to There are a total of 65 scheduling conditions.

[0018] The data integration and storage module accesses hydrological monitoring data in real time through the big data platform interface, and performs standardized cleaning and spatiotemporal fusion of multi-source heterogeneous data such as fish biological parameters, habitat spatial attributes and fish reproduction monitoring data to construct a unified aquatic ecological database.

[0019] Data integration is a fundamental component of the system, providing unified and reliable data support for model computation and decision analysis. For example... Figure 2 As shown, the system, relying on the data integration and storage module, integrates hydrological monitoring data, ecological observation data, topographic mapping data, and meteorological and water quality data to form a comprehensive data system with multiple dimensions and spatiotemporal scales. Ecological observation data covers biological parameters, habitat distribution, and fish reproduction monitoring data for the four major freshwater fish species and other typical fish. Topographic mapping data comes from underwater cross-section measurements and elevation remote sensing images, used to construct river topography and habitat spatial patterns. Simultaneously, meteorological and water quality data are combined to supplement the environmental background description of ecological processes. All types of data underwent standardization processing, including outlier removal, missing value imputation, and data normalization. After standardization and quality control, the data was integrated according to a unified spatial and temporal benchmark to form a structured database system.

[0020] The data integration and storage module forms a structured database system, including the following steps: B1. Obtain the irregular shoreline boundary of the study area (such as the meandering river channel from Yichang to Zhijiang in the middle reaches of the Yangtze River). Using the shoreline boundary as the constraint boundary, generate a set of unstructured triangular meshes within the study area using the Delaunay triangulation algorithm. ,in For the first Each unit is a triangular cell; in this embodiment, the grid size is set to approximately 50m, and the total number of grids is approximately 46,624.

[0021] Mapping topographic mapping data, hydrological monitoring data, ecological observation data, and meteorological and water quality data to the node attributes of the nodes and / or grid cells of the triangular unstructured grid, including: B2. Obtain a digital elevation model (DEM) of the river channel covering the study area. Use the elevation information of the DEM as a static geometric attribute and map it to the node attributes of the nodes and / or grid cells of the triangular unstructured mesh, forming topographic attribute data spatially registered with the mesh set Ω. For areas containing hydraulic structures such as groynes and revetments, specific nodes are manually corrected. Values ​​that ensure the accuracy of physical boundaries.

[0022] Among them, the digital elevation model (DEM) of the river channel is generated from topographic mapping data and includes spatial coordinates. and its corresponding elevation information Each node or cell in the mesh Riverbed elevation Obtained from the DEM mapping operator: in, This indicates the coordinates on the DEM dataset. Elevation values ​​obtained by nearest neighbor search interpolation or bilinear interpolation.

[0023] Static geometric properties also include: mesh ID, and the coordinates of the mesh's center point or nodes.

[0024] B3. Perform spatial interpolation, using the same sampling frequency from several similar data source monitoring points within the river channel of the study area. Discrete monitoring data acquired through synchronous sampling are transformed into continuous field data of the river channel covering the study area. This data is then mapped to the node attributes of the nodes and / or grid cells of the triangular unstructured grid, resulting in continuous spatial field data composed of monitoring data from the same data source at the grid nodes. Then, time interpolation is performed to convert continuous spatial field data from various data sources with different sampling frequencies. Aligned with equidistant target times on a unified target time axis, continuous spatiotemporal field data from multiple data sources covering the study area are generated, including hydrological monitoring data, fish reproduction monitoring data, and water quality data. Spatial interpolation includes: applying the inverse distance weighted (IDW) algorithm to the monitoring data that are synchronously sampled at the same sampling frequency. Spatial interpolation is performed to map the discrete observations of the monitoring points to the set of grid nodes, resulting in continuous spatial field data of the grid nodes covering the river channel in the study area. ; in, The distance is determined by the weighted average of all monitoring points. The closer the weight The larger.

[0025] For grid nodes to monitoring point The Euclidean distance. It is a power exponent (usually 2), used to control the rate at which the weight decays with distance.

[0026] Time interpolation includes: using linear interpolation to interpolate spatial initial field data at different sampling frequencies. Align to a unified timeline, generate continuous field data for the target time, and store it in the database; Because the sampling frequencies of data from different sources are inconsistent, the system uses linear interpolation to interpolate the continuous spatial field data before the target time. Continuous spatial field data after the target time Alignment to obtain continuous spatial field data at the target time. , ; Hydrological monitoring data includes flow rate and water temperature; water quality data includes transparency; and fish reproduction monitoring data includes the composition and quantity of eggs from different species of fish.

[0027] B4. The continuous field data of hydrological monitoring data is transformed into dynamic hydraulic attribute data using the River2D hydrodynamic model, and mapped onto the nodes and / or grid cells of the triangular unstructured grid. The dynamic hydraulic attribute data includes the water depth, velocity modulus, and vector direction of the grid nodes and / or grid cells at the corresponding riverbed elevation and flow rate. The water depth stored in the grid nodes and / or grid cells is then converted into water level rise; the water level rise is the difference between the water depth of the current day and the water depth of the previous day. The River2D hydrodynamic model is a two-dimensional finite element model used to simulate the supercritical / subcritical flow transition, ice cover, and variable wetland conditions of natural rivers.

[0028] B5. Based on the environmental data of each grid node and / or grid cell, retrieve the environmental data from the pre-constructed suitability curve parameter table, match the environmental data with the environmental data threshold range corresponding to the target fish, and read the individual suitability score of the environmental data in the matched suitability curve according to the matching result, which is used for the subsequent calculation of the comprehensive suitability index.

[0029] Environmental data includes water depth, flow velocity, water temperature, and water level rise obtained in steps B3 and B4. The suitability curve parameter table (as shown in Table 1) stores the environmental data thresholds for each fish species and the suitability scores mapped to the matched suitability curves. The suitability curves, suitability scores, and their environmental data thresholds are shown in Table 1. Figure 4 , 5 As shown in Figures 6 and 7, this includes the flow velocity suitability curve and the suitability score. ) and the minimum value of its flow velocity ( ) and maximum value ( ), water depth suitability curve and suitability score ( ) and the minimum value of its water depth ( ) and maximum value ( ), water temperature suitability curve and suitability score ( ) and the minimum value of its water temperature ( ) and maximum value ( ), water level rise suitability curve and suitability score ( ) and the minimum value of the water level rise ( ) and maximum value ( Each target fish species also corresponds to a breeding season and spawning type in the fish basic information table (such as Table 2).

[0030] Table 1 Table 2 Based on the matching results, read the individual fitness scores of the environmental data in the matching fitness curve, including: Check if the environmental data value of the current row falls between the minimum (Min) and maximum (Max) values ​​defined by the corresponding suitability curve. If so, return the suitability score corresponding to the environmental data value on the suitability curve; otherwise, return 0.

[0031] Step B5 above can filter out all grids that meet the environmental data range of the target fish (such as spawning current velocity (e.g., 0.2-3.5 m / s)) and their suitability scores. The data integration and storage module calculates the distribution of the suitability scores of each target fish in all grids under each scheduling condition.

[0032] The water ecology simulation and analysis module extracts water ecology digital data from the data integration and storage module, assesses habitat suitability, and finally predicts the number of fish spawnings using a random forest model, storing the results in the scheduling effect evaluation table.

[0033] Assessing habitat suitability includes calculating a comprehensive suitability index using the geometric mean method: Assessing habitat suitability also includes calculating the weighted usable area of ​​habitat. : In the formula: Let be the area of ​​the i-th grid cell, and n be the total number of grid cells.

[0034] The random forest model uses historical reproduction monitoring data and hydrological characteristic parameters stored in the data integration and storage module as training data, and generates data through Bootstrap resampling. Individual training set For each subset Construct a regression tree Each tree is trained independently based on randomly selected samples and features. The final output is the predicted egg production value, which is the average of the predictions from all regression trees. This allows for the prediction of fish spawning numbers under different scheduling schemes.

[0035] The hydrological feature parameters input to the random forest model include: habitat-weighted available area. Water level rise. This embodiment increases the weight of water level rise in the suitability index and in the random forest model, thereby improving the reliability of the fish spawning prediction model and reducing prediction errors.

[0036] The reliability of fish spawning prediction models was evaluated using the coefficient of determination, which was calculated as follows: In the formula: As the coefficient of determination, These are actual observations. These are predicted values.

[0037] The scheduling effectiveness evaluation table stores the weighted total usable area, the proportion of suitable area, and the number of spawning fish species predicted by the random forest model for each scheduling scheme. This table supports the scheduling scheme comparison function of the scheduling decision evaluation module.

[0038] The data integration and storage module achieves the fusion and sharing of multi-source information through a unified data management platform. The platform performs format conversion, spatial registration, and semantic association on heterogeneous data, and constructs a management structure with a spatial database as its core. It supports efficient querying and model calling of multi-temporal and spatiotemporal resolution data. The data integration and storage module maintains real-time interaction with the water ecological simulation analysis module and the scheduling decision evaluation module, providing a stable data interface for model coupling calculation, scenario simulation, and ecological scheduling evaluation, and realizing centralized storage, dynamic updating, and collaborative application of data.

[0039] like Figure 3 As shown, to verify the reliability of the Yangtze River Basin water ecological scheduling and analysis system, model verification and empirical analysis were conducted in the Yichang to Zhijiang section of the Yangtze River Basin, with the roughness factor set to a globally uniform value. A cross-section within the simulation area was selected for hydraulic model verification. The hydrodynamic simulation results were compared with measured flow velocity and water depth. Spawning volume was predicted from May 29th to June 5th, 2023. The coefficient of determination between the monitored and predicted values ​​was also analyzed. The average relative error of the prediction results was 8.06%, and the trend was basically consistent. The accuracy of the water ecology simulation analysis module met the needs of predicting the breeding process of drifting fish in the Yangtze River and supporting ecological scheduling decisions.

[0040] Furthermore, the scheduling decision evaluation module provides the ecological scheduling management department with dynamic assessment of spawning ground suitability and prediction of breeding scale. Under the coordination of the scheduling control module, it calls the verification data of the data integration and storage module. Based on the comparative analysis of breeding monitoring data such as fish egg abundance and spawning ground distribution with simulation results, the actual scheduling process effect is evaluated.

[0041] The scheduling decision-making and evaluation module, based on task instructions issued by the scheduling control module, utilizes hydrological and ecological data from the data integration and storage module and calculation results from the model center to complete the entire process, including selection of scheduling areas and target fish species, scheduling demand analysis, scheduling scheme simulation, and effect evaluation. The module dynamically displays the response relationship between "flow rate - habitat suitability - breeding scale," providing feedback on the ecological scheduling effect through visualized curves and spatial distribution maps. Based on comparative analysis of monitoring data such as fish egg abundance and spawning ground distribution, the module can quantitatively evaluate the scheduling implementation effect and generate result reports and optimization suggestions.

[0042] While the specific embodiments of this disclosure have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of this disclosure. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of this disclosure are still within the scope of protection of this disclosure.

Claims

1. A reservoir ecological regulation system for the natural reproduction of fish, characterized in that, It includes a scheduling control module, a data integration and storage module, a water ecology simulation and analysis module, and a scheduling decision evaluation module; The scheduling control module is communicatively connected to the data integration and storage module, the water ecology simulation and analysis module, and the scheduling decision evaluation module. It is configured to acquire hydrological monitoring data of the scheduling scheme. The hydrological monitoring data includes flow and water temperature monitoring records. The data integration and storage module constructs a spatial computing grid based on the physical boundary constraints and DEM elevation data of the river channel in the target study area. It acquires discrete monitoring data from multiple monitoring points distributed in the river channel of the study area in the scheduling scheme. Through spatial interpolation and temporal interpolation processing, the multi-source heterogeneous discrete monitoring data is transformed into continuous field data covering the spatial computing grid and eliminating temporal resolution differences. The monitoring data includes hydrological monitoring data, fish reproduction monitoring data, and water quality data. The continuous field data generated based on the hydrological monitoring data drives the hydrodynamic model to obtain dynamic hydraulic attribute data containing water depth and flow velocity vectors, and calculates the water level rise. Then, based on the dynamic hydraulic attribute data and environmental data including water level, water temperature, and water level rise, threshold matching and retrieval are performed through a preset suitability curve parameter table to quantify and output the single-factor suitability of each grid unit for the target fish. The aquatic ecosystem simulation and analysis module assesses habitat suitability based on the distribution of individual fitness scores for each target fish species under each scheduling condition in all grid cells, provided by the data integration and storage module. Finally, the random forest model predicts the fish reproduction scale. The scheduling decision evaluation module compares and evaluates scheduling schemes based on the prediction results output by the water ecology simulation analysis module.

2. The reservoir ecological regulation system for natural fish reproduction according to claim 1, characterized in that, The construction of a spatial computing grid based on the physical boundary constraints and DEM elevation data of the river channel in the target study area includes the following steps: generating a triangular unstructured grid with the irregular shoreline boundary of the river channel in the study area as a constraint, and mapping the DEM elevation data of the river channel in the study area to the grid cells to characterize the terrain features. Spatial interpolation and temporal interpolation processing includes the following steps: converting the discrete monitoring data into continuous spatial field data covering the spatial computing grid; The continuous spatial field data is subjected to time series standardization processing to eliminate the time resolution differences between heterogeneous discrete monitoring data sources and generate continuous field data with a unified spatiotemporal reference. The derived calculation of water level rise includes the step of converting the water depth stored in the grid cells into the difference between the water depth of the current day and the water depth of the previous day; Threshold matching and retrieval are performed using a pre-set suitability curve parameter table. The quantitative output of the single-factor suitability of each grid cell for the target fish includes the following steps: matching the environmental data stored in each grid cell with the environmental data threshold range corresponding to the target fish; checking whether the current environmental data value falls between the minimum and maximum values ​​defined by the suitability curve corresponding to the environmental data; if so, returning the suitability score corresponding to the environmental data value on the suitability curve to the grid cell; otherwise, returning 0.

3. The reservoir ecological regulation system for natural fish reproduction according to claim 1, characterized in that, The unstructured triangular mesh is generated using the Delaunay triangulation algorithm; DEM elevation data is fused into the mesh cells through nearest neighbor search or bilinear interpolation, and for areas containing hydraulic structures such as groynes and revetments, the riverbed elevation of specific meshes is manually corrected.

4. The reservoir ecological regulation system for natural fish reproduction according to claim 3, characterized in that, Spatial interpolation includes: using the inverse distance weighted IDW algorithm to interpolate discrete monitoring data sampled synchronously at the same sampling frequency. Spatial interpolation is performed to map the discrete observations of the monitoring points to the grid cells, thereby obtaining continuous spatial field data of the river channel covering the study area. ; in, The distance is determined by the weighted average of all monitoring points. The closer the weight The larger; For grid cells to monitoring point Euclidean distance, The exponent is set to 2, which controls the rate at which the weight decays with distance; Time interpolation includes: using linear interpolation to interpolate the target time. Previous continuous space field data and target time Subsequent continuous spatial field data Alignment to obtain continuous spatial field data at the target time. , ; 。 5. The reservoir ecological regulation system for natural fish reproduction according to claim 2, characterized in that, The suitability curves include flow velocity suitability curves, water depth suitability curves, water temperature suitability curves, and water level rise suitability curves. The assessment of habitat suitability includes calculating a comprehensive suitability index using the geometric mean method. ； The suitability score is the value of a point on the flow velocity suitability curve. The suitability score is the value of a point on the water depth suitability curve. The suitability score is the value of a point on the water temperature suitability curve. The suitability score for a point on the water level rise suitability curve; Assessing habitat suitability also includes calculating the weighted usable area of ​​habitat. : In the formula: Let be the area of ​​the i-th grid cell, and n be the total number of grid cells.

6. The reservoir ecological regulation system for natural fish reproduction as described in claim 1, characterized in that, The random forest model uses historical fish reproduction monitoring data and hydrological characteristic parameters stored in the data integration and storage module as training data, and is generated through Bootstrap resampling. Individual training set For each subset Construct a regression tree Each tree is trained independently based on randomly selected samples and features. The final output is the predicted egg production value, which is the average of the predictions from all regression trees. This enables the prediction of fish spawning numbers under different scheduling schemes; The hydrological feature parameters input to the random forest model include: habitat-weighted available area. Water level rise.

7. The reservoir ecological regulation system for natural fish reproduction as described in claim 1, characterized in that, The coefficient of determination of the trained random forest model is greater than a threshold. The coefficient of determination is calculated as follows: In the formula: As the coefficient of determination, These are actual observations. These are predicted values.

Images