Fish natural breeding ecological scheduling method based on natural weather process

Abstract

The invention belongs to the technical field of ecological scheduling of water conservancy projects, and relates to a fish natural breeding ecological scheduling method based on a natural weather process. The core of the method is to collect target fish breeding related data, analyze required hydrological parameters by using a statistical model, and formulate a conventional scheduling and ecological scheduling scheme according to hydrometeorological forecast. The method specifically comprises the steps of collecting spawning fish types, roe density and other information, calculating hydrological change characteristic indexes, and determining key hydrological indexes and a suitable range through a random forest and other models; collecting actual measurement information, predicting and simulating conventional scheduling discharged flow by using an artificial intelligence method, and constructing a correlation curve; parameters such as ecological scheduling time are determined by combining conditions such as water temperature, and the test is implemented after feasibility is analyzed from multiple aspects.

Description

Technical Field

[0001] This invention relates to the field of ecological scheduling technology for water conservancy projects, specifically to an ecological scheduling method for the natural reproduction of fish based on natural weather processes. Background Technology

[0002] During the construction and operation of water conservancy projects, the scheduling of reservoirs and other water facilities has multifaceted impacts on river ecosystems. On the one hand, the storage and release of water in reservoirs alters natural hydrological rhythms, leading to more homogenized flow processes and reduced water level fluctuations in many rivers. This poses a serious challenge to fish species that rely on specific hydrological conditions for natural reproduction. For example, some fish species require specific combinations of water temperature, flow, and water level changes to trigger spawning behavior, and reservoir scheduling often disrupts these natural water flow signals, resulting in decreased fish reproductive success rates and threatening population numbers.

[0003] Currently, although some measures have been implemented for fish conservation, numerous technical challenges remain in effectively coordinating reservoir operation with the natural reproductive needs of fish. Traditional scheduling methods often focus on the single function of water conservancy projects (such as maximizing power generation benefits), lacking a comprehensive consideration of the ecosystem, especially the reproductive needs of fish. Furthermore, existing ecological scheduling research faces problems in practical applications, including low data accuracy, limited model predictive capabilities, and a lack of specificity and operability in scheduling schemes. Therefore, there is an urgent need for an ecological scheduling method that can comprehensively consider natural weather processes, fish reproductive needs, and the actual operation of reservoirs to achieve the coordinated development of water conservancy projects and ecological protection. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the technical solution adopted by this invention is: a fish natural reproduction ecological regulation method based on natural weather processes, comprising the following steps:

[0005] Step 1: Collect basic information on the natural reproduction and hydrological conditions of the target fish species;

[0006] Step 2: Analyze the hydrological parameters required for the natural reproduction of the target fish species using statistical models;

[0007] Step 3: Based on hydrological and meteorological forecast information, analyze weather and runoff processes, and formulate a routine reservoir operation plan;

[0008] Step 4: Based on weather and runoff processes, develop an ecological scheduling plan that meets the hydrological needs for the natural reproduction of the target fish species;

[0009] Step 5: Analyze the feasibility of the ecological scheduling scheme, and then implement the ecological scheduling experiment.

[0010] Preferably, in step one, the natural reproduction and hydrological basic information of the target fish includes historical data and current data, and both include data on the species of spawning fish, data on fish egg density, data on water temperature at spawning, data on flow rate at spawning, and data on water level at spawning.

[0011] Preferably, in step two, hydrological change characteristic indicators are calculated based on the natural reproduction of the target fish and its basic hydrological information, a daily fish egg density and hydrological change characteristic indicator set is obtained, a random forest model method is used to identify hydrological change characteristic indicators that affect the natural reproduction of the target fish, a relationship model between fish egg density and hydrological change characteristic indicators is constructed, and the appropriate range of hydrological change characteristic indicators when the fish egg density is high is simulated and calculated.

[0012] Preferably, in step three, the hydrological and meteorological forecast information includes air temperature, precipitation, water temperature, and flow information. Artificial intelligence methods are used to analyze the weather and runoff projects, predict the changes in hydrological and meteorological forecast information for the next ten days, and simulate the reservoir discharge flow process under normal scheduling conditions by combining reservoir scheduling operation procedures and constraints. Through hydrological sequence data, a relationship curve between upstream reservoir discharge flow and downstream typical section flow is constructed.

[0013] Preferably, the method for constructing the relationship curve between the upstream reservoir discharge and the downstream typical cross-sectional discharge is as follows:

[0014] Step 3-1: Collect measured hydrological and meteorological information as input variables, including but not limited to temperature, precipitation, relative humidity, wind speed, and flow rate;

[0015] Step 3-2: Select predictor variables as input variables, including but not limited to water temperature, precipitation, and inflow. The output variables are the water temperature, precipitation, and inflow predicted by the machine learning model for the next ten days based on these predictor variables. Set the training period and the validation period, and set the forecast period length according to the actual scheduling plan requirements.

[0016] Step 3-3: Using the filtered and processed water temperature, precipitation, and inflow data sequences from Step 3-2 as input layer variables for the artificial neural network, the artificial neural network, support vector machine, and random forest methods are selected. The input variables are filtered based on mutual information. Subsequently, the input sequences of the training samples are normalized using the following formula:

[0017]

[0018] In the formula: x i and x i ′ and ′ are the i-th sample and the normalized value, respectively, x max and x min These are the maximum and minimum values, respectively.

[0019] Each node in an artificial neural network receives an input signal, processes it through weighted summaries and activation functions, and then outputs a signal. An artificial neural network consists of an input layer, hidden layers, and an output layer. The network adjusts its weights using the backpropagation algorithm. The data processing formulas for its hidden and output layers are as follows:

[0020] h = σ(W′x + b′)

[0021] y = σ(w″h + b″)

[0022] In the formula, W′ and W″ are the weight matrices connecting the input layer and the hidden layer, and the hidden layer and the output layer, respectively; b′ and b″ are the bias vectors of the hidden layer and the output layer, respectively; and x and y are the input and output variables, respectively.

[0023] Support Vector Machines (SVMs) improve the robustness and generalization ability of regression by finding an optimal hyperplane in a high-dimensional space and identifying the sample points closest to the hyperplane. The formula is as follows:

[0024]

[0025] In the formula, y is the model prediction value, and W T For weights, is a nonlinear mapping function, and e is the offset between the predicted value and the actual value;

[0026] Random forests achieve regression by constructing multiple decision trees and integrating their outputs. Each tree structure is trained on randomly sampled training data, and each split node only considers randomly selected features. The final prediction is obtained by averaging the predictions of all trees.

[0027] Step 3-4: Based on the input and output variables and loss function selected in steps 3-1, 3-2, and 3-3, train the machine learning model to obtain the prediction results;

[0028] Steps 3-5: Based on the prediction results of Step 3-4, and combined with the reservoir operation procedures and constraints, the reservoir discharge flow process under normal operation conditions is proposed according to the principle of water balance. The reservoir constraints are as follows:

[0029] Water balance constraints:

[0030]

[0031] Downflow constraint:

[0032] Q out,t,min ≤Q out,t ≤Q out,t,max

[0033] Q out,t =Q out,t,p+Q out,t,s

[0034] Reservoir water level constraints:

[0035] Z t,min ≤Z t ≤Z t,max

[0036] Unit output constraints:

[0037] P min ≤P t ≤P max

[0038] In the formula, V t and V t+1 Let Q be the reservoir capacity at times t and t+1. in and Q out These represent the inbound and outbound flow rates, respectively, where Δt is the time interval, and Q is the inbound flow rate. out,t,p and Q out,t,s Z represents the turbine generator flow rate and the water discharge rate at time t, respectively. t,min and Z t,max P represents the highest and lowest allowable water levels at time t of the reservoir. max and P min These refer to the maximum output limit and guaranteed output of the reservoir hydropower station, respectively.

[0039] Steps 3-6: Based on the discharge process in Step 3-5, construct the relationship curve between the upstream reservoir discharge and the discharge at a typical downstream section. The formula is as follows:

[0040] Q t,d =f(Q) out,t ),Q out,t＝ f'(Q t,d )

[0041] Z t,d =f(Q) out,t ),Q out,t＝ f'(Z t,d )

[0042] In the formula, Q t,d and Z t,d Let f be the flow rate and water level at the downstream control station, respectively, and f' be the relationship function and inverse function between the upstream reservoir discharge flow and the flow rate at a typical downstream section, respectively.

[0043] Preferably, an ecological regulation scheme is formulated to meet the natural reproductive needs of fish. First, based on literature review and current status surveys, the suitable water temperature range for the natural reproduction of target fish is obtained. Combined with the water temperature forecast for the next 10 days, the timing for implementing ecological regulation is preliminarily determined (if the water temperature remains stable at the lower limit of the natural reproductive temperature for fish for 7 consecutive days, this can be used as the initiation condition for ecological regulation). Second, based on upstream precipitation forecasts and reservoir regulation constraints (water level drawdown rate, maximum flow output, etc.), the available water volume and maximum discharge capacity of the reservoir for ecological regulation are calculated. Third, considering the reservoir outflow range under conventional regulation, the outflow before the start of ecological regulation is determined. Based on the key hydrological index domains that cause a large scale of target fish eggs (number of days of rising water, daily flow or water level growth rate, etc.), the daily outflow for ecological regulation that meets the downstream section's rising water conditions is formulated. The specific steps in step four for formulating an ecological regulation scheme to meet the natural reproductive needs of fish are as follows:

[0044] Step 4-1: Based on the current situation survey results, obtain the suitable water temperature range for the natural reproduction of the target fish, and determine the time to carry out ecological regulation in combination with the predicted water temperature.

[0045] Step 4-2: Based on the upstream predicted precipitation and reservoir scheduling constraints, calculate the amount of water available for ecological scheduling of the reservoir and the maximum discharge flow of the reservoir.

[0046] Step 4-3: Combine the reservoir outflow range under conventional scheduling to determine the outflow before ecological scheduling is initiated. Based on the key hydrological index domains corresponding to a large scale of target fish eggs, formulate the daily outflow for ecological scheduling that meets the downstream section's rising water conditions.

[0047] Assuming the current water level at the downstream section is Z0, if the downstream section experiences a continuous rise in water level for 5 days with a daily water level increase of 0.5m, the upstream reservoir outflow can be deduced as f'(Z0+5*0.5) based on the relationship function between the upstream reservoir discharge and the downstream typical section flow / water level from step 3. The daily flow rate is:

[0048]

[0049] Preferably, in step five, the feasibility analysis steps of the ecological solution are as follows:

[0050] Step 5-1: Based on hydrological and meteorological data, analyze whether the hydrological conditions during the target fish's breeding season meet the basic requirements for scheduling activation, and check whether the reservoir's operation meets the flow conditions and water level control requirements of the ecological scheduling plan.

[0051] Step 5-2: Compare the current and future predicted inflows to confirm whether the reservoir can support the ecological scheduling process, and assess the impact of scheduling on power generation, irrigation, and flood control needs.

[0052] Step 5-3: Based on the reproductive needs of the target fish species, analyze the impact of ecological regulation on downstream water levels, temperatures, and flow rates to ensure the regulation plan meets suitable ecological environment requirements. Compare ecological regulation data implemented in other regions and years to analyze the target fish species under conditions of prolonged flooding (>4 days) and daily flow rate increase (>0.5 m³). 3 The response under scheduling conditions such as / s and water temperature of 18-24℃ was used to predict the potential effect of scheduling on fish reproduction.

[0053] Step 5-4: In response to extreme weather events and abnormal flow fluctuations, assess the risk resistance of the dispatching plan and formulate emergency control measures and adjustment mechanisms in case of emergencies.

[0054] The beneficial effects of this invention are as follows:

[0055] 1. By accurately collecting basic information on the natural reproduction of target fish species, including detailed historical and current data such as spawning fish species, egg density, water temperature, flow rate, and water level during spawning, and combining this with advanced statistical models (such as random forest models and classification regression tree models), we can deeply analyze the hydrological parameters required for natural fish reproduction. This allows us to accurately identify the most critical hydrological indicators affecting fish reproduction and simulate the suitable range of key hydrological indicators when egg density is high. In actual scheduling, ecological scheduling plans developed based on this can effectively create suitable hydrological conditions for fish reproduction. For example, the spawning scale of fish such as the four major freshwater fish species is expected to increase significantly, and the hatching rate of fish eggs and the survival rate of juvenile fish will also be improved, thereby promoting the growth of fish populations and the stability of the ecosystem.

[0056] 2. In formulating the ecological scheduling plan, factors such as upstream rainfall forecasts, reservoir scheduling constraints (e.g., water level drawdown rate, maximum flow output), and the reservoir outflow range under conventional scheduling were fully considered. This allows the reservoir to rationally calculate the available water volume and maximum discharge flow when implementing ecological scheduling, avoiding water resource waste and over-exploitation. For example, while meeting the needs of fish reproduction, the plan can also take into account other functions of the reservoir such as power generation and irrigation, achieving multi-objective optimized utilization of water resources and improving water resource utilization efficiency.

[0057] 3. Artificial intelligence methods (such as artificial neural networks, support vector machines, and random forests) are used to predict hydrological and meteorological information. Combined with reservoir operation procedures and constraints, a relationship curve between upstream reservoir discharge and downstream typical cross-section flow / water level is constructed, providing scientific data support and decision-making basis for ecological dispatching schemes. Simultaneously, the feasibility of ecological dispatching schemes is comprehensively analyzed from multiple aspects, including hydrological and reservoir operating conditions, water resources, environmental adaptability, and scheme risks. Emergency control measures and adjustment mechanisms are formulated to effectively address various actual situations, improve the reliability and stability of the dispatching scheme in practical applications, and ensure that the dispatching scheme meets the needs of fish reproduction while guaranteeing the safe operation and comprehensive benefits of the reservoir.

[0058] 4. This ecological scheduling method provides an innovative solution for the ecological operation of water conservancy projects. By implementing scientific and reasonable ecological scheduling during the fish breeding season, it is possible to effectively protect fish resources in the river ecosystem, enhance the integrity and biodiversity of the river ecosystem, promote the harmonious coexistence of water conservancy projects and the ecological environment, and provide strong support for sustainable development without affecting the basic functions of water conservancy projects. Attached Figure Description

[0059] Figure 1 This is an index diagram of the decomposition of the artificial flood peak process in this invention;

[0060] Figure 2 This invention is a key hydrological index map affecting the spawning scale of the four major Chinese carps;

[0061] Figure 3 This is a partial effect diagram of key indicators affecting the spawning scale of the four major freshwater fish species;

[0062] Figure 4 This is a graph showing the relationship between the water temperature range and the number of fish egg collections according to the present invention;

[0063] Figure 5 This invention is a diagram showing the outflow process of the Three Gorges Reservoir in May during ecological scheduling.

[0064] Figure 6 This is a diagram showing the water level and temperature processes at key hydrological stations during the ecological scheduling period of this invention.

[0065] Figure 7 This invention provides a 24-hour precipitation forecast map for the Yangtze River basin.

[0066] Figure 8 This invention provides a 48-hour precipitation forecast map for the Yangtze River basin.

[0067] Figure 9 This invention provides a 72-hour precipitation forecast map for the Yangtze River basin.

[0068] Figure 10This invention provides a 96-hour precipitation forecast map for the Yangtze River basin.

[0069] Figure 11 This invention provides a 120-hour precipitation forecast map for the Yangtze River basin.

[0070] Figure 12 This invention provides a 144-hour precipitation forecast map for the Yangtze River basin.

[0071] Figure 13 This invention provides a 168-hour precipitation forecast map for the Yangtze River basin.

[0072] Figure 14 This invention provides a 192-hour precipitation forecast map for the Yangtze River basin.

[0073] Figure 15 This invention provides a 216-hour precipitation forecast map for the Yangtze River basin.

[0074] Figure 16 This invention provides a 240-hour precipitation forecast map for the Yangtze River basin. Detailed Implementation

[0075] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. The embodiments of the present invention are given for illustrative and descriptive purposes only, and are not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described to better illustrate the principles and practical application of the invention, and to enable those skilled in the art to understand the invention and design various embodiments with various modifications suitable for a particular purpose.

[0076] Example:

[0077] A method for ecological regulation of fish reproduction based on natural weather processes includes the following steps:

[0078] From May to July each year, fish eggs of the four major freshwater fish species (Crocodile, Caucasus, and Carinthus fragrans) are collected at designated locations along a section of the Yangtze River. Daily data on egg runoff, water temperature, flow rate, and water level during the sampling period are obtained. Hydrological process curves for May to July are plotted to determine the start and end times of each flood peak, and hydrological characteristic indicators of each flood peak are calculated accordingly. Figure 1 Using the total number of fish eggs generated during a single flood peak as the dependent variable and nine hydrological characteristic indicators as independent variables, an eco-hydrological dataset corresponding to the flood peak series was obtained. An eco-hydrological response model was established using the random forest model method. Analysis revealed that the duration of the flood rise, the daily rate of increase in flow, and the difference in water level between the preceding and following flood peaks are key hydrological indicators affecting the abundance of fish eggs of the four major freshwater fish species in the Shashi section of the Yangtze River. Simulations showed that the suitable hydrological conditions for a large fish egg population are: flood rise duration > 4 days, daily water level rise rate > 0.5 m, and water level difference between the preceding and following flood peaks > 1.4 m. Figure 2 , Figure 3 ).

[0079] Water temperature is a key factor affecting the natural reproduction of fish. Under natural conditions, the lower limit of water temperature for the natural reproduction of the four major Chinese freshwater fish species is 18℃, and the suitable water temperature is 18-24℃. Based on existing observational data, frequency analysis shows that when the water temperature in a river section exceeds 19℃, the four major Chinese freshwater fish species enter their peak reproductive period (spawning frequency gradually increases, see...). Figure 4 Therefore, the timing of ecological regulation can be determined by measuring the water temperature on-site.

[0080] According to hydrological and meteorological forecasts in early May, the water temperature in the river section has remained stable above 18℃ for several consecutive days in mid-May, indicating that the ecological regulation window period can be entered.

[0081] To formulate a joint ecological dispatching plan, machine learning models such as random forests were first used to predict the changes in hydrological and meteorological forecasts for the next ten days. Combined with the principles of water balance, reservoir operation procedures, and constraints, the reservoir discharge process under conventional dispatching conditions was simulated. Based on upstream precipitation forecasts and reservoir dispatching constraints, a daily outflow rate for ecological dispatching that meets downstream water level rise conditions was determined.

[0082] Combination Figures 7-16 The Yangtze River Basin precipitation forecast map shown includes the following specific data on upstream precipitation: Two precipitation events are expected in the Yangtze River Basin over the next 10 days. On the 19th, moderate to locally heavy rain is expected in the middle and lower reaches of the Jinsha River, the Wanyi section of the Three Gorges Dam, the southern part of the Han River main stream, the western part of the middle reaches of the Yangtze River, and the southern part of the Two Lakes system. Daily rainfall is expected to be 10-20 mm in the middle and lower reaches of the Jinsha River, the Wanyi section, the Han River, the Qing River, the Jianghan Plain, the Lishui River, the Xiang River, and the Gan River. From the 24th to the 28th, a moderate to locally heavy rain event will occur in the Yangtze River Basin from northwest to southeast. The cumulative rainfall is expected to be 30-50 mm in the lower reaches of the Jinsha River, the Wujiang River, the Xiangcun section, the Cunwan section, the Lishui River, and the Two Lakes system. Water temperatures are expected to rise over the next 10 days.

[0083] Based on the precipitation forecast map of the Yangtze River Basin, the following ecological dispatching plan is formulated:

[0084] Considering the anticipated rainfall and upstream reservoir operations, the inflow to the reservoir in the coming week is expected to be around 3100 m³. 3 The reservoir has been conducting ecological regulation trials since May 18th, with an average daily outflow of 2600 m³ / s. 3 / s, from the 19th to the 22nd, at a rate of 300-500m per day 3 With the increase in outflow, the reservoir water level is expected to be around 374.8m on the 24th.

[0085] Considering the ecological regulation of the Jinxia cascade reservoir and the anticipated rainfall, the inflow to the Three Gorges Reservoir is expected to fluctuate, with an average daily inflow of 11,500 m³ / h from the 19th to the 24th. 3 Approximately [number] m³ / s. Considering the ecological dispatching experiment starting on the 19th, with an average daily outflow of 10,000 m³ / s. 3 / s, from the 20th to the 23rd, at a daily rate of 1200-1800m 3 With the increase in outflow, the reservoir water level is expected to be around 152.6m on the 24th. Figure 5 ).

[0086] Following the Three Gorges ecological regulation experiment, the outflow from the Three Gorges Reservoir gradually increased, and the water level at the downstream Shashi hydrological station gradually rose. Water temperatures at the Shashi and Yichang stations fluctuated around 21℃, meeting the natural reproductive needs of the four major freshwater fish species. Figure 6 ).

[0087] Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art and related fields based on the embodiments of the present invention without inventive effort should fall within the scope of protection of the present invention. Structures, devices, and operating methods not specifically described and explained in the present invention, unless otherwise specified or limited, shall be implemented according to conventional means in the art.

Claims

1. A method for ecological regulation of fish reproduction based on natural weather processes, characterized in that, Includes the following steps: Step 1: Collect basic information on the natural reproduction and hydrological conditions of the target fish species; Step 2: Analyze the hydrological parameters required for the natural reproduction of the target fish species using statistical models; Step 3: Based on hydrological and meteorological forecast information, analyze weather and runoff processes, and formulate a routine reservoir operation plan; Step 4: Based on weather and runoff processes, develop an ecological scheduling plan that meets the hydrological needs for the natural reproduction of the target fish species; Step 5: Analyze the feasibility of the ecological scheduling scheme, and then implement the ecological scheduling experiment.

2. The method for ecological regulation of fish natural reproduction based on natural weather processes according to claim 1, characterized in that: In step one, the natural reproduction and hydrological basic information of the target fish includes historical data and current data, both of which include data on the species of spawning fish, fish egg density, water temperature at spawning, flow rate at spawning, and water level at spawning.

3. The method for ecological regulation of fish natural reproduction based on natural weather processes according to claim 1, characterized in that: In step two, hydrological change characteristic indicators are calculated based on the natural reproduction of the target fish and its basic hydrological information. A set of daily fish egg density and hydrological change characteristic indicators is obtained. The random forest model method is used to identify hydrological change characteristic indicators that affect the natural reproduction of the target fish, construct a model of the relationship between fish egg density and hydrological change characteristic indicators, and simulate and calculate the appropriate range of hydrological change characteristic indicators when the fish egg scale is large.

4. The method for ecological regulation of fish natural reproduction based on natural weather processes according to claim 1, characterized in that: In step three, the hydrological and meteorological forecast information includes air temperature, precipitation, water temperature and flow information. Artificial neural networks, support vector machines and random forests are selected to analyze weather and runoff processes, predict the changes in hydrological and meteorological forecast information for the next ten days, and simulate the reservoir discharge flow process under normal scheduling conditions by combining reservoir scheduling operation procedures and constraints. Through hydrological sequence data, the relationship curve between upstream reservoir discharge flow and downstream typical section flow is constructed.

5. The method for ecological regulation of fish natural reproduction based on natural weather processes according to claim 4, characterized in that: The method for constructing the relationship curve between the upstream reservoir discharge and the discharge at a typical downstream section is as follows: Step 3-1: Collect measured hydrological and meteorological information as input variables, including but not limited to temperature, precipitation, relative humidity, wind speed, and flow rate; Step 3-2: Select predictor variables as input variables, including but not limited to water temperature, precipitation, and inflow. The output variables are the water temperature, precipitation, and inflow predicted by the machine learning model for the next ten days based on these predictor variables. Set the training period and the validation period, and set the forecast period length according to the actual scheduling plan requirements. Step 3-3: Using the filtered and processed water temperature, precipitation, and inflow data sequences from Step 3-2 as input layer variables for the artificial neural network, the artificial neural network, support vector machine, and random forest methods are selected. The input variables are filtered based on mutual information. Subsequently, the input sequences of the training samples are normalized using the following formula: In the formula: x i and x i x represents the i-th sample and the normalized value, respectively. max and x min These are the maximum and minimum values, respectively. Each node in an artificial neural network receives an input signal, processes it through weighted summaries and activation functions, and then outputs a signal. An artificial neural network consists of an input layer, hidden layers, and an output layer. The network adjusts its weights using the backpropagation algorithm. The data processing formulas for its hidden and output layers are as follows: h = σ(W′x + b′) y = σ(W″h + b″) In the formula, W′ and W″ are the weight matrices connecting the input layer and the hidden layer, and the hidden layer and the output layer, respectively; b′ and b″ are the bias vectors of the hidden layer and the output layer, respectively; and x and y are the input and output variables, respectively. Support Vector Machines (SVMs) improve the robustness and generalization ability of regression by finding an optimal hyperplane in a high-dimensional space and identifying the sample points closest to the hyperplane. The formula is as follows: In the formula, y is the model prediction value, and W T For weights, is a nonlinear mapping function, and e is the offset between the predicted value and the actual value; Random forests achieve regression by constructing multiple decision trees and integrating their outputs. Each tree structure is trained on randomly sampled training data, and each split node only considers randomly selected features. The final prediction is obtained by averaging the predictions of all trees. Step 3-4: Based on the input and output variables and loss function selected in steps 3-1, 3-2, and 3-3, train the machine learning model to obtain the prediction results; Steps 3-5: Based on the prediction results of Step 3-4, and combined with the reservoir operation procedures and constraints, the reservoir discharge flow process under normal operation conditions is proposed according to the principle of water balance. The reservoir constraints are as follows: Water balance constraints: Downflow constraint: Q out,t,min ≤Q out,t ≤Q out,t,max Q out,t =Q out,t,p +Q out,t,s Reservoir water level constraints: WITH t,min ≤Z t ≤Z t,max Unit output constraints: P min ≤P t ≤P max In the formula, V t and V t+1 Let Q be the reservoir capacity at times t and t+1. in and Q out These represent the inbound and outbound flow rates, respectively, where Δt is the time interval, and Q is the inbound flow rate. out,t,p and Q out,t,s Z represents the turbine generator flow rate and the water discharge rate at time t, respectively. t,min and Z t,max P represents the highest and lowest allowable water levels at time t of the reservoir. max and P min These refer to the maximum output limit and guaranteed output of the reservoir hydropower station, respectively. Steps 3-6: Based on the discharge process in Step 3-5, construct the relationship curve between the upstream reservoir discharge and the discharge at a typical downstream section. The formula is as follows: Q t,d =f(Q out,t ),Q out,t＝ f’(Q t,d ) Z t,d =f(Q out,t ),Q out,t＝ f’(Z t,d ) In the formula, Q t,d and Z t,d Let f be the flow rate and water level at the downstream control station, respectively, and f' be the relationship function and inverse function between the upstream reservoir discharge flow and the flow rate at a typical downstream section, respectively.

6. The method for ecological regulation of fish natural reproduction based on natural weather processes according to claim 5, characterized in that: In step four, the specific steps for formulating an ecological scheduling plan that meets the natural reproductive needs of fish are as follows: Step 4-1: Based on the current situation survey results, obtain the suitable water temperature range for the natural reproduction of the target fish, and determine the time to carry out ecological regulation in conjunction with the predicted water temperature process. Step 4-2: Based on the upstream predicted precipitation and reservoir scheduling constraints, calculate the amount of water available for ecological scheduling of the reservoir and the maximum discharge flow of the reservoir. Step 4-3: Combine the reservoir outflow range under conventional scheduling to determine the outflow before ecological scheduling is initiated. Based on the key hydrological index domains corresponding to a large scale of target fish eggs, formulate the daily outflow for ecological scheduling that meets the downstream section's rising water conditions.

7. The method for ecological regulation of fish natural reproduction based on natural weather processes according to claim 6, characterized in that: In step five, the feasibility analysis steps for the ecological solution are as follows: Step 5-1: Based on hydrological and meteorological data, analyze whether the hydrological conditions during the target fish's breeding season meet the basic requirements for scheduling activation, and check whether the reservoir's operation meets the flow conditions and water level control requirements of the ecological scheduling plan. Step 5-2: Compare the current and future predicted inflows to confirm whether the reservoir can support the ecological scheduling process, and assess the impact of scheduling on power generation, irrigation, and flood control needs. Step 5-3: Based on the reproductive needs of the target fish species, analyze the impact of ecological regulation on downstream water level, temperature, and flow velocity conditions to ensure that the regulation plan meets suitable ecological environment requirements. Compare ecological regulation data implemented in other regions and years to analyze the target fish species under conditions of continuous flooding for >4 days and daily flow rate increase >0.5 m³ / h. 3 The response under scheduling conditions of / s and water temperature of 18-24℃ was used to predict the potential effect of scheduling on fish reproduction. Step 5-4: In response to extreme weather events and abnormal flow fluctuations, assess the risk resistance of the dispatching plan and formulate emergency control measures and adjustment mechanisms in case of emergencies.

Images