A method and device for predicting a water bloom in a backwater area

Abstract

The application provides a backwater area water bloom prediction method and device, and relates to the technical field of environmental science. The method comprises the following steps: acquiring geographical hydrological data and environmental monitoring data of a target hydrological environment, and constructing a backwater area three-dimensional hydrodynamic water temperature model according to the geographical hydrological data and the environmental monitoring data; constructing a backwater area three-dimensional eutrophication model based on water temperature, water depth, horizontal direction flow velocity and wind speed obtained from the backwater area three-dimensional hydrodynamic water temperature model; constructing a BP neural network model based on the backwater area three-dimensional eutrophication model and training the BP neural network model; and obtaining a water bloom risk prediction result of the backwater area of the target hydrological environment by using the neural network model. The application provides data support for chlorophyll a calculation of the backwater area eutrophication model by constructing the backwater area three-dimensional hydrodynamic water temperature model, without the need to add a large number of high-frequency detection equipment, thereby greatly reducing the model operation time and improving the model prediction efficiency.

Description

Technical Field

[0001] This application relates to the field of environmental science and technology, and in particular to a method and apparatus for predicting algal blooms in backwater areas. Background Technology

[0002] Water resource management and environmental science are crucial fields in modern society, particularly in the construction of reservoirs and dams, which greatly promote the rational utilization and allocation of water resources. Reservoirs and dams not only address basic human needs for water, such as water supply, irrigation, power generation, flood control, and navigation, but also significantly improve water resource utilization efficiency and alleviate the problems of uneven water distribution and water scarcity. Furthermore, these water conservancy facilities bring significant social and economic benefits, but they also bring water pollution and water quality safety issues, which are increasingly exacerbated by climate change and increased human activities.

[0003] Especially during reservoir impoundment and water transfer, reservoirs and their closely connected backwater areas face complex water quality challenges. Rising water levels and slower flow rates in these areas reduce water dispersion capacity and prolong pollutant retention time, frequently leading to problems such as eutrophication and algal blooms. These issues not only affect water quality but may also endanger drinking water safety and public health, making early warning and prediction of algal blooms a crucial task for ensuring water quality safety.

[0004] Currently, algal bloom prediction mainly relies on two types of models: process mechanism models and data-driven models. Process mechanism models simulate the growth environment of phytoplankton by analyzing the interaction between phytoplankton and environmental factors, while data-driven models rely on statistical analysis of a large number of environmental variables to predict phytoplankton growth. However, these models have encountered many challenges in their specific applications in backwater areas. Due to the special hydrological and hydrodynamic conditions in backwater areas, general river and lake models are often unsuitable, necessitating the development of specialized algal bloom prediction models.

[0005] The main drawback of existing technologies lies in the fact that process mechanism models rely on a large amount of high-frequency monitoring data to ensure the accuracy of simulations, but this is often limited by economic and technological conditions in practice. Furthermore, the computational process of such models is complex and time-consuming, affecting the timeliness of predictions. While data-driven models can improve prediction efficiency, they require a large amount of high-quality and diverse data as support in the early stages, which is difficult to achieve when data collection is limited. These limitations highlight the inadequacies of existing algal bloom prediction technologies in addressing algal bloom problems in backwater areas, particularly in the challenges of rapidly responding to environmental changes and providing accurate early warnings. Summary of the Invention

[0006] The purpose of this invention is to provide a method and apparatus for predicting algal blooms in backwater areas. The method for predicting algal blooms in backwater areas can achieve efficient and accurate prediction of algal blooms in reservoir backwater areas under limited data conditions.

[0007] In order to achieve the above-mentioned objectives of the present invention, the following technical solution is adopted:

[0008] In a first aspect, the present invention provides a method for predicting algal blooms in backwater areas, comprising:

[0009] Obtain geographic hydrological data and environmental monitoring data of the target hydrological environment, and construct a three-dimensional hydrodynamic and temperature model of the backwater area based on the geographic hydrological data and the environmental monitoring data;

[0010] Based on the water temperature, water depth, horizontal flow velocity and wind speed obtained from the three-dimensional hydrodynamic and temperature model of the backwater area, a three-dimensional eutrophication model of the backwater area is constructed.

[0011] Based on the three-dimensional eutrophication model of the backwater area, a BP neural network model was constructed and trained.

[0012] The risk prediction results of algal blooms in the backwater area of ​​the target hydrological environment are obtained by using a trained neural network model.

[0013] In an optional implementation, the step of acquiring geographic hydrological data and environmental monitoring data of the target hydrological environment, and constructing a three-dimensional hydrodynamic and temperature model of the backwater area based on the geographic hydrological data and the environmental monitoring data, includes:

[0014] Collect the geographic hydrological data and environmental monitoring data of the target hydrological environment, construct a one-dimensional hydrological model of the inflow river, and calculate the riverbed roughness data in the target hydrological environment;

[0015] Based on the riverbed roughness data and the outflow data of the upstream reservoir of the inflow river in the target hydrological environment, a one-dimensional hydrodynamic model of the inflow river is constructed, and the inflow data and upstream flow data are calculated.

[0016] Based on the inflow data and the measured water level data of the reservoir, a two-dimensional hydrodynamic model of the reservoir is constructed, and the downstream water level data of the inflow river is calculated.

[0017] Using the upstream flow data and the measured water level data of the reservoir, a three-dimensional hydrodynamic and temperature model of the backwater zone is constructed.

[0018] In an optional implementation, the step of collecting the geographic hydrological data and environmental monitoring data of the target hydrological environment, constructing a one-dimensional hydrological model of the inflow river, and calculating the riverbed roughness data of the target hydrological environment includes:

[0019] The geographic hydrological data and environmental monitoring data of the target hydrological environment are collected; wherein, the geographic hydrological data includes the surface type of the rivers flowing into the reservoir, the reservoir topographic map, and land use remote sensing data; the environmental monitoring data includes rainfall data and evaporation data;

[0020] Based on the aforementioned geographical and hydrological data and environmental monitoring data, the inflowing river is divided into multiple catchment areas;

[0021] A one-dimensional hydrological model of the inflow river based on the NAM hydrological model is constructed, and the riverbed roughness data is obtained by calibrating each catchment area of ​​the inflow river using the one-dimensional hydrological model of the inflow river.

[0022] In an optional implementation, the step of constructing a one-dimensional hydrodynamic model of the inflow river based on the riverbed roughness data and the outflow data of the upstream reservoir of the inflow river in the target hydrological environment, and calculating the inflow data and upstream flow data, includes:

[0023] Obtain the river system map and topological relationship map of the rivers flowing into the reservoir, and perform generalization processing on the river channels of the rivers flowing into the reservoir and the reservoir.

[0024] Based on the generalized river channels and reservoirs, the external and internal boundary conditions of the model are determined; wherein, the external boundary conditions include the upstream water boundary of the river channel and the downstream water level boundary of each river channel; the internal boundary conditions include the setting of hydraulic structures on the river channel.

[0025] Based on the external and internal boundary conditions, the MIKE 11HD hydrodynamic calculation model is coupled with the one-dimensional hydrological model of the inflow river to construct the one-dimensional hydrodynamic model of the inflow river, and the upstream flow data and the inflow flow data of the reservoir are calculated.

[0026] In an optional implementation, the calculation expression of the MIKE 11HD hydrodynamic calculation model is: Where x and t represent the spatial and temporal coordinates of the calculation point, respectively; A represents the cross-sectional area of ​​the water passage; Q represents the flow rate; h represents the water level; q represents the lateral inflow flow rate; C represents the Chezy coefficient; R represents the hydraulic radius; α represents the momentum correction coefficient; and g represents the gravitational acceleration.

[0027] In an optional implementation, the step of constructing a two-dimensional hydrodynamic model of the reservoir based on the inflow data and the measured water level data of the reservoir, and calculating the downstream water level data of the inflow river, includes:

[0028] The reservoir is then processed into a two-dimensional grid using the reservoir topographic map.

[0029] Based on the reservoir topographic map processed by two-dimensional gridding, the land boundary conditions and water boundary conditions of the model are determined; wherein, the land boundary conditions are obtained by the inflow data of the inflow river based on the one-dimensional hydrodynamic model of the inflow river; the water boundary conditions are determined by the inflow process of the reservoir at the beginning of the century, the measured water level and reservoir capacity curves.

[0030] A two-dimensional hydrodynamic model of the reservoir is constructed. Feature points are selected in the reservoir, and the measured water level data is used to calibrate the constructed two-dimensional hydrodynamic model of the reservoir to obtain the downstream water level data of the inflow river.

[0031] In an optional implementation, the step of constructing a three-dimensional hydrodynamic and temperature model of the backwater zone using the upstream flow data and the measured water level data of the reservoir includes:

[0032] The backwater area of ​​the target hydrological environment is subjected to three-dimensional meshing; wherein, the three-dimensional meshing includes horizontal triangular and quadrilateral meshes, as well as vertical meshes; the middle channel of the river channel of the inflowing river adopts the quadrilateral mesh; the area outside the middle channel of the river channel adopts the triangular mesh; the vertical mesh adopts a hybrid layered form of sigma and Z-level;

[0033] Based on the three-dimensional meshing of the backwater area, a three-dimensional hydrodynamic model of the backwater area is constructed.

[0034] The upstream and downstream boundary conditions of the backwater area are determined; wherein, the upstream boundary conditions are obtained by using the upstream flow data calculated by the one-dimensional hydrodynamic model of the inflow river as input conditions; and the boundary between the backwater area and the reservoir is used as the downstream boundary conditions.

[0035] Using the three-dimensional hydrodynamic model of the backwater area, and based on the upstream and downstream boundary conditions, the water depth and horizontal velocity of the backwater area are calculated.

[0036] Based on the three-dimensional hydrodynamic model of the backwater area combined with the water temperature model, a three-dimensional hydrodynamic and water temperature model of the backwater area is constructed.

[0037] In an optional implementation, the transport equation for the temperature model in the three-dimensional hydrodynamic temperature model of the backwater zone is: in, The turbulent diffusion coefficient represents the vertical direction; F represents the source term generated by heat exchange with the atmosphere; tRepresents the horizontal diffusion term; T represents temperature; u, v, and w represent the velocity components of the water flow in the x, y, and z directions, respectively;

[0038] The horizontal diffusion term F t The equation is: in, Represents the turbulent diffusion coefficient in the horizontal direction;

[0039] The boundary condition for surface temperature is: when Z = η, in, η represents the turbulent diffusion coefficient in the vertical direction; z represents the vertical position coordinate; η represents the water surface position. Q represents the temperature gradient in the vertical direction; n ρ0 represents the net heat flux, a measure of heat exchange between water and the atmosphere; ρ0 represents the density of water; c p Represents the specific heat capacity of water;

[0040] The bottom boundary condition is: when z = -d,

[0041] The heat exchange between water and air surfaces is calculated based on the following physical process: Q n =Q v +Q c +βq sr,net +Q lr,net ; Among them, Q n Represents the net heat flux of the surface; c p q represents the specific heat capacity of water; v Represents latent heat flux; q c q represents sensible heat flux; sr,net Represents net shortwave radiation; q lr,net Represents net longwave radiation; Heat received or lost per unit mass of water.

[0042] In an optional implementation, the construction of a three-dimensional eutrophication model of the backwater area based on the water temperature, water depth, horizontal flow velocity, and wind speed obtained from the three-dimensional hydrodynamic and temperature model of the backwater area includes:

[0043] A three-dimensional eutrophication model of the backwater zone is constructed based on a preset eutrophication template.

[0044] The boundary conditions and model forces of the three-dimensional eutrophication model of the backwater area are determined; the boundary conditions include the upstream river boundary conditions and the downstream reservoir confluence boundary conditions.

[0045] Based on the boundary conditions, the water temperature, water depth, horizontal flow velocity, and wind speed obtained from the three-dimensional hydrodynamic and temperature model of the backwater area, as well as the light intensity obtained from the meteorological station, are used as the forces acting on the model, and the three-dimensional eutrophication model of the backwater area is calibrated.

[0046] In an optional implementation, the step of constructing a BP neural network model based on the three-dimensional eutrophication model of the backwater area and training the BP neural network model includes:

[0047] Based on downstream water level, upstream flow rate, air temperature, light intensity, upstream water quality, and downstream water quality, different scenario schemes are constructed. Based on the different scenario schemes, chlorophyll a concentration is calculated through the three-dimensional eutrophication model of the backwater area to form a training set.

[0048] Construct a BP neural network model;

[0049] The BP neural network model is trained by using the parameters in the training set as the model input and the chlorophyll a concentration to be fitted as the model output, thus obtaining the trained BP neural network model.

[0050] In an optional implementation, training the BP neural network model further includes:

[0051] Principal component analysis is used to reduce the dimensionality of the parameters in the training set to obtain the model input terms;

[0052] The model input terms are standardized to obtain standardized model input terms; wherein the expression for the standardization process is: Where x represents the original model input; xˋ represents the standardized model input; μ represents the data mean; σ represents the data standard deviation;

[0053] Based on the cross-validation strategy, the standardized model input is divided into k folds, where k-1 folds are used as the training data set, and the 1-fold samples are combined with NSE, RMSE, and MAE evaluation metrics to form the validation data set.

[0054] The following training process is executed k times: the BP neural network model is trained based on the training data set and validated using the validation data set to obtain the validation result;

[0055] The verification results from k trials are taken as the training score of the BP neural network model, and the BP neural network model is adjusted based on the training score.

[0056] In an optional implementation, the construction of the BP neural network model further includes:

[0057] The number of hidden layers in the BP neural network model is set to a single hidden layer; wherein the number of neurons in the single hidden layer is set to 1 / 3 to 2 times that of the input layer neurons;

[0058] The evaluation results were obtained by using the NSE index.

[0059] If the evaluation result shows that the NSE value reaches the preset threshold, then the performance is determined to have met expectations.

[0060] If the evaluation result is that the NSE value does not reach the preset threshold, it is determined that the performance has not met expectations, and the number of hidden layers is set to double hidden layers; wherein, in the double hidden layers, the number of neurons in the first layer is set to 1 / 3 to 2 times that of the input layer neurons; and the number of neurons in the second layer is set to 2 / 3 times that of the input layer neurons.

[0061] In an optional implementation, the construction of the BP neural network model further includes:

[0062] The weights of the BP neural network model are initialized based on the Xavier method; wherein the expression for weight initialization is: Where ω represents the weight matrix; U represents a uniform distribution; n in n represents the number of neurons in the input layer of the neural network. out μ represents the number of neurons in the output layer of the neural network; μ is determined by the type of activation function selected; the activation function is any one of Tanh, Sigmoid, and Softsign.

[0063] If the activation function is Tanh, then μ = 6; if the activation function is Sigmoid, then μ = 96; if the activation function is ReLU, then μ = 12.

[0064] The BP neural network model is trained using different activation functions, and evaluated using performance evaluation metrics to obtain the evaluation results for each activation function.

[0065] The activation function Tanh is: The activation function Sigmoid is: The activation function Softsign is: Where x represents the sum of the input signals, including the weighted and biased terms of the previous layer;

[0066] Based on the evaluation results, the activation function used in the BP neural network model is determined.

[0067] In an optional implementation, the construction of the BP neural network model further includes:

[0068] Based on the chlorophyll a concentration, a time series model is established; wherein, the expression of the time series model is: Where xt represents the actual observed value at time point T in the time series; Represents the autoregressive coefficient; ε t ε represents the error term at time point t; t-1 ε t-2 ε t-q This represents the first q error terms in the time series;

[0069] Using lag time as the input variable, the chlorophyll a concentration data at the lag time is calculated based on the time series model and used as the time training set;

[0070] The BP neural network model is trained based on the time training set and evaluated using performance evaluation metrics to obtain evaluation results corresponding to different lag times. The BP neural network model is then adjusted based on the evaluation results corresponding to different lag times.

[0071] In an optional implementation, the algal bloom risk prediction result includes: if the chlorophyll a concentration Ccompound of the backwater area of ​​the target hydrological environment is ≤10, then the algal bloom risk prediction result is: algal bloom severity level 1; if the chlorophyll a concentration Ccompound of the backwater area of ​​the target hydrological environment is 10<C≤15, then the algal bloom risk prediction result is: algal bloom severity level 2; if the chlorophyll a concentration Ccompound of the backwater area of ​​the target hydrological environment is 15<C≤50, then the algal bloom risk prediction result is: algal bloom severity level 3; if the chlorophyll a concentration Ccompound of the backwater area of ​​the target hydrological environment is 50<C≤100, then the algal bloom risk prediction result is: algal bloom severity level 4; if the chlorophyll a concentration Ccompound of the backwater area of ​​the target hydrological environment is C>100, then the algal bloom risk prediction result is: algal bloom severity level 5.

[0072] Secondly, the present invention also provides a device for predicting algal blooms in backwater areas, comprising:

[0073] The environmental data module is used to acquire geographic hydrological data and environmental monitoring data of the target hydrological environment, and to construct a three-dimensional hydrodynamic and temperature model of the backwater area based on the geographic hydrological data and the environmental monitoring data.

[0074] The eutrophication module is also used to construct a three-dimensional eutrophication model of the backwater area based on the water temperature, water depth, horizontal flow velocity and wind speed obtained from the three-dimensional hydrodynamic and temperature model of the backwater area.

[0075] The neural network module is also used to construct a BP neural network model based on the three-dimensional eutrophication model of the backwater area, and to train the BP neural network model.

[0076] The risk prediction module is used to obtain the algal bloom risk prediction results of the backwater area of ​​the target hydrological environment using a trained neural network model.

[0077] The algal bloom prediction method for backwater areas provided in this application constructs a three-dimensional hydrodynamic and temperature model of the backwater area, providing data support for chlorophyll a calculation in the eutrophication model of the backwater area without requiring a large number of additional high-frequency detection devices. A neural network big data model trained on mechanistic model data reconstructs the chlorophyll a prediction calculation mechanism, significantly reducing model computation time and improving model prediction efficiency while ensuring prediction accuracy (the Nash efficiency coefficient of chlorophyll a prediction results 3 days in advance reaches 0.972). Attached Figure Description

[0078] To more clearly illustrate the technical solutions of this application, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of this application and therefore should not be considered as a limitation on the scope of protection of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0079] Figure 1 This is a flowchart illustrating Embodiment 1 of the algal bloom prediction method for backwater areas of the present invention;

[0080] Figure 2 This is a detailed flowchart of step S100 in Embodiment 2 of the algal bloom prediction method for backwater areas of the present invention;

[0081] Figure 3 This is a detailed flowchart of step S200 in Embodiment 3 of the algal bloom prediction method for backwater areas of the present invention;

[0082] Figure 4 This is a detailed flowchart of step S300 in Example 4 of the algal bloom prediction method for backwater areas of the present invention;

[0083] Figure 5 This is a schematic diagram of the module connection of the algal bloom prediction device in the backwater area in an embodiment of the present invention. Detailed Implementation

[0084] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0085] The components of the embodiments of this application described and illustrated in the accompanying drawings can be arranged and designed in a variety of different configurations. Therefore, the following detailed description of the embodiments of this application provided in the drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0086] In the following, the terms “comprising,” “having,” and their cognates, which may be used in various embodiments of this application, are intended only to indicate a particular feature, number, step, operation, element, component, or combination thereof, and should not be construed as excluding, firstly, the presence of one or more other features, numbers, steps, operations, elements, components, or combinations thereof, or adding the possibility of one or more features, numbers, steps, operations, elements, components, or combinations thereof.

[0087] Furthermore, the terms "first," "second," and "third" are used only to distinguish descriptions and should not be interpreted as indicating or implying relative importance.

[0088] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of this application pertain. Terms (such as those defined in commonly used dictionaries) shall be interpreted as having the same meaning as in their contextual meaning in the relevant technical field and shall not be construed as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of this application.

[0089] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0090] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0091] Example 1:

[0092] refer to Figure 1 Embodiment 1 of this application provides a method for predicting algal blooms in backwater areas, including:

[0093] Step S100: Obtain geographic hydrological data and environmental monitoring data of the target hydrological environment, and construct a three-dimensional hydrodynamic and temperature model of the backwater area based on the geographic hydrological data and the environmental monitoring data.

[0094] The target hydrological environment, as mentioned above, refers to the area where algal bloom prediction is required. This area can possess certain hydrological and environmental characteristics, such as naturally formed or artificial inflowing rivers, waterways, reservoirs, and other topographical and environmental features. Furthermore, it can also target wetlands, estuaries and nearshore sea areas, lakes and reservoirs, etc.

[0095] Each type of hydrological environment has its unique hydrodynamic characteristics and environmental conditions, which determine the risks and management strategies of eutrophication. Therefore, algal bloom prediction methods need to take these environmental characteristics into account and be adjusted and optimized according to the specific hydrological environment. Selecting an appropriate target hydrological environment is crucial to ensuring the effectiveness and applicability of the prediction model.

[0096] As mentioned above, this step involves collecting geographic hydrological data and environmental monitoring data of the target hydrological environment.

[0097] Geographic and hydrological data can include the geographical location of water bodies, drainage area, river course, and riverbed topography, while environmental monitoring data covers water temperature, flow velocity, water level, and rainfall.

[0098] These data were used to construct a three-dimensional hydrodynamic and temperature model that can describe the physical and thermodynamic behavior of water bodies, such as the direction and velocity of water flow, and the distribution and variation of water temperature.

[0099] This model can accurately simulate the dynamic and thermodynamic characteristics of specific water bodies, helping to understand complex water flow and temperature changes and providing a foundation for subsequent algal bloom prediction.

[0100] Step S200: Based on the water temperature, water depth, horizontal flow velocity and wind speed obtained from the three-dimensional hydrodynamic and temperature model of the backwater area, construct a three-dimensional eutrophication model of the backwater area.

[0101] Based on the hydrodynamic and temperature model obtained above, this step further constructs a three-dimensional eutrophication model of the backwater area. This model uses the water temperature, water depth, horizontal flow velocity, and wind speed data obtained in the previous step to simulate the movement of nutrients such as nitrogen and phosphorus and the dynamics of phytoplankton growth, thereby predicting the risk of eutrophication and algal blooms.

[0102] By combining with hydrodynamic and temperature models, eutrophication models can more accurately simulate the behavior of nutrients and phytoplankton under complex hydrological conditions, increasing the accuracy of predictions.

[0103] Ecological modeling software such as AQUATOX or EcoLab, combined with the output of hydrodynamic models, can be used to simulate nutrient cycling and biological growth processes.

[0104] Step S300: Based on the three-dimensional eutrophication model of the backwater area, construct a BP neural network model and train the BP neural network model.

[0105] Based on the output of the three-dimensional eutrophication model of the backwater area, this step constructs a BP neural network model for training. This model will use data provided by the eutrophication model as input, such as chlorophyll a concentration and nutrient salinity, to learn the relationship between these parameters and algal blooms.

[0106] Neural networks can capture complex nonlinear relationships, improving the accuracy and efficiency of predicting algal bloom risks. Specifically, machine learning libraries such as TensorFlow or PyTorch can be used to train and validate the network by setting up a multilayer perceptron (MLP) and selecting appropriate activation functions, loss functions, and optimizers.

[0107] Step S400: Use the trained neural network model to obtain the algal bloom risk prediction result of the backwater area of ​​the target hydrological environment.

[0108] The above describes the prediction of algal bloom risk in the backwater area of ​​a target hydrological environment using a trained neural network model. The model outputs the probability and severity of algal blooms, helping decision-makers to take appropriate management measures. This step directly provides feasible prediction results, facilitating timely response to potential environmental risks and optimizing resource allocation and protection measures.

[0109] A trained neural network model can be used to input real-time or the latest collected hydrological and water quality data to obtain prediction results, which may also include the formulation of warning levels and recommended response strategies.

[0110] The steps in the above method, through precise models and advanced technologies, ensure the scientific validity and practicality of the entire algal bloom prediction method, providing an efficient technical support tool for water resource management.

[0111] In summary, the algal bloom prediction method for backwater areas provided in this embodiment constructs a three-dimensional hydrodynamic and temperature model of the backwater area, providing data support for chlorophyll a calculation in the eutrophication model of the backwater area without requiring a large number of additional high-frequency detection devices. The neural network big data model trained based on the mechanistic model data reconstructs the chlorophyll a prediction calculation mechanism, significantly reducing model computation time and improving model prediction efficiency while ensuring prediction accuracy (the Nash efficiency coefficient of chlorophyll a prediction results 3 days in advance reaches 0.972).

[0112] Example 2:

[0113] refer to Figure 2Based on the above embodiment 1, embodiment 2 of this application provides a method for predicting algal blooms in backwater areas. Step S100 involves acquiring geographic hydrological data and environmental monitoring data of the target hydrological environment, and constructing a three-dimensional hydrodynamic and temperature model of the backwater area based on the geographic hydrological data and the environmental monitoring data, including:

[0114] Step S110: Collect the geographic hydrological data and environmental monitoring data of the target hydrological environment, construct a one-dimensional hydrological model of the inflow river, and calculate the riverbed roughness data in the target hydrological environment.

[0115] The steps described above form the foundation of the entire model, involving the collection of relevant data on the target hydrological environment, including topography, land use, and hydrological cycle data (such as rainfall and evaporation). This data is used to construct a one-dimensional hydrological model, primarily for simulating changes in the inflow river's flow and water quality. This step yields fundamental data such as river flow conditions and riverbed roughness, crucial for understanding hydrodynamics and subsequent model development. This method comprehensively reflects the natural and artificial characteristics of the river, contributing to the accurate simulation and prediction of water body behavior.

[0116] Geographic information can be processed using GIS software, combined with hydrological measurement tools and techniques (such as current meters) to collect hydrological data. Hydrological models can be constructed and simulated using software such as HEC-HMS or SWAT.

[0117] As mentioned above, in hydrodynamic models, riverbed roughness is a key parameter used to describe the resistance or friction of the riverbed surface to water flow. This parameter is an important factor in evaluating water flow velocity and behavior because it directly affects the dynamic characteristics of the flow, including flow velocity and direction. Riverbed roughness reflects the influence of riverbed materials (such as sand, sediment, vegetation, etc.) and riverbed surface morphology (such as smoothness, roughness, presence or absence of slope, and tortuosity) on water flow. The higher the riverbed roughness, the greater the frictional resistance to the water flow and the slower the flow velocity; conversely, the lower the roughness, the faster the flow velocity.

[0118] Step S120: Based on the riverbed roughness data and the outflow data of the upstream reservoir of the inflow river in the target hydrological environment, construct a one-dimensional hydrodynamic model of the inflow river and calculate the inflow data and upstream flow data.

[0119] The above describes the construction of a one-dimensional hydrodynamic model of the inflow river using the riverbed roughness data obtained in the first step and the outflow data from the upstream reservoir. This model is mainly used to simulate the river's physical characteristics such as flow velocity and flow rate, thereby obtaining more detailed inflow and upstream flow data, providing a foundation for understanding the reservoir's hydrodynamic behavior.

[0120] This model can provide an accurate description of the physical behavior of water flow, which helps to predict water flow changes and potential water quality problems in the backwater zone.

[0121] Hydrodynamic simulation software such as MIKE 11 or HEC-RAS can be used to simulate water flow, which can accurately calculate various physical parameters of water flow.

[0122] Step S130: Based on the inflow data and the measured water level data of the reservoir, construct a two-dimensional hydrodynamic model of the reservoir and calculate the downstream water level data of the inflow river.

[0123] As described above, this step uses inflow and measured water level data to construct a two-dimensional reservoir hydrodynamic model to simulate the flow and level behavior of the reservoir. The model provides information on reservoir water level changes and flow conditions, which is crucial for managing reservoir operations and predicting water quality changes in the reservoir area.

[0124] Two-dimensional models provide more detailed information on spatial variations than one-dimensional models, facilitating a more comprehensive understanding of the dynamic changes in water bodies. Specifically, this can be achieved using MIKE 21 or other two-dimensional hydrodynamic simulation software, which can handle complex water body boundary and internal flow problems.

[0125] Step S140: Using the upstream flow data and the measured water level data of the reservoir, a three-dimensional hydrodynamic and temperature model of the backwater zone is constructed.

[0126] Based on the above, a three-dimensional hydrodynamic and temperature model was constructed by combining upstream flow data and measured water level data from the reservoir to simulate the three-dimensional distribution of water flow, water level, and water temperature in the backwater zone. This model can provide detailed three-dimensional dynamics of the water body in the backwater zone, including water temperature distribution, providing necessary basic data for subsequent eutrophication models.

[0127] Three-dimensional models can provide the most comprehensive description of water dynamics, especially in complex backwater areas, and help to gain a deeper understanding of the various physical, chemical and biological processes of water bodies.

[0128] Specifically, 3D models typically require advanced computing power and software support, such as MIKE 3D or DELFT3D, which can handle complex 3D hydrodynamic and water temperature problems.

[0129] The method presented in this embodiment integrates geographic hydrological and environmental monitoring data to construct hydrodynamic models from one-dimensional to three-dimensional, providing a systematic approach to predict and manage algal bloom risks in backwater areas. First, by collecting geographic hydrological and environmental monitoring data and constructing a one-dimensional hydrological model of the inflow river, this step utilizes existing data sources for decision support, optimizes resource use, and reduces the need for new monitoring facilities. Next, a one-dimensional hydrodynamic model constructed using riverbed roughness and inflow data accurately simulates water flow dynamics, providing crucial information on water flow behavior and laying an accurate foundation for water quality management. Furthermore, by combining inflow and measured water level data, a two-dimensional hydrodynamic model comprehensively analyzes the hydrodynamic and temperature distribution within the reservoir, improving water management strategies. Finally, a three-dimensional hydrodynamic and temperature model integrates environmental factors, enhancing the responsiveness to complex environmental changes and providing necessary input parameters for eutrophication models. Overall, this series of steps not only enhances the scientific rigor and predictive capabilities of water management but also strengthens the ability to predict and respond to potential algal bloom events in backwater areas, providing an effective tool for sustainable water resource management.

[0130] In summary, the above method can provide high-frequency data inputs such as flow, water level, water temperature, and water quality for the three-dimensional eutrophication model of the backwater area by constructing a coupling of the inflow river (one-dimensional), reservoir (two-dimensional), and backwater area (three-dimensional) based on existing meteorological stations, water quality monitoring stations, and hydrological stations around the backwater area, thereby driving the prediction of algal blooms in the backwater area without the need for additional high-frequency monitoring equipment.

[0131] Further, step S110 involves collecting the geographic hydrological data and environmental monitoring data of the target hydrological environment, constructing a one-dimensional hydrological model of the inflow river, and calculating the riverbed roughness data in the target hydrological environment, including:

[0132] Step S111: Collect the geographic hydrological data and environmental monitoring data in the target hydrological environment; wherein, the geographic hydrological data includes the surface type of the inflow river, the reservoir topographic map, and land use remote sensing data; the environmental monitoring data includes rainfall data and evaporation data;

[0133] The steps described above involve collecting fundamental data on the target hydrological environment, including geographic hydrological data and environmental monitoring data. Geographic hydrological data typically includes river surface types, reservoir topographic maps, land use data, etc., which help to depict the physical framework and geographical features of the entire water body. Environmental monitoring data, such as rainfall and evaporation, provide crucial dynamic information on the hydrological cycle.

[0134] This step provides access to a comprehensive dataset, which forms the basis for building hydrological models and predicting water behavior, thus ensuring the accuracy and reliability of the models as they are built upon real and detailed data input.

[0135] Specifically, this information can be collected through satellite remote sensing, ground surveys, and meteorological station data. GIS tools and remote sensing technologies can be used to analyze topographic maps and land use, while standard meteorological instruments are used to collect rainfall and evaporation data.

[0136] Step S112: Based on the geographic hydrological data and the environmental monitoring data, the inflow river is divided into multiple catchment areas;

[0137] As described above, using collected geographic hydrological data, the rivers flowing into the reservoir are divided into multiple catchment areas based on natural or artificial characteristics. This division helps to more precisely understand and simulate the behavior and characteristics of water flow in different areas. Each catchment area can be evaluated and simulated independently, resulting in higher resolution hydrological models and more targeted predictions. This zoning method allows for detailed analysis of the hydrological dynamics of specific areas, helping to identify potential risk points and key intervention areas.

[0138] Specifically, hydrological modeling software such as HEC-HMS or SWAT can be used to partition the hydrology and model the flow and runoff in each region.

[0139] Step S113: Construct a one-dimensional hydrological model of the inflow river based on the NAM hydrological model, and calibrate each catchment area of ​​the inflow river using the one-dimensional hydrological model of the inflow river to obtain the riverbed roughness data.

[0140] This step employs the NAM (Nedbor-Afstromnings Model), a popular rainfall-runoff model used to simulate the hydrological behavior of rivers flowing into reservoirs. The model utilizes zoned data to predict water flow and quality in each catchment area, thereby obtaining detailed hydrological parameters for each catchment area, such as flow rate, water level, and potential water quality issues.

[0141] The NAM model can take into account various hydrological factors and land use influences, providing accurate hydrological response predictions, especially under extreme climate events. Collected rainfall and topographic data are input into the NAM model, which calculates riverbed roughness and flow behavior.

[0142] It should be noted that the NAM hydrological model is a lumped deterministic conceptual model used to simulate rainfall generation and runoff processes within a watershed. It divides soil moisture content into four parts: snow storage, surface storage, lower zone storage, and groundwater storage (e.g., ...). Figure 2 As shown in the figure, continuous calculations are performed to simulate the corresponding hydrological processes in the watershed.

[0143] The meteorological input data for the NAM model include rainfall, evaporation capacity, and temperature (if a snowmelt model is used). The main outputs are runoff and other hydrological elements such as soil moisture content and groundwater recharge. The key parameters of the NAM model include the maximum surface aquifer water content (Umax), the maximum root zone aquifer water content (Lmax), the surface runoff concentration coefficient (CQOF), the interflow concentration time (CKIF), the surface runoff concentration time (CK12), the surface runoff generation critical value (TOF), the interflow generation critical value (TIF), the root zone groundwater recharge critical value (TG), and the baseflow concentration time (CKBF). The initial conditions of the model include the relative soil moisture content of the surface and root zone aquifers at the start time, as well as the initial values ​​for surface runoff, interflow, and baseflow. The simulation results of the NAM model include surface runoff, interflow, and baseflow processes.

[0144] In this embodiment, firstly, for the target hydrological environment, topographic maps of the reservoir and its surrounding area, land use remote sensing data, and rainfall data and evaporation data from the environmental monitoring data are collected from the geographic hydrological data. Then, based on the existing topographic data, land use types, and water system distribution, the rivers flowing into the reservoir are divided into multiple catchment areas. Finally, the NAM model is used to set and adjust the hydrological parameter values ​​(including Umax, Lmax, CQOF, CK12, etc.) of each river flowing into the reservoir, and model calibration will be carried out subsequently.

[0145] In summary, this embodiment divides the inflow rivers into zones based on their elevation data, channel distribution, and land use types. Combined with rainfall and evaporation data, the MIKE11 NAM model is used to simulate the rainfall-runoff process of each inflow river. Subsequently, this is coupled with a one-dimensional hydrodynamic model of the inflow rivers to calculate the runoff process of each river, providing source or lateral inflow data for the inflow rivers.

[0146] Further, in step S120, based on the riverbed roughness data and the outflow data of the upstream reservoir of the inflow river in the target hydrological environment, a one-dimensional hydrodynamic model of the inflow river is constructed, and the inflow data and upstream flow data are calculated, including:

[0147] Step S121: Obtain the river system map and topological relationship map of the river flowing into the reservoir, and perform generalization processing on the river channel of the river flowing into the reservoir and the reservoir.

[0148] The steps described above are intended to simplify and generalize the geographical and hydrodynamic characteristics of the rivers flowing into the reservoirs and their associated waterways. Generalization involves simplifying complex river systems into formats that are easier to simulate and analyze, typically including simplifications of river length, width, depth, and flow direction.

[0149] By generalizing, a simplified model of the river channel and reservoir can be obtained, which highlights the main path and key features of the water flow, facilitating subsequent hydrodynamic analysis.

[0150] Simplify complex natural water systems, reduce model complexity, improve computational efficiency, and maintain sufficient accuracy to predict key hydrological dynamics.

[0151] Specifically, GIS tools and hydrodynamic simulation software (such as MIKE 11 or HEC-RAS) can be used to generalize the river model. These tools can help simulators identify key parts of the model and set the corresponding simulation parameters.

[0152] Step S122: Based on the generalized river channels and the reservoir, determine the external and internal boundary conditions of the model; wherein, the external boundary conditions include the upstream water boundary of the river channel and the downstream water level boundary of each river channel; the internal boundary conditions include the setting of hydraulic structures on the river channel.

[0153] As mentioned above, setting correct boundary conditions in the hydrodynamic model is crucial for simulating water flow behavior. External boundary conditions typically involve the river's inlet and outlet, such as upstream inflow and downstream water level; internal boundary conditions may include the placement of hydraulic structures such as dams and sluices. Setting these boundary conditions allows for a more accurate simulation of water flow within the river system, including flow velocity and water level changes. This step ensures that the model accurately reflects real-world water flow behavior, especially in situations involving hydraulic structures.

[0154] Boundary conditions can be set and adjusted using MIKE 11HD or similar hydrodynamic calculation software. These software programs allow for detailed configuration of various physical and technical parameters of the model.

[0155] Step S123: Based on the external boundary conditions and the internal boundary conditions, the MIKE 11HD hydrodynamic calculation model is coupled with the one-dimensional hydrological model of the inflow river to construct the one-dimensional hydrodynamic model of the inflow river, and the upstream flow data and the inflow flow data of the reservoir are calculated.

[0156] The MIKE 11HD model, as described above, was used to further refine the simulation of river hydrodynamic behavior. By coupling it with the previously constructed one-dimensional hydrological model, this model can more comprehensively predict changes in river flow, water level, and related water quality. The coupled model provides detailed predictions of flow, water level, and potential water quality changes, offering crucial data for algal bloom risk prediction. This enhances the model's predictive power and accuracy, enabling it to handle both hydrological and hydrodynamic factors simultaneously, which is particularly important for managing and predicting the impact of reservoirs on the downstream hydrological environment.

[0157] Hydrological data and hydrodynamic parameters can be integrated into the MIKE 11HD software to perform detailed hydrodynamic simulations and analyses.

[0158] Furthermore, the calculation expression of the MIKE 11HD hydrodynamic calculation model is as follows:

[0159]

[0160] Where x and t represent the spatial and temporal coordinates of the calculation point, respectively; A represents the cross-sectional area of ​​the water passage; Q represents the flow rate; h represents the water level; q represents the lateral inflow flow rate; C represents the Chezy coefficient; R represents the hydraulic radius; α represents the momentum correction coefficient; and g represents the gravitational acceleration.

[0161] It should be noted that the MIKE11 HD hydrodynamic calculation model used in this embodiment is based on the vertical integral of the mass and momentum conservation equations, namely the one-dimensional unsteady flow Saint-Venant equations, to simulate the flow state of a river or estuary.

[0162] It should be noted that the one-dimensional hydrodynamic model of the river channel was constructed using MIKE 11HD, and coupled with the NAM hydrological model of the inflow river for joint calibration. This provides inflow flow data for the construction of the two-dimensional hydrodynamic model of the reservoir, and upstream boundary flow data for the three-dimensional hydrodynamic model of the backwater area of ​​the inflow river.

[0163] For example, firstly, river system maps and their topological relationships are collected to generalize the rivers and reservoirs; then, internal and external boundaries are set, with external boundary conditions including upstream inflow boundaries and downstream water level boundaries of each inflow river; internal boundary conditions include the setting of hydraulic structures on the river; finally, measured flow data from important hydrological stations in the reservoir basin and reservoir inflow are selected as calibration targets, and the NAM model is coupled to carry out model calibration.

[0164] Further, step S130, based on the inflow data and the measured water level data of the reservoir, constructs a two-dimensional hydrodynamic model of the reservoir and calculates the downstream water level data of the inflow river, including:

[0165] Step S131: Using the reservoir topographic map, perform two-dimensional grid processing on the reservoir;

[0166] In this embodiment, based on the simulation results of the one-dimensional hydrological and hydrodynamic model of the inflow river, a two-dimensional hydrodynamic model of the reservoir is constructed to provide downstream boundary water level data for the three-dimensional hydrodynamic model of the backwater area of ​​the inflow river.

[0167] The steps described above involve creating a two-dimensional meshed model using a topographic map of the reservoir. Two-dimensional meshing is a standard technique for simulating water flow and related dynamic processes; it transforms complex aquatic topography into a more computationally efficient mesh structure. This generates a detailed mesh model that accurately describes the reservoir's physical morphology and hydrodynamic behavior in two-dimensional space.

[0168] Two-dimensional meshing enables the model to capture details such as flow velocity, flow direction, and water quality changes within the water body, improving the model's spatial resolution and prediction accuracy.

[0169] Topographic maps can be gridded using GIS software and hydrodynamic simulation software (such as MIKE 21 or Hec-RAS2D), and an appropriate grid size can be set to balance computational efficiency and accuracy.

[0170] Step S132: Based on the reservoir topographic map processed by two-dimensional gridding, determine the land boundary conditions and water boundary conditions of the model; wherein, the land boundary conditions are obtained by the inflow data of the inflow river obtained by the one-dimensional hydrodynamic model of the inflow river; the water boundary conditions are determined by the inflow process of the reservoir at the beginning of the century, the measured water level and the reservoir capacity curve.

[0171] As mentioned above, setting appropriate boundary conditions is crucial for constructing an accurate hydrodynamic model. Land boundary conditions involve the boundary where the water body contacts the land, such as the shape and characteristics of the shoreline; water boundary conditions deal with the flow characteristics inside the water body, such as inlet flow rate and outlet water level.

[0172] This step ensures that the model boundaries reflect actual conditions, providing accurate inputs and outputs for flow simulation. Appropriate boundary condition settings ensure the reliability and accuracy of the model output, which has direct practical significance for water quality and algal bloom management. Boundary conditions can be set by integrating flow data from the inflowing river and water level data from the reservoir, and then applied using hydrodynamic simulation software.

[0173] Step S133: Construct a two-dimensional hydrodynamic model of the reservoir, select feature points in the reservoir, and use the measured water level data to calibrate the constructed two-dimensional hydrodynamic model of the reservoir to obtain the downstream water level data of the inflow river.

[0174] As described above, after completing the gridding process and setting boundary conditions, a two-dimensional hydrodynamic model of the reservoir is constructed, and the model is calibrated using measured water level data. Calibration is the process of adjusting model parameters to match measured data, ensuring that the model can accurately predict future hydrological conditions. Through the calibration process, the model parameters are adjusted to ensure that its output matches the actual observations, thereby improving the model's predictive reliability. The calibrated model can more accurately reflect the actual hydrodynamic conditions, providing accurate basic data for algal bloom prediction, especially under changing climate and water conditions.

[0175] Hydrodynamic simulation software (such as MIKE 21HD or other similar tools) can be used to calibrate the model by inputting measured water levels and other hydrological data. Optimization algorithms are then used to adjust the model parameters to minimize the difference between the model output and the observed data.

[0176] Furthermore, step S140, using the upstream flow data and the measured water level data of the reservoir, constructs a three-dimensional hydrodynamic and temperature model of the backwater zone, including:

[0177] Step S141: Perform three-dimensional meshing on the backwater area of ​​the target hydrological environment; wherein, the three-dimensional meshing includes horizontal triangular and quadrilateral meshes, as well as vertical meshes; the middle channel of the river channel of the inflowing river adopts the quadrilateral mesh; the area outside the middle channel of the river channel adopts the triangular mesh; the vertical mesh adopts a hybrid layered form of sigma and Z-level;

[0178] The above steps are the starting point for constructing a three-dimensional hydrodynamic model. We can first construct a three-dimensional hydrodynamic model of the backwater area, and then, based on this model, construct a water temperature model, i.e., a three-dimensional hydrodynamic-temperature model of the backwater area. In this step, we first need to convert the physical space of the backwater area into a three-dimensional mesh. This meshing process allows the model to depict the water flow and temperature distribution in detail in space. This results in a three-dimensional spatial model containing horizontal triangular and quadrilateral meshes, as well as vertical meshes (Z-Level and sigma layers).

[0179] Three-dimensional meshing provides a high-resolution characterization of complex hydrological environments, enabling the capture of detailed changes in water dynamics, such as flow velocity and water temperature gradient.

[0180] Meshing can be performed using 3D hydrodynamic simulation software (such as MIKE 3D or DELFT3D). These software programs can automatically or manually create suitable mesh structures based on physical terrain and hydrological conditions.

[0181] Step S142: Based on the three-dimensional meshed backwater area, construct a three-dimensional hydrodynamic model of the backwater area;

[0182] Step S143: Determine the upstream and downstream boundary conditions of the backwater area; wherein, the upstream boundary conditions are obtained by using the upstream flow data calculated by the one-dimensional hydrodynamic model of the inflow river as input conditions; and the downstream boundary conditions are obtained by using the boundary between the backwater area and the reservoir.

[0183] To ensure accurate model operation, precise upstream and downstream boundary conditions need to be set. These conditions are based on actual hydrological data to simulate the behavior of water flow entering and leaving the simulation area. It is crucial to ensure that the flow data in the model during the simulation remains consistent with actual conditions, especially flow rate and water level.

[0184] Setting the correct boundary conditions is crucial to ensuring the accuracy of simulation results, especially when simulating large-scale hydrological dynamics. Upstream boundary conditions can be set using flow data obtained from a one-dimensional hydrodynamic model of the inflow river, while downstream boundary conditions can be set using the point of intersection with the reservoir. This can be implemented using hydrodynamic simulation software.

[0185] Step S144: Using the three-dimensional hydrodynamic model of the backwater area, and based on the upstream and downstream boundary conditions, calculate the water depth and horizontal flow velocity of the backwater area.

[0186] Based on the constructed 3D mesh and defined boundary conditions, the model calculates the water depth and horizontal flow velocity in the backwater zone, obtaining detailed water depth and velocity data. This data has a direct impact on understanding and predicting algal bloom formation. This step provides comprehensive information on hydrodynamics, helping to identify areas and conditions that may lead to algal blooms.

[0187] These calculations can be performed using three-dimensional hydrodynamic simulation software, which runs dynamic equations based on actual physical and hydrological data to produce results.

[0188] Step S145: Based on the three-dimensional hydrodynamic model of the backwater area and the water temperature model, construct the three-dimensional hydrodynamic and water temperature model of the backwater area.

[0189] The final step, as described above, is to integrate water temperature factors into the three-dimensional hydrodynamic model, creating a comprehensive hydrodynamic-temperature model that can simultaneously simulate water flow and temperature changes. This generates a comprehensive model capable of simulating water flow and temperature dynamics, providing necessary environmental data for subsequent algal bloom risk prediction. Water temperature is one of the key factors influencing algal bloom occurrence; this model can accurately predict water temperature changes, helping to improve the accuracy of algal bloom predictions. Specifically, water temperature change equations and related thermodynamic parameters can be added to the three-dimensional hydrodynamic model, and simulation software such as MIKE 3D or other similar tools can be used to perform these calculations.

[0190] For example, in this embodiment, the three-dimensional hydrodynamic and water quality coupled model of the backwater area is built using MIKE 3FM. First, the three-dimensional hydrodynamic and water temperature coupled model of the backwater area is established using the MIKE 3HD module. Based on this, the three-dimensional ECOLab module is coupled to establish a three-dimensional eutrophication model. Using the three-dimensional hydrodynamic and water temperature coupled model of the backwater area, the upstream flow data, downstream water level data, and meteorological station, hydrological station, and water quality monitoring data of the backwater area are processed to provide spatiotemporal distribution data input for the construction of the eutrophication model of the backwater area. The water quality and chlorophyll a concentration of the backwater area are output through model simulation calculation.

[0191] The MIKE 3FM model is built under the assumptions of shallow water and Boussinesq to solve the Navier-Stokes equations for incompressible fluids. The horizontal mesh in the model uses an unstructured mesh, while the vertical mesh uses coordinates.

[0192] For the construction of the three-dimensional hydrodynamic model of the backwater area, firstly, a three-dimensional mesh of the backwater area is set up, using triangles and quadrilaterals for meshing. The channel part of the middle part of the river uses quadrilateral meshes, while other areas use triangular meshes to meet the needs of the mesh to depict the actual terrain. The vertical mesh adopts a hybrid layered form of sigma and Z-level.

[0193] Then, key parameters such as roughness type and horizontal eddy viscosity coefficient were set, and wind speed and direction data from meteorological stations were used as the wind field conditions for the model.

[0194] Finally, the upstream river inflow is used as the upstream open boundary of the backwater area, and the confluence of the backwater area and the Danjiangkou Reservoir area is used as the downstream open boundary. The inflow flow conditions calculated by the one-dimensional hydrodynamic model of the inflow river are used as the input conditions for the upstream open boundary, and the measured water level of the reservoir is used as the input conditions for the downstream open boundary. Data such as water depth and horizontal flow velocity in the backwater area are calculated by the model.

[0195] After constructing the three-dimensional hydrodynamic model of the backwater area, a three-dimensional hydrodynamic and water temperature model of the backwater area was constructed based on the three-dimensional hydrodynamic model of the backwater area and the water temperature model.

[0196] Furthermore, in the three-dimensional hydrodynamic and temperature model of the backwater zone, the transport equation of the temperature model is: in, The turbulent diffusion coefficient represents the vertical direction; F represents the source term generated by heat exchange with the atmosphere; t The horizontal diffusion term F represents the horizontal diffusion term; T represents temperature; u, v, and w represent the velocity components of the water flow in the x, y, and z directions, respectively; the horizontal diffusion term F t The equation is:

[0197] in, Represents the turbulent diffusion coefficient in the horizontal direction;

[0198] It should be noted that the turbulent diffusion coefficients in the horizontal and vertical directions are constant values ​​or determined proportionally by the eddy viscosity coefficient.

[0199] In this embodiment, a water temperature model is introduced based on a three-dimensional hydrodynamic model. A complex water temperature transport equation is established to simulate the spatial distribution and temporal variation of water temperature. This equation includes factors such as convection, diffusion, and heat exchange between the water body and the atmosphere, comprehensively simulating the dynamic changes in water temperature.

[0200] The constructed model provides a detailed description of the temperature distribution in a water body in three-dimensional space and its changes over time, taking into account various thermodynamic processes. This model allows for accurate prediction of ecological and chemical changes caused by water temperature variations, particularly the role of water temperature in algal blooms. It is crucial for managing water health and predicting algal bloom risks.

[0201] Regarding boundary conditions:

[0202] (1) The boundary condition for surface temperature is: when Z = η, in, η represents the turbulent diffusion coefficient in the vertical direction; z represents the vertical position coordinate; η represents the water surface position. Q represents the temperature gradient in the vertical direction; n ρ0 represents the net heat flux, a measure of heat exchange between water and the atmosphere; ρ0 represents the density of water (typically around 1000 kg / m2). 3 );c p This represents the specific heat capacity of water [given as 4217 J / (kg·K)].

[0203] The bottom boundary condition is: when z = -d,

[0204] The heat exchange between water and air surfaces is calculated based on the following physical processes:

[0205] Q n =q v +q c +βq sr,net +q lr,net ;(Formula 7) Among them, Q n Represents the net surface heat flux (the sum of all heat flux components); c p =4217 J / (kg°K) represents the specific heat of water; q v Represents latent heat flux (latent heat flux, related to the evaporation or condensation of water); q c Represents sensible heat flux (related to direct heat conduction between water and the atmosphere); q sr,net Represents net shortwave radiation (net solar radiation input); q lr,net Represents net longwave radiation (the exchange of infrared radiation between water and the atmosphere); This represents the amount of heat received or lost per unit mass of water (used for calculating temperature changes).

[0206] These formulas collectively form the basis of the thermodynamic boundary conditions in water temperature models, accurately describing how water exchanges heat with its surrounding environment through various physical processes. These models allow for precise prediction of water temperature changes under different environmental conditions, which is crucial for ecological models, water quality management, and environmental monitoring.

[0207] The above-mentioned model can be constructed by using the measured air temperature and relative humidity of the meteorological station as input conditions for atmospheric temperature and relative humidity, respectively; and using the measured water temperature data of the water quality monitoring station as input conditions for the upstream and downstream water temperature boundaries. The water temperature of different layers can be calculated by using the three-dimensional hydrodynamic water temperature model of the backwater area.

[0208] Example 3:

[0209] refer to Figure 3 Based on the above embodiment 1, embodiment 3 of this application provides a method for predicting algal blooms in backwater areas. Step S200 involves constructing a three-dimensional eutrophication model of the backwater area based on the water temperature, water depth, horizontal flow velocity, and wind speed obtained from the three-dimensional hydrodynamic and temperature model of the backwater area. This includes:

[0210] Step S210: Construct a three-dimensional eutrophication model of the backwater area based on a preset eutrophication template;

[0211] The aforementioned three-dimensional eutrophication model of the backwater area was built by modifying a predefined eutrophication template in ECO Lab.

[0212] It should be noted that the eutrophication template is a system composed of a series of differential equations. This system consists of differential equations describing the changes in 12 state variables of phytoplankton, chlorophyll a, zooplankton, detritus, dissolved oxygen, and inorganic nutrients in the water body, as shown in Table 4.3. The entire model is calculated based on the cycling and transformation of carbon between different nutrient pools (phytoplankton, zooplankton, detritus, and sediment), and assumes that the amount of carbon is sufficient during the calculation.

[0213] Table 1. List of state variables in the three-dimensional eutrophication model and sediment system template of the backwater zone

[0214] Chinese name Model Abbreviation Description in the model unit Carbon content of phytoplankton PC Phytoplankton C <![CDATA[gC / m 3 ]]> Nitrogen content of phytoplankton PN Phytoplankton N <![CDATA[gN / m 3 ]]> Phosphorus content of phytoplankton PP Phytoplankton P <![CDATA[gPm 3 ]]> Chlorophyll a CH Chlorophyll-a <![CDATA[g Chl / m 3 ]]> Carbon content of zooplankton ZC Zooplankton C <![CDATA[gC / m 3 ]]> Carbon content in debris DC Detritus C <![CDATA[gC / m 3 ]]> nitrogen content in debris DN Detritus N <![CDATA[gN / m 3 ]]> Phosphorus content in debris DP Detritus P <![CDATA[gP / m 3 ]]> ammonia nitrogen <![CDATA[NH3]]> Ammonia N <![CDATA[gN / m 3 ]]> Nitrogen <![CDATA[NO3]]> Nitrate N <![CDATA[gN / m 3 ]]> Inorganic phosphorus IP Inorganic phosphorus <![CDATA[gP / m 3 ]]> Dissolved oxygen DO Dissolved oxygen <![CDATA[g DO / m 3 ]]>

[0215] The life activities of phytoplankton in water mainly consist of the following parts: phytoplankton absorbing nutrients to maintain their growth, phytoplankton consuming nutrients in their bodies for metabolism, phytoplankton photosynthesis to produce oxygen, phytoplankton being preyed upon by zooplankton, and phytoplankton settling and death. In the model, these are represented by the processes of growth, predation, settling, and death.

[0216] Nutrients, light, and temperature all influence phytoplankton growth, and the effects of these environmental factors are reflected in the equations through constraint functions. Phytoplankton are represented by four state variables: phytoplankton carbon, phytoplankton nitrogen, phytoplankton phosphorus, and phytoplankton chlorophyll a. The nutrient pools (carbon, nitrogen, and phosphorus) within the phytoplankton cells in the model are state variables because their uptake motive force is separated from the phytoplankton carbon uptake motive force, resulting in time-varying nitrogen-carbon and phosphorus-carbon ratios. For carbon-based phytoplankton cell nutrient pools, their source and sink terms are proportional to the corresponding carbon content.

[0217] In the steps described above, a eutrophication model is constructed using data (water temperature, water depth, current velocity, and wind speed) obtained from a three-dimensional hydrodynamic and temperature model. This model uses a pre-defined eutrophication template, which defines how to assess the behavior of nutrients (such as nitrogen and phosphorus) and their impact on algal blooms based on hydrological and meteorological data. This creates a model capable of simulating and predicting nutrient dynamics and the potential occurrence of algal blooms in a specific water body.

[0218] Specifically, ecological modeling software or custom numerical simulation programs can be used in conjunction with Geographic Information Systems (GIS) to process spatial data and run these complex ecological models.

[0219] Step S220: Determine the boundary conditions and model forces of the three-dimensional eutrophication model of the backwater area; the boundary conditions include the upstream river boundary conditions and the downstream reservoir confluence boundary conditions.

[0220] The above-mentioned open boundary conditions of the three-dimensional eutrophication model of the backwater area include two boundary conditions: the confluence of the upstream river channel and the downstream reservoir. The water quality conditions of the two boundaries are based on the measured water quality data of the water quality monitoring station. The water quality monitoring indicators include total nitrogen, total phosphorus, ammonia nitrogen, and dissolved oxygen.

[0221] The model's influencing forces include water temperature, light intensity, water depth, horizontal flow velocity, vertical water stratification thickness, and wind speed. Among these, water temperature, water depth, horizontal flow velocity, and wind speed are built-in forces. Water temperature is obtained from a three-dimensional water temperature model of the backwater area; water depth, horizontal flow velocity, and wind speed are read from a three-dimensional hydrodynamic model of the backwater area; and light intensity is obtained from surrounding meteorological stations.

[0222] Set parameters related to algal growth and algal bloom processes; some parameters will be adjusted during model calibration and validation.

[0223] Step S230: Based on the boundary conditions, the water temperature, water depth, horizontal flow velocity, and wind speed obtained from the three-dimensional hydrodynamic and temperature model of the backwater area, as well as the light intensity obtained from the meteorological station, are used as the model forces, and the three-dimensional eutrophication model of the backwater area is calibrated.

[0224] Calibration is the process of adjusting model parameters to ensure that the model output is consistent with actual observation data. This step uses actual observed water quality data to adjust the model to ensure its accuracy in future applications. Data from two water quality monitoring stations upstream and downstream of the backwater area were used to calibrate and validate the established three-dimensional eutrophication model of the backwater area.

[0225] In summary, this embodiment presents a systematic solution for constructing a three-dimensional eutrophication model to assess and manage algal blooms in backwater areas. This method combines detailed hydrological data, ecological models, and field monitoring data to scientifically predict algal bloom occurrences and provide decision support for management.

[0226] Example 4:

[0227] refer to Figure 4 Based on Embodiment 1 above, Embodiment 4 of this application provides a method for predicting algal blooms in backwater areas. Step S300 involves constructing a BP neural network model based on the three-dimensional eutrophication model of the backwater area and training the BP neural network model, including:

[0228] Step S310: Based on downstream water level, upstream flow rate, air temperature, light intensity, upstream water quality, and downstream water quality, construct different scenario schemes, and based on the different scenario schemes, calculate chlorophyll a concentration through the three-dimensional eutrophication model of the backwater area to form a training set.

[0229] As described above, this step involves using a three-dimensional eutrophication model to simulate chlorophyll a concentration based on various environmental parameters (downstream water level, upstream flow rate, air temperature, light intensity, upstream water quality, and downstream water quality). This data reflects potential algal bloom development under various environmental conditions and is used to generate a training set for the neural network model. This results in a dataset containing multiple predictor and response variables (chlorophyll a concentration), which will be used to train the neural network model.

[0230] By simulating chlorophyll a concentration under different environmental conditions, neural network models can learn the complex relationship between environmental factors and algal bloom development, thereby enhancing the model's generalization ability and prediction accuracy.

[0231] Specifically, environmental simulation software or custom numerical simulation scripts can be used to run eutrophication models and output datasets for training. Data preprocessing, such as normalization, is usually required to optimize the training performance of the neural network.

[0232] Step S320: Construct the BP neural network model;

[0233] After obtaining the training dataset, the next step is to build a BP (backpropagation) neural network model. This is a widely used supervised learning network that can learn the complex nonlinear relationship between the input (environmental parameters) and the output (chlorophyll a concentration).

[0234] Step S330: Using the parameters in the training set as model input and the chlorophyll a concentration to be fitted as model output, the BP neural network model is trained to obtain the trained BP neural network model.

[0235] In this step, the dataset generated in step S310 is used to train the constructed BP neural network model. The training process involves adjusting network parameters (such as weights and biases) to minimize the difference between the predicted output and the actual data. The trained neural network model is able to accurately predict the concentration of chlorophyll a based on the input environmental parameters.

[0236] The training process allows the model to adapt to specific data features and patterns, improving the accuracy and reliability of predictions.

[0237] Model parameters can be iteratively adjusted during training by defining a loss function (such as mean squared error) and an optimization algorithm (such as stochastic gradient descent). Training can be performed on high-performance computing resources to handle large amounts of data and complex network structures.

[0238] Furthermore, step S330, training the BP neural network model, further includes:

[0239] Step S331: Principal component analysis is used to reduce the dimensionality of the parameters in the training set to obtain the model input terms;

[0240] It should be noted that, due to the relatively low volatility and limited implicit information in the dataset, and the large number of water quality parameters, using all of them as input could easily lead to the "curse of dimensionality," significantly increasing the computational load on the model. Therefore, in this embodiment, the data is preprocessed. Specifically, Principal Component Analysis (PCA) can be used to reduce the dimensionality of the data before it is used as model input. The principal components are selected based on a cumulative contribution rate higher than 85%.

[0241] As mentioned above, PCA is used to reduce the number of features in the dataset before training the neural network, which helps reduce model complexity and avoid overfitting. In this embodiment, principal components are extracted from the original dataset to cumulatively explain at least 85% of the data variance, and used as input to the neural network. This simplifies the dataset, retaining most of the information, but with lower dimensionality.

[0242] It simplifies the model's input, speeds up training, and may improve the model's generalization ability.

[0243] Specifically, the PCA algorithm can be used in the data preprocessing stage, usually implemented in Python using the PCA class from the scikit-learn library.

[0244] Step S332: Standardize the model input terms to obtain standardized model input terms; wherein, the expression for the standardization process is:

[0245] Where x represents the original model input; xˋ represents the standardized model input; μ represents the data mean; σ represents the data standard deviation;

[0246] As mentioned above, neural network models are sensitive to the features of the input data. To eliminate the order-of-magnitude differences between data of different dimensions, standardization is used to transform the input data into a standard distribution with a mean of 0 and a variance of 1. This ensures that each feature contributes equally to the model, preventing certain features with large dimensions from dominating the learning process. The mean and standard deviation of each feature are calculated, and these statistics are used to transform all data points, ensuring all features have the same scale, potentially leading to faster convergence during model training.

[0247] This step is used to prevent gradient vanishing or exploding and to improve model stability.

[0248] Step S333: Based on the cross-validation strategy, the standardized model input items are divided into k folds, where k-1 folds are used as the training data set, and the 1-fold samples are combined with the NSE, RMSE, and MAE evaluation metrics to form the validation data set.

[0249] The above-mentioned raw data, after PCA dimensionality reduction and standardization, were used to train a BP neural network model for the backwater area. The differences between the simulation results of the two methods were compared through cross-validation. The main evaluation indicators were: NSE, RMSE, MAE, MRE, and MRE_1. At the same time, in order to analyze the fluctuation of the cross-validation results, the standard error of the NSE score was calculated, which is denoted as Std(NSE).

[0250] A k-folders cross-validation strategy is adopted to ensure full utilization of the sample data while separating the training and validation sets. The entire training set is divided into k folds. Training is performed on the training set consisting of k-1 folds, and the model training is evaluated on the validation set consisting of 1 folds using NSE, RMSE, and MAE metrics. This process is repeated k times, and the average of the k iterations is taken as the final model score. This score is used as the basis for evaluating parameter tuning results.

[0251] In this step, by dividing the data into multiple parts and alternately using one part as the validation set and the rest as the training set, the generalization ability of the model can be better evaluated. Specifically, the data is divided into k subsets, with one subset set as validation data each time and the rest used as training data. This allows us to obtain the model's performance on various different data subsets, reducing the model's dependence on a specific data distribution.

[0252] Building a validation dataset reduces the risk of model overfitting and improves model reliability.

[0253] k-fold cross-validation can be used during model training, typically using the cross_val_score function from the scikit-learn library.

[0254] Step S334, execute the following training process k times: train the BP neural network model based on the training data set, and verify it through the verification data set to obtain the verification result;

[0255] As described above, by repeatedly training and validating, the model's performance on different training and validation datasets can be evaluated, helping to identify the model's stability and reliability. In each iteration, different training and validation datasets are selected for training and testing the model. This ensures that every data point has the opportunity to serve as validation data, thus enabling a more comprehensive evaluation of the model.

[0256] After each training iteration, performance metrics (such as accuracy and loss) are obtained by evaluating the model on a validation set. These metrics can be used to measure the model's generalization ability. This provides a comprehensive view of the model's performance and helps avoid overfitting to a single dataset.

[0257] Step S335: Take the verification results of k times as the training score of the BP neural network model, and adjust the BP neural network model based on the training score.

[0258] As described above, collecting the results of multiple training and validation runs yields the model's average performance across different subsets. This helps evaluate the model's overall performance and allows for parameter tuning based on this information. Summarizing the results of k validation runs, the average score is typically used to evaluate the model's overall performance. Based on the average score from the k runs, the quality of the model parameters can be more accurately determined, facilitating model optimization.

[0259] This step makes model adjustments more scientific and systematic, based on objective data rather than the results of a single experiment.

[0260] During the model training process, the network structure or learning parameters can be adjusted based on the average value of evaluation metrics (such as accuracy, recall, F1 score, etc.), and these parameters can be optimized using the backpropagation algorithm.

[0261] The method described in this embodiment ensures that the BP neural network not only learns the inherent patterns in the data but also possesses good generalization ability and excellent predictive performance for unknown data. In practice, these steps typically require the use of specialized machine learning and statistical software libraries to ensure computational efficiency and the accuracy of the results.

[0262] Furthermore, step S320, constructing the BP neural network model, further includes:

[0263] Step S3211: Set the number of hidden layers in the BP neural network model to a single hidden layer; wherein the number of neurons in the single hidden layer is set to 1 / 3 to 2 times that of the input layer neurons;

[0264] It should be noted that the structure of the hidden layer in a neural network model includes the number of hidden layers and the number of neurons in a single layer. Generally speaking, the more hidden neurons and hidden layers there are, the stronger its ability to mine and extract data features. However, the existence of the curse of dimensionality also exponentially increases the time and storage space required for model computation, and the complex hidden layer structure may lead to overfitting, affecting the model's generalization ability.

[0265] In theory, a single-hidden-layer neural network using a non-linear activation function is sufficient to extract the non-linear mapping relationship between Chl-a and the various environmental variables of the input items.

[0266] In this embodiment, the applicability of a single hidden layer neural network to a two-point model will be examined first. Based on empirical methods and considering the computational load of the model, the range of neurons in the single hidden layer is set between 1 / 3 and 2 times that of the neurons in the input layer (i.e., 2-12), and the model performance is verified and evaluated.

[0267] In this step, a single-hidden-layer neural network is first set up to simplify the model's structure, which facilitates rapid implementation and initial observation of the model's performance. The neural network structure contains only one hidden layer, and the number of neurons can be adjusted according to the size of the input layer, thus creating a single-hidden-layer BP neural network model. Simplifying the model reduces computational resource consumption and helps avoid overfitting.

[0268] Step S3212: Use the NSE index to evaluate and obtain the evaluation results;

[0269] As mentioned above, since NSE is more sensitive to parameter settings in practice, it is used as the main evaluation index. Furthermore, considering the volatility of the model's fit, a polynomial regression curve is fitted to the NSE score to represent its main trends.

[0270] Step S3213: If the evaluation result is that the NSE value reaches the preset threshold, then the performance is determined to have met expectations.

[0271] Step S3214: If the evaluation result is that the NSE value does not reach the preset threshold, it is determined that the performance has not met expectations, and the number of hidden layers is set to double hidden layers; wherein, in the double hidden layers, the number of neurons in the first layer is set to 1 / 3 to 2 times that of the input layer neurons; and the number of neurons in the second layer is set to 2 / 3 times that of the input layer neurons.

[0272] As mentioned above, if even reaching the upper limit of the number of hidden neurons is insufficient to explain the correlation between variables, then constructing a two-hidden-layer neural network should be considered. Two-hidden-layer neural networks have stronger feature extraction capabilities than single-hidden-layer networks, while effectively reducing the required number of hidden neurons.

[0273] Similarly, based on empirical methods, the number of neurons in the first hidden layer is set to a range of 2-12, and the number of neurons in the second hidden layer is set to 2 / 3 of the number in the first layer (rounded up), making it slightly smaller than the first layer to improve the ability to summarize features. Finally, the optimization effect of the dual hidden layer neurons is evaluated by calculating the model's NSE score.

[0274] In this embodiment, the above method embodies an iterative and adaptive approach to optimize the structure and parameters of the neural network model, ensuring that the model is neither too simple nor too complex, thereby achieving optimal performance within a given application context. This approach ensures that the model possesses good generalization ability and prediction accuracy for specific problems.

[0275] Furthermore, step S320, constructing the BP neural network model, further includes:

[0276] Step S3221: Based on the Xavier method, initialize the weights of the BP neural network model; wherein the expression for weight initialization is: Where ω represents the weight matrix; U represents a uniform distribution; n in n represents the number of neurons in the input layer of the neural network. out μ represents the number of neurons in the output layer of the neural network; μ is determined by the type of activation function selected; the activation function is any one of Tanh, Sigmoid, and Softsign; if the activation function is Tanh, then μ = 6; if the activation function is Sigmoid, then μ = 96; if the activation function is ReLU, then μ = 12.

[0277] It should be noted that, for neural network models, using a random initialization strategy for the weights of each neuron helps to break the symmetry between neuron nodes and improve the feature extraction capability of the neural network.

[0278] Therefore, in constructing the pre-defined model, this embodiment employs the Xavier method for parameter initialization. Proper weight initialization helps prevent gradient vanishing or exploding problems in the early stages of training, promoting faster and more efficient network convergence. Using the Xavier initialization method (such as the weight initialization expression), the initial weight size is automatically adjusted based on the number of neurons in the input and output layers. This results in a more uniform activation and gradient distribution in the early stages of neural network training, contributing to improved training efficiency and stability.

[0279] Step S3222: Train the BP neural network model using different activation functions, and evaluate it using performance evaluation metrics to obtain the evaluation results for each activation function;

[0280] The activation function Tanh is: The activation function Sigmoid is: The activation function Softsign is: (Formula 12) where x represents the sum of the input signals, including the weighted sum and bias term of the previous layer;

[0281] Step S3223: Based on the evaluation results, determine the activation function used in the BP neural network model.

[0282] The Xavier method avoids the problem of gradient explosion or gradient vanishing leading to model training failure by setting the initial weights of each neuron to a reasonable standard deviation or boundary under a normal or uniform distribution, in conjunction with the model size.

[0283] The gradient is a function of the derivative of the activation function; therefore, the Xavier method is sensitive to the activation function. Its assumptions make it more suitable for activation functions symmetric about 0, such as Tanh, Sigmoid, and Softsign. For models with a pre-defined ReLU, its suitability for activation functions is generally limited. Softsign has similar functional characteristics to Tanh, but its reciprocal calculation takes longer.

[0284] The activation function ReLU is: f(x) = max(0,x). (Equation 13)

[0285] In the above embodiment, ReLU, Sigmoid, and Tanh are used as activation functions to train the neural network model, and the model scores are compared and analyzed. The evaluation index values ​​include: NSE, Std, RMSE, MAE, MRE, and MRE_1.

[0286] Furthermore, the construction of the BP neural network model also includes:

[0287] Step S3231: Based on the chlorophyll a concentration, establish a time series model; wherein, the expression of the time series model is: Where xt represents the actual observed value at time point t in the time series; Represents the autoregressive coefficient; ε t ε represents the error term at time point t; t-1 ε t-2 ε t-q This represents the first q error terms in the time series;

[0288] It should be noted that the study found that data records obtained from long-term continuous observation of Chl-a concentration (chlorophyll a concentration) in water bodies exhibit significant time-series characteristics, making them suitable for fitting and prediction using time-series models. That is, there is a correlation between current Chl-a fluctuations and past Chl-a data; this hidden data dependency helps improve the model's fitting and predictive performance. Furthermore, in scenarios requiring short-term Chl-a prediction, past Chl-a monitoring data is often known. Therefore, incorporating Chl-a observations that lag behind the model's output as input data into the model allows for the examination of model performance changes and further optimization.

[0289] The above time series model: Autoregressive Moving Average (ARMA) model, its formula can contain two parts: (1) The left side of the equation is composed of the independent variable x t and its lag time series x t-p The autoregressive (AR) part is composed of; (2) the right side of the equation consists of the error term, also known as the disturbance term ε. t and its lag time series ε t-q The moving average (MA) component.

[0290] In other words, when applied to chlorophyll a simulation, the AR component shows the correlation between Chl-a and its own past concentration levels, while the MA component shows the correlation between Chl-a and various external environmental disturbances, namely the various environmental variable inputs in the aforementioned modeling and their past data.

[0291] Therefore, in order to determine the time series characteristics of chlorophyll a within the range of two feature points t1 and h2 in the Shending River Basin, this embodiment uses the idea of ​​Box-Jenkins ARMA time series modeling to perform pattern recognition on the chlorophyll a time series.

[0292] Step S3232: Using the lag time as an input variable, calculate the chlorophyll a concentration data under the lag time based on the time series model, and use it as the time training set;

[0293] Step S3233: Based on the time training set, train the BP neural network model, evaluate it using performance evaluation metrics, obtain evaluation results corresponding to different lag times, and adjust the BP neural network model based on the evaluation results corresponding to different lag times.

[0294] The above calculations yielded the autocorrelation function (ACF) and partial autocorrelation function (PACF) of Chl-a relative to its lagged data, and the mean values ​​of the scenario schemes were used for analysis.

[0295] We selected 1-day time-lag data as input to further train the model and analyzed its score. Secondly, considering that the results obtained from pattern recognition using autocorrelation and partial correlation coefficient plots in time series analysis are often approximate and require further verification, we added 4-day and 7-day time-lag data to the model and calculated and analyzed its score.

[0296] Furthermore, the algal bloom risk prediction results include:

[0297] If the chlorophyll a concentration Ccompound in the backwater area of ​​the target hydrological environment is ≤10, then the predicted risk of algal bloom is: Algal bloom severity level 1; if the chlorophyll a concentration Ccompound in the backwater area of ​​the target hydrological environment is 10 < C ≤ 15, then the predicted risk of algal bloom is: Algal bloom severity level 2; if the chlorophyll a concentration Ccompound in the backwater area of ​​the target hydrological environment is 15 < C ≤ 50, then the predicted risk of algal bloom is: Algal bloom severity level 3; if the chlorophyll a concentration Ccompound in the backwater area of ​​the target hydrological environment is 50 < C ≤ 100, then the predicted risk of algal bloom is: Algal bloom severity level 4; if the chlorophyll a concentration Ccompound in the backwater area of ​​the target hydrological environment is C > 100, then the predicted risk of algal bloom is: Algal bloom severity level 5.

[0298] In this embodiment, the BP neural network model for predicting chlorophyll a in the backwater area, trained based on mechanistic model data, achieves a Nash efficiency coefficient of 0.972 for chlorophyll a prediction results three days in advance. Based on the predicted chlorophyll a concentration in the backwater area, the degree of algal bloom in the backwater area can be evaluated with reference to the table below, providing a basis for subsequent algal bloom control measures.

[0299] Table 2. Classification of Algal Bloom Severity Based on Chlorophyll a Prediction Results

[0300] Algal bloom severity level Chlorophyll a concentration C (μg / L) Algal bloom severity level Chlorophyll a concentration C (μg / L) Level 1 C＜10 Level 4 50＜C＜100 Level 2 10＜C＜15 Level 5 C＞100 Level 3 15＜C＜50 / /

[0301] It is understood that the device in this embodiment corresponds to the algal bloom prediction method in the backwater area of ​​the above embodiment, and the options in the above embodiment are also applicable to this embodiment, so they will not be described again here.

[0302] In addition, refer to Figure 5This application also provides a backwater area algal bloom prediction device, comprising: an environmental data module 10, used to acquire geographic hydrological data and environmental monitoring data of the target hydrological environment, and construct a three-dimensional hydrodynamic and temperature model of the backwater area based on the geographic hydrological data and the environmental monitoring data; a eutrophication module 20, further used to construct a three-dimensional eutrophication model of the backwater area based on the water temperature, water depth, horizontal flow velocity and wind speed obtained from the three-dimensional hydrodynamic and temperature model of the backwater area; a neural network module 30, further used to construct a BP neural network model based on the three-dimensional eutrophication model of the backwater area, and train the BP neural network model; and a risk prediction module 40, used to obtain the algal bloom risk prediction result of the backwater area of ​​the target hydrological environment using the trained neural network model.

[0303] This application also provides a computer-readable storage medium for storing the computer program used in the aforementioned computer device. For example, the computer-readable storage medium may include, but is not limited to, various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0304] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that, in alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0305] In addition, the functional modules or units in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0306] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a smartphone, personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0307] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for predicting algal blooms in backwater areas, characterized in that, include: Acquiring geographic hydrological data and environmental monitoring data of the target hydrological environment, and constructing a three-dimensional hydrodynamic and temperature model of the backwater area based on the geographic hydrological data and the environmental monitoring data, includes: collecting the geographic hydrological data and environmental monitoring data of the target hydrological environment, constructing a one-dimensional hydrological model of the inflow river, and calculating the riverbed roughness data in the target hydrological environment; based on the riverbed roughness data and the outflow data of the upstream reservoir of the inflow river in the target hydrological environment, constructing a one-dimensional hydrodynamic model of the inflow river, and calculating the inflow data and upstream flow data; based on the inflow data and the measured water level data of the reservoir, constructing a two-dimensional hydrodynamic model of the reservoir, and calculating the downstream water level data of the inflow river; and using the upstream flow data and the measured water level data of the reservoir, constructing a three-dimensional hydrodynamic and temperature model of the backwater area. Based on the water temperature, water depth, horizontal flow velocity, and wind speed obtained from the three-dimensional hydrodynamic and temperature model of the backwater area, a three-dimensional eutrophication model of the backwater area is constructed, including: constructing the three-dimensional eutrophication model of the backwater area based on a preset eutrophication template; determining the boundary conditions and model forces of the three-dimensional eutrophication model of the backwater area; the boundary conditions include the upstream river boundary conditions and the downstream reservoir confluence boundary conditions; and calibrating the three-dimensional eutrophication model of the backwater area based on the boundary conditions, using the water temperature, water depth, horizontal flow velocity, and wind speed obtained from the three-dimensional hydrodynamic and temperature model of the backwater area, as well as the light intensity obtained from the meteorological station, as model forces. Based on the three-dimensional eutrophication model of the backwater area, a BP neural network model is constructed and trained, including: using the three-dimensional eutrophication model to simulate chlorophyll a concentration according to different environmental parameters, generating a dataset containing environmental parameters and chlorophyll a concentration; and using the dataset to train the BP neural network model; wherein the input of the BP neural network model is the environmental parameters; and the output is the chlorophyll a concentration. The risk prediction results of algal blooms in the backwater area of ​​the target hydrological environment are obtained by using a trained neural network model.

2. The method for predicting algal blooms in backwater areas as described in claim 1, characterized in that, The process of collecting geographic hydrological data and environmental monitoring data of the target hydrological environment, constructing a one-dimensional hydrological model of the inflow river, and calculating the riverbed roughness data of the target hydrological environment includes: The geographic hydrological data and environmental monitoring data of the target hydrological environment are collected; wherein, the geographic hydrological data includes the surface type of the rivers flowing into the reservoir, the reservoir topographic map, and land use remote sensing data; the environmental monitoring data includes rainfall data and evaporation data; Based on the aforementioned geographical and hydrological data and environmental monitoring data, the inflowing river is divided into multiple catchment areas; A one-dimensional hydrological model of the inflow river based on the NAM hydrological model is constructed, and the riverbed roughness data is obtained by calibrating each catchment area of ​​the inflow river using the one-dimensional hydrological model of the inflow river.

3. The method for predicting algal blooms in backwater areas as described in claim 1, characterized in that, Based on the riverbed roughness data and the outflow data of the upstream reservoir of the inflow river in the target hydrological environment, a one-dimensional hydrodynamic model of the inflow river is constructed, and the inflow data and upstream flow data are calculated, including: Obtain the river system map and topological relationship map of the rivers flowing into the reservoir, and perform generalization processing on the river channels of the rivers flowing into the reservoir and the reservoir. Based on the generalized river channels and reservoirs, the external and internal boundary conditions of the model are determined; wherein, the external boundary conditions include the upstream water boundary of the river channel and the downstream water level boundary of each river channel; the internal boundary conditions include the setting of hydraulic structures on the river channel. Based on the external and internal boundary conditions, the MIKE 11 HD hydrodynamic calculation model is coupled with the one-dimensional hydrological model of the inflow river to construct the one-dimensional hydrodynamic model of the inflow river, and the upstream flow data and the inflow flow data of the reservoir are calculated. The calculation expression of the MIKE 11 HD hydrodynamic calculation model is as follows: ； ； Where x and t represent the spatial and temporal coordinates of the calculation point, respectively; A represents the cross-sectional area of ​​the water passage; Q represents the flow rate; h represents the water level; q represents the lateral inflow rate; C represents the Chezy coefficient; R represents the hydraulic radius; α represents the momentum correction coefficient; and g represents the gravitational acceleration.

4. The method for predicting algal blooms in backwater areas as described in claim 1, characterized in that, The process of constructing a two-dimensional hydrodynamic model of the reservoir based on the inflow data and the measured water level data of the reservoir, and calculating the downstream water level data of the inflow river, includes: The reservoir is then processed into a two-dimensional grid using the reservoir topographic map. Based on the reservoir topographic map processed by two-dimensional gridding, the land boundary conditions and water boundary conditions of the model are determined; wherein, the land boundary conditions are obtained by the inflow data of the inflow river based on the one-dimensional hydrodynamic model of the inflow river; the water boundary conditions are determined by the inflow process of the reservoir at the beginning of the century, the measured water level and reservoir capacity curves. A two-dimensional hydrodynamic model of the reservoir is constructed. Feature points are selected in the reservoir, and the measured water level data is used to calibrate the constructed two-dimensional hydrodynamic model of the reservoir to obtain the downstream water level data of the inflow river.

5. The method for predicting algal blooms in backwater areas as described in claim 1, characterized in that, The process of constructing a three-dimensional hydrodynamic and temperature model of the backwater zone using the upstream flow data and the measured water level data of the reservoir includes: The backwater area of ​​the target hydrological environment is subjected to three-dimensional meshing; wherein, the three-dimensional meshing includes horizontal triangular and quadrilateral meshes, as well as vertical meshes; the middle channel of the river channel of the inflowing river adopts the quadrilateral mesh; the area outside the middle channel of the river channel adopts the triangular mesh; the vertical mesh adopts a hybrid layered form of sigma and Z-level; Based on the three-dimensional meshing of the backwater zone, a three-dimensional hydrodynamic model of the backwater zone is constructed. The upstream and downstream boundary conditions of the backwater area are determined; wherein, the upstream boundary conditions are obtained by using the upstream flow data calculated by the one-dimensional hydrodynamic model of the inflow river as input conditions; and the boundary between the backwater area and the reservoir is used as the downstream boundary conditions. Using the three-dimensional hydrodynamic model of the backwater area, and based on the upstream and downstream boundary conditions, the water depth and horizontal velocity of the backwater area are calculated. Based on the three-dimensional hydrodynamic model of the backwater area combined with the water temperature model, a three-dimensional hydrodynamic and water temperature model of the backwater area is constructed. In the three-dimensional hydrodynamic and temperature model of the backwater zone, the transport equation of the temperature model is: ； in, The turbulent diffusion coefficient represents the vertical direction; Represents the source term generated by heat exchange with the atmosphere; Represents the horizontal diffusion term; T represents temperature; u, v, and w represent the velocity components of the water flow in the x, y, and z directions, respectively; The horizontal diffusion term The equation is: ； in, Represents the turbulent diffusion coefficient in the horizontal direction; The boundary condition for surface temperature is as follows: When 𝑧=𝜂 ; in, η represents the turbulent diffusion coefficient in the vertical direction; z represents the vertical position coordinate; η represents the water surface position. Q represents the temperature gradient in the vertical direction; n ρ0 represents the net heat flux, a measure of heat exchange between water and the atmosphere; ρ0 represents the density of water; c p Represents the specific heat capacity of water; The bottom boundary conditions are: When 𝑧=-d ; The heat exchange between water and air surfaces is calculated based on the following physical processes: ； ； Among them, Q n Represents the net heat flux at the surface; c p q represents the specific heat capacity of water; v Represents latent heat flux; q c q represents sensible heat flux; sr,net Represents net shortwave radiation; q lr,net Represents net longwave radiation; Heat received or lost per unit mass of water.

6. The method for predicting algal blooms in backwater areas as described in claim 1, characterized in that, The process of constructing a BP neural network model based on the three-dimensional eutrophication model of the backwater area and training the BP neural network model includes: Based on downstream water level, upstream flow rate, air temperature, light intensity, upstream water quality, and downstream water quality, different scenario schemes are constructed. Based on the different scenario schemes, chlorophyll a concentration is calculated through the three-dimensional eutrophication model of the backwater area to form a training set. Construct a BP neural network model; The BP neural network model is trained by using the parameters in the training set as the model input and the chlorophyll a concentration to be fitted as the model output, so as to obtain the trained BP neural network model. The training of the BP neural network model further includes: Principal component analysis is used to reduce the dimensionality of the parameters in the training set to obtain the model input terms; The model input terms are standardized to obtain standardized model input terms; wherein the expression for the standardization process is: ； Where x represents the original model input; x` represents the standardized model input; μ represents the data mean; σ represents the data standard deviation; Based on the cross-validation strategy, the standardized model input is divided into k folds, where k-1 folds are used as the training data set, and the 1-fold samples are combined with NSE, RMSE, and MAE evaluation metrics to form the validation data set. The following training process is executed k times: the BP neural network model is trained based on the training data set and validated using the validation data set to obtain the validation result; The verification results from k trials are taken as the training score of the BP neural network model, and the BP neural network model is adjusted based on the training score. The construction of the BP neural network model also includes: The number of hidden layers in the BP neural network model is set to a single hidden layer; wherein the number of neurons in the single hidden layer is set to 1 / 3 to 2 times that of the input layer neurons; The evaluation results were obtained by using the NSE index. If the evaluation result shows that the NSE value reaches the preset threshold, then the performance is determined to have met expectations. If the evaluation result is that the NSE value does not reach the preset threshold, it is determined that the performance has not met expectations, and the number of hidden layers is set to double hidden layers; wherein, in the double hidden layers, the number of neurons in the first layer is set to 1 / 3 to 2 times that of the input layer neurons; and the number of neurons in the second layer is set to 2 / 3 times that of the input layer neurons; The construction of the BP neural network model also includes: The weights of the BP neural network model are initialized based on the Xavier method; wherein the expression for weight initialization is: ； Where ω represents the weight matrix; U represents a uniform distribution; n in n represents the number of neurons in the input layer of the neural network. out μ represents the number of neurons in the output layer of the neural network; μ is determined by the type of activation function selected; the activation function is any one of Tanh, Sigmoid, and Softsign. If the activation function is Tanh, then μ=6; if the activation function is Sigmoid, then μ=96; if the activation function is ReLU, then μ=12. The BP neural network model is trained using different activation functions, and evaluated using performance evaluation metrics to obtain the evaluation results for each activation function. The activation function Tanh is: ； The activation function Sigmoid is: ； The activation function Softsign is: ； Where x represents the sum of the input signals, including the weighted and biased terms of the previous layer; Based on the evaluation results, the activation function used in the BP neural network model is determined; The construction of the BP neural network model also includes: Based on the chlorophyll a concentration, a time series model is established; wherein, the expression of the time series model is: ； Where xt represents the actual observed value at time point t in the time series; ∅ represents the autoregressive coefficient; ε t ε represents the error term at time point t; t-1 ε t-2 ε t-q This represents the first q error terms in the time series; Using lag time as the input variable, the chlorophyll a concentration data at the lag time is calculated based on the time series model and used as the time training set; The BP neural network model is trained based on the time training set and evaluated using performance evaluation metrics to obtain evaluation results corresponding to different lag times. The BP neural network model is then adjusted based on the evaluation results corresponding to different lag times.

7. The method for predicting algal blooms in backwater areas as described in claim 1, characterized in that, The algal bloom risk prediction results include: If the chlorophyll a concentration Ccompound in the backwater area of ​​the target hydrological environment is ≤10, then the prediction result of the algal bloom risk is: algal bloom severity level is 1. If the chlorophyll a concentration C in the backwater area of ​​the target hydrological environment meets the condition 10 < C ≤ 15, then the predicted result of the algal bloom risk is: the algal bloom severity level is 2. If the chlorophyll a concentration C in the backwater area of ​​the target hydrological environment meets the condition 15 < C ≤ 50, then the predicted result of the algal bloom risk is: the algal bloom severity level is 3. If the chlorophyll a concentration C in the backwater area of ​​the target hydrological environment meets the condition 50 < C ≤ 100, then the predicted result of the algal bloom risk is: the algal bloom severity level is 4. If the chlorophyll a concentration C in the backwater area of ​​the target hydrological environment meets the condition C > 100, then the predicted result of the algal bloom risk is: the algal bloom severity level is 5.

8. A device for predicting algal blooms in a backwater area, characterized in that, include: An environmental data module is used to acquire geographic hydrological data and environmental monitoring data of the target hydrological environment, and to construct a three-dimensional hydrodynamic and temperature model of the backwater area based on the geographic hydrological data and the environmental monitoring data. This includes: collecting the geographic hydrological data and environmental monitoring data of the target hydrological environment, constructing a one-dimensional hydrological model of the inflow river, and calculating the riverbed roughness data in the target hydrological environment; based on the riverbed roughness data and the outflow data of the upstream reservoir of the inflow river in the target hydrological environment, constructing a one-dimensional hydrodynamic model of the inflow river, and calculating the inflow data and upstream flow data; based on the inflow data and the measured water level data of the reservoir, constructing a two-dimensional hydrodynamic model of the reservoir, and calculating the downstream water level data of the inflow river; and using the upstream flow data and the measured water level data of the reservoir, constructing a three-dimensional hydrodynamic and temperature model of the backwater area. The eutrophication module is also used to construct a three-dimensional eutrophication model of the backwater area based on the water temperature, water depth, horizontal flow velocity, and wind speed obtained from the three-dimensional hydrodynamic and temperature model of the backwater area. This includes: constructing the three-dimensional eutrophication model of the backwater area based on a preset eutrophication template; determining the boundary conditions and model forces of the three-dimensional eutrophication model of the backwater area; the boundary conditions include upstream river boundary conditions and downstream reservoir confluence boundary conditions; and calibrating the three-dimensional eutrophication model of the backwater area based on the boundary conditions, using the water temperature, water depth, horizontal flow velocity, and wind speed obtained from the three-dimensional hydrodynamic and temperature model of the backwater area, as well as the light intensity obtained from the meteorological station, as model forces. The neural network module is also used to construct a BP neural network model based on the three-dimensional eutrophication model of the backwater area, and to train the BP neural network model, including: using the three-dimensional eutrophication model to simulate the concentration of chlorophyll a according to different environmental parameters, generating a dataset containing environmental parameters and the concentration of chlorophyll a; and using the dataset to train the BP neural network model; wherein the input of the BP neural network model is the environmental parameters; and the output is the concentration of chlorophyll a. The risk prediction module is used to obtain the algal bloom risk prediction results of the backwater area of ​​the target hydrological environment using a trained neural network model.

Images