River flood season pollution intensity prediction and early warning method and system

Abstract

The invention provides a river flood season pollution intensity prediction and early warning method and system, and relates to the technical field of hydraulic engineering, and the method comprises the steps: obtaining multi-source data in a drainage basin, and carrying out the preprocessing of the multi-source data; constructing a hydrological water quality model for calculating the hydrological cycle process of the whole drainage basin from the upstream to the downstream; calculating point source and non-point source pollution loads to obtain water quality concentration; constructing a prediction data set according to the water quality concentration, and training the XGBoost machine learning model by using the prediction data set; and utilizing the trained XGBoost machine learning model to predict the pollutant concentration in the flood season under the influence of different meteorological conditions and carry out standard exceeding concentration early warning. The method can predict a short-term river water quality change peak value in advance by fusing weather forecast, a hydrological model and real-time monitoring data, realizes early warning of water quality change in advance, and is finally integrated to a drainage basin management decision-making platform, so that advanced regulation and control of a management department and advanced prevention and control of pollution are realized.

Description

Technical Field

[0001] The present invention belongs to the field of water conservancy projects, and particularly relates to a method and system for predicting and warning the pollution intensity during the flood season of rivers. Background Art

[0002] As a typical natural-economic-social complex system, urban and rural areas are facing the water environment pressure, water pollution and the resulting environmental risks brought about by the rapid growth of the population and the intensification of social and economic activities, which have become one of the key factors restricting the coordinated development of regional economy, society and environment. Although remarkable achievements have been made in the control of point source pollution, non-point source pollution and its variability are still deep-seated problems to be solved urgently. In particular, nitrogen and phosphorus have become the main pollutants in the Yangtze River Basin, coastal waters and many lakes, and non-point source pollution is gradually rising to become the key contradiction restricting the continuous improvement of the water ecological environment.

[0003] The pollution intensity during the flood season, as an important indicator to measure the change degree of pollutant concentration in rivers or lakes during the flood season (i.e., the period with more rainfall), is of great significance for evaluating the severity of water pollution during the flood season. However, existing hydrological models and pollution prediction methods often cannot fully cope with the non-linear complexity of non-point source pollution. Especially when the hydrological conditions and meteorological changes have a greater impact on the pollution load, the limitations of traditional methods become more prominent. Therefore, it is particularly urgent to carry out the simulation of the pollution intensity during the flood season driven by "meteorological conditions - pollutant concentration".

[0004] The pollution intensity during the flood season not only reflects the change of water environment quality, but also reflects the intensity of factors such as urban planning, infrastructure construction, pollution control measures and the ability to respond to climate change. If the pollution intensity during the flood season is too high, it will cause serious damage to the river ecosystem and directly threaten the drinking water safety and ecological health of humans. Especially in economically developed and densely populated areas, the acceleration of the urbanization process and the continuous growth of sewage discharge have made the pollution problem during the flood season increasingly serious, which has become a major problem in water environmental protection and ecological civilization construction.

[0005] In the face of the pollution problem during the flood season of rivers, the current detection means often have lag and limitations, and it is difficult to detect and effectively respond to sudden pollution incidents in a timely manner. Carrying out research on pollution simulation and warning during the flood season is of great significance for scientifically evaluating the degree of water pollution, predicting the change trend of pollution intensity and taking effective control measures in advance. Although the common hydrological models nowadays can better simulate the transformation and migration process of pollutants in the basin, they cannot predict the pollution situation during the flood season according to the changes of various influencing factors, and the hydrological models cannot accurately depict the non-linear relationships existing in some factors. Machine learning models are powerful tools for mining the laws of complex non-linear data, but the construction of machine learning models requires a large amount of data sets and is not universal for many areas lacking data.

[0006] Therefore, the method for predicting and warning pollution intensity during the flood season driven by meteorological conditions and pollutant concentrations has extremely important application value. Summary of the Invention

[0007] Aiming at the above deficiencies in the prior art, the present invention provides a method and system for predicting and warning pollution intensity during the flood season of rivers. The present invention proposes an innovative prediction and warning method, which can utilize a refined hydrological model and advanced machine learning technology to overcome the deficiencies of traditional methods, and accurately predict the pollution intensity during the flood season in real time, so as to provide effective warning support for decision-makers, take control measures in advance, and ensure water environment safety.

[0008] In order to achieve the above objectives, the technical solution adopted by the present invention is: A method for predicting and warning pollution intensity during the flood season of rivers, comprising the following steps: S1. Obtain multi-source data within the basin and preprocess the multi-source data; S2. Construct a hydrological and water quality model for calculating the hydrological cycle process from the upstream to the downstream of the entire basin; S3. Use the constructed hydrological and water quality model to calculate the point source and non-point source pollution loads to obtain the water quality concentration; S4. According to the water quality concentration, construct a prediction dataset and use the prediction dataset to train the XGBoost machine learning model; S5. Use the trained XGBoost machine learning model to predict the pollutant concentration during the flood season under the influence of different meteorological conditions and issue an early warning for the exceeded concentration, thereby completing the prediction and warning of the pollution intensity during the flood season of the river.

[0009] The beneficial effects of the present invention: The present invention provides a method for predicting and warning pollution intensity during the flood season of rivers based on the combination of a mechanism model and machine learning. First, based on any hydrological and water quality model, the hydrological and water quality changes at different time and space scales in the basin are simulated, and then the machine learning XGBoost model is used to achieve the prediction of pollutant concentration during the flood season under the influence of different meteorological conditions and the early warning of the exceeded concentration. The present invention can utilize a refined hydrological model and advanced machine learning technology to overcome the deficiencies of traditional methods, accurately predict the pollution intensity during the flood season in real time, so as to provide effective warning support for decision-makers, take control measures in advance, and ensure water environment safety.

[0010] Further, the specific content of S1 is as follows: S101. Obtain multi-source data within the basin and perform cleaning and correction processing on the multi-source data; S102. Identify and remove the outliers in the multi-source data processed by S101 through the use of statistical analysis methods, and fill in the missing values; S103. For the discontinuous time series of meteorological data and water quality monitoring data in the multi-source data processed in S102, use the linear interpolation method to insert several new time points between two adjacent dates, and use the trend of existing data points to calculate the meteorological data values and water quality concentrations at the new time points. S104. According to the data processed in S103, fuse the multi-source data in the basin to form a standard data set. Each data at each time point in the standard data set includes multi-dimensional information, and the preprocessing of the multi-source data is completed.

[0011] The beneficial effects of the above further solution are as follows: The above process aims to provide high-quality input data for the hydrological and water quality model and ensure the high compatibility and strong flexibility of the model. In this process, through machine learning and data processing methods, the high quality of the original data is ensured, laying a foundation for subsequent model calculation and prediction.

[0012] Furthermore, the hydrological cycle process in S2 includes atmospheric water process, soil water process, groundwater process and surface water process. Among them, the soil water balance formula followed by the soil water process is as follows:

[0013] Among them, represents the soil water storage volume on the current day, represents the soil water storage volume on the previous day, represents the irrigation water volume, represents the infiltration volume, represents the soil evaporation volume, represents the vegetation transpiration volume, represents the deep leakage volume, represents the subsurface flow; The groundwater process includes shallow groundwater process and deep groundwater process. The shallow groundwater balance formula followed by the shallow groundwater process is as follows:

[0014] Among them, represents the shallow groundwater storage volume on the current day, represents the shallow groundwater storage volume on the previous day, represents the shallow recharge volume, represents the base flow generation volume, represents the phreatic evaporation volume, represents the shallow water usage volume; The deep groundwater balance formula followed by the deep groundwater process is as follows:

[0015] Among them, represents the deep groundwater storage volume of the current day, represents the deep groundwater storage volume of the previous day, represents the deep recharge volume, represents the deep water usage volume; The water storage balance formula of the pond / wetland in the surface water process is as follows:

[0016] where, represents the pond / wetland water storage volume of the current day, represents the pond / wetland water storage volume of the previous day, represents the inflow volume, represents the outflow volume, represents the rainfall volume, represents the evaporation volume, represents the seepage volume; The water volume evolution process balance formula of the main river channel in the surface water process is as follows:

[0017] where, represents the main river channel storage volume of the current day, represents the main river channel storage volume of the previous day, represents the upstream river channel inflow volume, represents the downstream outflow volume, represents the riverbank storage volume, represents the irrigation and consumption volume, represents the evaporation volume of the river channel water surface, represents the seepage volume.

[0018] The beneficial effects of the above further solution are as follows: By refining the hydrological cycle process into four dynamic subsystems of atmospheric water, soil water, groundwater, and surface water, the refined modeling and multi-level coupling of the water cycle process are realized, and the water volume migration path of "precipitation - infiltration - evapotranspiration - runoff - recharge" is completely characterized.

[0019] Furthermore, the expression of the hydrological water quality model in S2 is as follows:

[0020] where, represents the soil water storage volume, represents time, represents the precipitation volume, represents the evaporation volume, represents the infiltration volume, represents the surface runoff volume.

[0021] The beneficial effects of the above further solution are as follows: By constructing a soil water dynamic equation based on the water balance, the physical process of the hydrological cycle is integrated with the pollutant migration path, realizing the collaborative simulation of the hydrological process and water quality evolution.

[0022] Furthermore, the specific content of S3 is as follows: S301. Based on the hydrological response unit, use a hydrological water quality model to simulate the hydrological process, and consider the generation, transformation, and migration processes of water quality pollutants. Among them, the pollutant generation and inflow amounts are calculated through the following formula:

[0023] Where, represents the pollutant load of the i th control unit HRU, n represents the number of pollution sources in the control unit HRU, and the pollution sources include point source pollution and non-point source pollution, represents the area of the control unit HRU, represents the pollutant concentration; S302. According to the changes in pollutant concentration and inflow amount in each control unit HRU, use the hydrological water quality model to calculate the pollutant inflow amount into the river in each control unit HRU. Among them, for point source pollution, the pollutant emission amount is calculated from the point source flow and concentration. For non-point source pollution, calculate the pollutant inflow amount into the river in each control unit HRU according to soil erosion, runoff, and pollution load:

[0024] Where, represents the total pollutant inflow amount into the river in the overall basin, represents the summation of all control units HRU; S303. According to the pollutant inflow amount into the river in each control unit HRU, calculate the water quality concentration, and complete the calculation of the point source and non-point source pollution loads:

[0025] Where, represents the water quality concentration, represents the pollution load amount, represents the sewage flow, represents the time conversion coefficient.

[0026] The beneficial effects of the above further solution are as follows: By using a hydrological water quality model to finely simulate the emission amounts of pollution control factors, the present invention can effectively obtain the pollutant generation amounts in each control unit.

[0027] Furthermore, the specific content of S4 is as follows: S401. Use the six meteorological data and the Julian date of historical years as the prediction variables of the XGBoost machine learning model, and construct the prediction datasets for permanganate index and total phosphorus concentration respectively; S402. Perform linear interpolation on the six meteorological data and the Julian date, and evenly insert three new time points between two adjacent Julian dates. For each new time point, perform linear interpolation; S403. Add the time series data obtained after linear interpolation, its corresponding meteorological data, and water quality concentration to the constructed prediction datasets for permanganate index and total phosphorus concentration to form a new dataset, and use the new dataset as the prediction dataset; S404. Divide the prediction dataset obtained in S403 into a training set and a validation set, and use the training set and the validation set to train and validate the XGBoost machine learning model; S405. According to the six meteorological data in the short-term future, use the verified XGBoost machine learning model to obtain the values of permanganate index, total phosphorus concentration, and pollution intensity at the control section during the flood season, and complete the training of the XGBoost machine learning model.

[0028] The beneficial effect of the above further solution is that by constructing a water quality prediction model based on XGBoost machine learning and combining meteorological driving factors with time series characteristics, the dynamic prediction of the concentrations of key pollutants (permanganate index, TP) during the flood season is realized.

[0029] Furthermore, the expression of the loss function of the XGBoost machine learning model is as follows:

[0030]

[0031] Among them, represents the loss function of the XGBoost machine learning model, i represents i samples, n represents the total number of samples, represents the predicted value and the true value the error between them, represents the regularization parameter, which is used to balance the loss function and the regularization term, k represents the k th decision tree, K represents the total number of trees in the model, represents the XGBoost machine learning model parameters; The expression of the regularization term of the XGBoost machine learning model is as follows:

[0032] Among them, represents the regularization penalty term of a single decision tree, represents the number of leaf nodes, represents the leaf node weight, j represents the index of the leaf node, represents the penalty coefficient of the number of leaf nodes.

[0033] The beneficial effects of the above further solution are as follows: Through the deep integration of mathematical constraints and algorithm characteristics, the XGBoost model not only ensures the prediction accuracy but also has professional rationality and engineering practicability, providing an intelligent prediction tool with both statistical significance and physical interpretability for water quality management.

[0034] Furthermore, the S5 includes the following steps: S501. Use the trained XGBoost machine learning model to predict the pollutant concentration during the flood season under different meteorological conditions; S502. According to the prediction results of S501, combined with the set warning threshold, trigger the warning mechanism for those exceeding the safety standard; S503. According to the triggered warning mechanism in S502, initiate emergency response measures and decision support to warn of the excessive concentration, and complete the prediction and warning of the pollution intensity during the river flood season.

[0035] The beneficial effects of the above further solution are as follows: By constructing a "prediction - warning - response" decision-making system, the early identification and precise control of the pollution risk during the river flood season are realized. This solution dynamically analyzes the non-linear relationship between meteorological factors and pollutant concentration based on the trained XGBoost model, and outputs the pollution intensity prediction results with high spatio-temporal resolution (such as hourly level), breaking through the lag bottleneck of traditional monitoring means; by setting the warning threshold and automatically triggering warnings in combination with real-time prediction values, the emergency response efficiency is improved from "post-disposal" to "pre-intervention". The present invention provides a prediction and warning system for river flood season pollution intensity. The prediction and warning system for river flood season pollution intensity is used to execute the prediction and warning method for river flood season pollution intensity, and includes: The first processing module is used to obtain multi-source data within the basin and preprocess the multi-source data; The second processing module is used to construct a hydrological and water quality model for calculating the hydrological cycle process from the upstream to the downstream of the entire basin; The third processing module is used to calculate the point source and non-point source pollution loads by using the constructed hydrological and water quality model to obtain the water quality concentration; The fourth processing module is used to construct a prediction data set according to the water quality concentration and train the XGBoost machine learning model using the prediction data set. The fifth processing module is used to predict the pollutant concentration during the flood season under different meteorological conditions and issue an early warning for the exceeded concentration by using the trained XGBoost machine learning model, so as to complete the prediction and early warning of the pollution intensity of the river during the flood season.

[0036] Advantages of the present invention: The present invention provides a method for predicting and early warning the pollution intensity of a river during the flood season based on the combination of a mechanism model and machine learning. First, the present invention simulates the hydrological and water quality changes at different time and space scales in the basin based on an arbitrary hydrological and water quality mechanism model, and then uses the machine learning XGBoost model to realize the prediction of the pollutant concentration during the flood season under different meteorological conditions and the early warning of the exceeded concentration. The present invention can utilize a refined hydrological model and advanced machine learning technology to overcome the deficiencies of traditional methods, predict the pollution intensity during the flood season in real time and accurately, so as to provide effective early warning support for decision-makers, take control measures in advance, and ensure the safety of the water environment. Description of the Drawings

[0037] Figure 1 It is a flowchart of the method of the present invention.

[0038] Figure 2 It is a technical roadmap of the present invention.

[0039] Figure 3 It is a schematic diagram of the hydrological cycle structure of the hydrological and water quality model of the present invention.

[0040] Figure 4 It is a schematic diagram of the control unit division in this embodiment.

[0041] Figure 5 It is a schematic diagram of the calibration and verification results of the flow rate and water quality concentration of the hydrological and water quality model.

[0042] Figure 6 It is a schematic diagram of the water quality concentration training and verification results of the XGBoost machine learning model.

[0043] Figure 7 It is a schematic diagram of the flood season pollution intensity prediction and early warning system in this embodiment.

[0044] Figure 8 It is a schematic diagram of the system structure of the present invention. Detailed Embodiments

[0045] The specific embodiments of the present invention will be described below to facilitate those skilled in the art of the present technology to understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those of ordinary skill in the art of the present technology, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions created using the concept of the present invention are within the scope of protection.

[0046] Embodiment 1 As Figure 1 - Figure 2 shown, the present invention provides a method for predicting and warning the pollution intensity during the flood season of rivers, and the implementation method is as follows: S1. Obtain multi-source data within the basin and preprocess the multi-source data. The implementation method is as follows: S101. Obtain multi-source data within the basin and perform cleaning and correction processing on the multi-source data; S102. Identify and remove outliers from the multi-source data processed by S101 by using statistical analysis methods, and fill in the missing values; S103. According to the multi-source data processed by S102, for the case where the time series of meteorological data and water quality monitoring data is discontinuous, use linear interpolation method to insert several new time points between two adjacent dates, and use the trend of existing data points to calculate the meteorological data values and water quality concentrations at the new time points; S104. According to the data processed by S103, fuse the multi-source data within the basin to form a standard data set. Among them, the data at each time point in the standard data set includes multi-dimensional information, and the preprocessing of the multi-source data is completed.

[0047] In this embodiment, data collection and preprocessing: Data collection and preprocessing is a crucial part of the present invention, aiming to provide high-quality data support for subsequent hydrological and water quality models and machine learning models.

[0048] Data collection involves collecting relevant data within the basin from multiple sources, including meteorological data (such as precipitation, temperature, humidity, solar radiation, wind speed, etc.), hydrological data (such as river flow, evaporation, groundwater level, etc.), water quality monitoring data (such as permanganate index, total phosphorus concentration, etc.), and pollution source data (such as industrial emissions, agricultural non-point source pollution, urban non-point source pollution, etc.). The data usually comes from different monitoring stations, meteorological networks, and remote sensing images, etc. Since the collected data may have missing, noisy data, or outliers, effective cleaning and correction are required in the data preprocessing stage. Identify and remove outliers by using statistical analysis methods (such as Z-Score or IQR method), and fill in the missing values by using linear interpolation or machine learning methods (such as KNN interpolation) to ensure the integrity and consistency of the data.

[0049] The time series of data may also be discontinuous. In particular, meteorological data and water quality monitoring data may not be recorded at every time point. To address this issue, a linear interpolation method is adopted to insert several new time points between two adjacent dates and use the trend of existing data points to infer the meteorological data values and water quality concentrations at these new time points. In this way, not only are the data gaps filled, but also the time continuity and accuracy of the data are ensured. To avoid the impact of scale differences between different data sources on the performance of machine learning models, all data are standardized or normalized, usually using Z-Score standardization or Min-Max normalization, so that the ranges of each feature are the same, thereby improving the training efficiency and stability of the model.

[0050] In the data fusion process, all the collected data are integrated to form a unified data set, where the data at each time point contains information in multiple dimensions, such as meteorological data, flow rate, pollution source data, etc. These data will be used as input features for training in subsequent hydrological and water quality simulations and machine learning models. At the same time, to further improve the accuracy and generalization ability of the model, feature engineering methods are used to extract meaningful features from the original data. For example, aggregated data at different time scales (such as daily, monthly, and quarterly average precipitation, temperature, etc.) are extracted, and through the construction of lag features, the pollutant concentrations or meteorological data of the past few days are used as inputs for current predictions to reflect the delay effect of pollutants. Through these data preprocessing steps, the quality and consistency of the input data are ensured, providing strong support for machine learning models and improving the accuracy and generalization ability of the models in subsequent prediction processes.

[0051] S2. Construct a hydrological and water quality model for calculating the hydrological cycle process of the entire basin from upstream to downstream; In this embodiment, the construction and integration of the hydrological and water quality model: Using geospatial information technology, comprehensively considering the processes of atmospheric precipitation and irrigation input to the soil in the hydrological cycle, as well as the processes of evaporation, infiltration, runoff, etc. output from the soil. The model includes calculations of the atmosphere, soil, rivers, reservoirs, and management measures, etc. The hydrological cycle structure of the hydrological and water quality model is as Figure 3 shown, calculating the hydrological cycle of the entire basin from upstream to downstream. The principle of the water cycle process can be subdivided into 4 categories, including atmospheric water process, soil water process, groundwater process, and surface water process: (1) The atmospheric water process is mainly divided into rainfall and snowfall, and the snowfall process will subsequently affect the water cycle through snow accumulation and snowmelt processes.

[0052] (2) The soil water process is divided into 6 processes: infiltration, soil surface evaporation, vegetation transpiration, subsurface flow, deep percolation, and irrigation, following the soil water balance formula:

[0053] Among them, represents the soil water storage on the current day, with the unit of mm; represents the soil water storage on the previous day, with the unit of mm; represents the irrigation water volume, with the unit of mm; represents the infiltration amount, with the unit of mm; represents the soil evaporation amount, with the unit of mm; represents the vegetation transpiration amount, with the unit of mm; represents the deep percolation amount, with the unit of mm; represents the interflow, with the unit of mm.

[0054] (3) The groundwater simulation by the hydrological and water quality model has the following characteristics: ① Groundwater is divided into shallow groundwater and deep groundwater; ② Only the water balance is simulated, and the groundwater level is not simulated; ③ The groundwater between sub-basins is an independent relationship, and the influence of lateral runoff is not considered.

[0055] Among them, the main processes of shallow groundwater include 5 processes: recharge, base flow, phreatic evaporation, irrigation, and consumption, following the shallow groundwater water balance formula:

[0056] Among them, represents the shallow groundwater storage on the current day, with the unit of mm; represents the shallow groundwater storage on the previous day, with the unit of mm; represents the shallow recharge amount, with the unit of mm; represents the base flow generation amount, with the unit of mm; represents the phreatic evaporation amount, with the unit of mm; represents the shallow usage amount (irrigation / consumption), with the unit of mm.

[0057] The main processes of deep groundwater are divided into 3 processes: recharge, irrigation, and consumption, following the deep groundwater water balance formula:

[0058] Among them, represents the deep groundwater storage on the current day, with the unit of mm; represents the deep groundwater storage on the previous day, with the unit of mm; represents the deep recharge amount, with the unit of mm; represents the deep usage amount (irrigation / consumption), with the unit of mm.

[0059] (4) The surface water processes of the hydrological and water quality model include three parts: ① the sub-channel transmission process; ② the storage process in reservoirs, ponds, wetlands, and depressions; ③ the water volume evolution process in the main channel. Among them, the sources of surface water include: ① the surface runoff from the hydrological response unit; ② the subsurface flow generated by the hydrological response unit; ③ the base flow generated by the shallow groundwater in the sub-basin.

[0060] There are three paths in total for the transmission process of surface water in the sub-basin: ① The surface runoff from the hydrological response unit first undergoes the sub-channel process, resulting in transmission losses; ② A part of the remaining surface runoff, subsurface flow, and groundwater base flow enters the ponds / wetlands in the sub-basin, and a part enters the main channel of the sub-basin; ③ The outflow from the ponds / wetlands enters the main channel of the sub-basin.

[0061] The storage in ponds / wetlands includes the processes of inflow, rainfall and evaporation, infiltration, outflow, and consumption (only for ponds). Each process follows the water balance formula for ponds / wetlands:

[0062] Among them, represents the storage volume of ponds / wetlands on the current day, with the unit of m³; represents the storage volume of ponds / wetlands on the previous day, with the unit of m³; represents the inflow volume, with the unit of m³; represents the outflow volume, with the unit of m³; represents the rainfall volume, with the unit of m³; represents the evaporation volume, with the unit of m³; represents the leakage volume, with the unit of m³.

[0063] The water volume evolution process in the main channel includes the inflow from the upstream channel, the evaporation on the river surface, the storage in the riverbank, the leakage, the outflow to the downstream, the consumption, and the irrigation. Each process follows the water balance formula for the main channel:

[0064] Among them, represents the storage volume of the main channel on the current day, with the unit of m³; represents the storage volume of the main channel on the previous day, with the unit of m³; represents the inflow volume from the upstream channel, with the unit of m³; represents the outflow volume to the downstream, in m³; represents the storage volume in the riverbank, with the unit of m³; represents the irrigation and consumption volume, with the unit of m³; represents the evaporation volume on the river surface, with the unit of m³; represents the leakage volume, with the unit of m³.

[0065] In this embodiment, for the simulation of hydrological mechanisms: a hydrological and water quality mechanism model is used to model the hydrological process and the process of pollutant migration and transformation in the basin. Assume that the basic expression of the hydrological and water quality model is as follows:

[0066] Where, represents the soil water storage, represents time, represents precipitation, represents evaporation, represents infiltration, represents surface runoff. This equation describes the water change in the hydrological cycle and takes into account different input and output processes.

[0067] In this embodiment, for the simulation of pollution load during the flood season: the simulation of pollutants is also based on the hydrological and water quality model, and a pollution load model is used for simulation. A common water quality concentration model can be expressed as:

[0068] Where, represents time the pollutant concentration at time represents the initial concentration, represents the pollutant degradation rate, represents the pollutant emission. The model simulates the change of pollutants over time and their degradation process.

[0069] S3. Using the constructed hydrological and water quality model, calculate the point source and non-point source pollution loads to obtain the water quality concentration. The implementation method is as follows: S301. Based on the hydrological response unit, use the hydrological and water quality model to simulate the hydrological process, and consider the generation, transformation and migration process of water quality pollutants; S302. According to the change of pollutant concentration and inflow in each control unit HRU, use the hydrological and water quality model to calculate the pollutant inflow into the river in each control unit HRU. Among them, for point source pollution, the pollutant emission is calculated from the point source flow and concentration. For non-point source pollution, the pollutant inflow into the river in each control unit HRU is calculated according to soil erosion, runoff and pollution load; S303. According to the pollutant inflow into the river in each control unit HRU, calculate the water quality concentration to complete the calculation of point source and non-point source pollution loads.

[0070] In this embodiment, the refined accounting of point source and non-point source pollution loads is carried out. Based on the control units HRUs, a hydrological and water quality mechanism simulation model is constructed to simulate the hydrological processes and pollutant migration and transformation processes of each hydrological response unit under different meteorological conditions. Through the hydrological cycle and pollutant load calculation of the model, the pollutant inflow into the river in each control unit (i.e., HRUs) is obtained. The specific steps are as follows: ① Input of hydrological and water quality model: Based on the hydrological response units, a hydrological and water quality model is used to simulate hydrological processes such as precipitation, evaporation, runoff, and infiltration, and the generation, transformation, and migration processes of water quality pollutants are considered. The generation and inflow of pollutants are calculated by the following formula:

[0071] Among them, represents the pollutant load of the i th control unit HRU, n represents the number of pollution sources in the control unit HRU, and the pollution sources include point source pollution and non-point source pollution (such as agricultural non-point source pollution, urban non-point source pollution, etc.), represents the area of the control unit HRU, represents the pollutant concentration.

[0072] ② Calculation of pollutant load inflow into the river: According to the changes in pollutant concentration and flow in each control unit HRU, the hydrological and water quality model calculates the pollutant inflow into the river in each control unit HRU. For point source pollution (such as industrial enterprises, sewage treatment plants, etc.), the pollutant emissions are calculated from the point source flow and concentration; for non-point source pollution (such as agricultural non-point source and urban non-point source pollution), the pollutant inflow into the river in each control unit HRU is calculated according to soil erosion, runoff, and pollution load. The specific formula for pollutant inflow into the river is as follows:

[0073] Among them, represents the total pollutant inflow into the river in the whole basin, represents the accumulation of all control units HRU.

[0074] ③ Data processing of pollutant inflow into the river during the flood season: The calculated pollutant inflow data needs to be processed and then used to simulate the pollution concentration of the basin water body to further evaluate the changing trend of pollution intensity during the flood season. Specifically, these pollutant inflow data will be used as inputs in the flow-water quality simulation stage, combined with the hydrological and water quality model to predict the water quality changes in the basin, and provide data support for the machine learning model, so as to achieve high-precision prediction of pollution intensity during the flood season.

[0075] The hydrological and water quality model simulates the flow by simulating processes such as rainfall infiltration, evapotranspiration, soil moisture, groundwater, and runoff for each hydrological response unit through the imported meteorological data (daily cumulative rainfall, daily maximum temperature, daily minimum temperature, daily relative humidity, daily average solar radiation, daily average wind speed). At the same time, the input of pollutants is set, including point source and non-point source pollution, to simulate the whole process of pollutants from generation to entering the river to obtain the pollution load into the river. The pollution load into the river is used to calculate the water quality data through the following formula, and the model is calibrated and verified based on the measured data to make the simulation accuracy meet the model requirements:

[0076] Among them, represents the water quality concentration, with the unit of mg / L, represents the pollution load, with the unit of kg / d, represents the sewage flow, with the unit of m³ / d, represents the time conversion coefficient.

[0077] S4. According to the water quality concentration, construct a prediction dataset, and use the prediction dataset to train the XGBoost machine learning model. The implementation method is as follows: S401. Take the 6 meteorological data and Julian date of historical years as the prediction variables of the XGBoost machine learning model, and construct the prediction datasets for permanganate index and total phosphorus concentration respectively; S402. Perform linear interpolation on the 6 meteorological data and Julian date, and evenly insert 3 new time points between two adjacent Julian dates. For each new time point, perform linear interpolation; S403. Add the time series data obtained after linear interpolation, its corresponding meteorological data, and water quality concentration to the constructed prediction datasets for permanganate index and total phosphorus concentration to form a new dataset, and use the new dataset as the prediction dataset; S404. Divide the prediction dataset obtained in S403 into a training set and a validation set, and use the training set and the validation set to train and validate the XGBoost machine learning model; S405. According to the 6 meteorological data in the short term in the future, use the verified XGBoost machine learning model to obtain the values of permanganate index, total phosphorus concentration, and pollution intensity at the control section during the flood season, and complete the training of the XGBoost machine learning model.

[0078] In this embodiment, six meteorological data of historical years and the Julian date are used as the prediction variables of the XGBoost machine learning model, and prediction datasets for permanganate index and TP concentration are constructed respectively. To further expand the data volume of the machine learning model, linear interpolation is performed on the Julian date and six meteorological data. An average of three new time points are inserted between two adjacent Julian dates. For each new time point, six meteorological data values, permanganate index, and TP concentration are obtained using the linear interpolation method. The interpolated time series data and its corresponding meteorological data and water quality concentration are added to the original dataset to form a new dataset, which serves as the prediction dataset for permanganate index and TP concentration. The prediction dataset is divided into a training set and a validation set in proportion for training and validating the machine learning model. Among them, 80% of the dataset is used as the training set, and 20% is used as the validation set. Finally, based on the six meteorological data announced in the short-term future in the weather forecast, the values of permanganate index, total phosphorus concentration, and pollution intensity at the control section during the flood season can be obtained from the verified XGBoost machine learning model. To improve the prediction ability of the XGBoost machine learning model and prevent overfitting, the present invention uses the objective function of the XGBoost machine learning model to optimize the performance of the model. The objective function of the XGBoost machine learning model usually includes a loss function and a regularization term, and its goal is to minimize the following expression:

[0079] The loss function usually takes forms such as mean squared error (MSE) or logarithmic loss (Log Loss), and the specific selection is determined according to the nature of the problem and the distribution of the data. In this model, the loss function is specifically:

[0080] Among them, represents the loss function of the XGBoost machine learning model, i represents i samples, n represents the total number of samples, represents the predicted value and the true value the error between them, represents the regularization parameter, which is used to balance the loss function and the regularization term, k represents the k th decision tree, K represents the total number of trees in the model, represents the XGBoost machine learning model parameters; The regularization term is used to control the model complexity and prevent overfitting:

[0081] Among them, represents the regularization penalty term of a single decision tree, represents the number of leaf nodes, represents the weights of leaf nodes, j represents the indices of leaf nodes, represents the penalty coefficient for the number of leaf nodes.

[0082] The regularization term helps to adjust the complexity of the model, enabling the model to maintain good generalization ability while ensuring accuracy.

[0083] Finally, the model optimizes the objective function through Gradient Boosting, that is, by iteratively training a series of decision trees to gradually correct the errors in the previous round of predictions and optimize the objective function value, thereby obtaining higher prediction accuracy.

[0084] In this embodiment, the self-adaptive learning characteristic of the XGBoost machine learning model enables the model to dynamically adjust according to the pollution characteristics under different regions, seasons, and meteorological conditions. By validating the training sets of different regions and seasons, the model can continuously optimize its prediction ability and improve its adaptability and flexibility in a changing environment.

[0085] S5. Use the trained XGBoost machine learning model to predict the pollutant concentrations during the flood season under the influence of different meteorological conditions and issue early warnings for excessive concentrations, thereby completing the prediction and early warning of the pollution intensity of the river during the flood season. The implementation method is as follows: S501. Use the trained XGBoost machine learning model to predict the pollutant concentrations during the flood season under the influence of different meteorological conditions; S502. According to the prediction results of S501 and in combination with the set early warning threshold, trigger the early warning mechanism for exceeding the safety standard; S503. According to the triggered early warning mechanism of S502, initiate emergency response measures and decision support to issue early warnings for excessive concentrations, thereby completing the prediction and early warning of the pollution intensity of the river during the flood season.

[0086] In this embodiment, risk early warning and decision support are one of the key applications of the present invention, aiming to provide timely and scientific decision-making basis for environmental protection and water quality management departments by accurately predicting the changes in the pollution intensity during the flood season. After the data collection, preprocessing, simulation, and prediction model are completed, the predicted results of the pollution intensity during the flood season will be directly used as the input of the decision support system to provide early warning information for relevant departments. The system accurately evaluates the changing trend of the pollutant concentrations during the flood season and timely identifies potential pollution risks by analyzing historical data, real-time monitoring data, and future meteorological predictions in combination with the output of the machine learning model.

[0087] ① In the risk warning process, first, based on the pollution intensity data predicted by the model and combined with the set warning threshold, it automatically determines whether the safety standard is exceeded, thereby triggering the warning mechanism. For example, when the predicted pollution intensity reaches or exceeds the set warning threshold, the system generates a warning report and promptly sends the warning information to relevant management departments through an automated platform. These warning reports will include key information such as the predicted values of pollutant concentrations, the possible distribution of pollution sources, and the warning levels, helping decision-makers take countermeasures as early as possible, such as strengthening the supervision of pollution sources, initiating emergency responses, or releasing public warning information.

[0088] In practical applications, the risk warning system not only relies on the input of historical data and real-time data but also can be dynamically adjusted according to the latest weather forecasts (such as future rainfall, temperature changes, etc.). For example, when the weather forecast predicts heavy rainfall in the coming days, the system automatically adjusts the parameters of the warning model, taking into account the impact of precipitation on pollutant runoff, thereby improving the accuracy and timeliness of the prediction. This adaptive prediction adjustment ability enables the warning system of the present invention to operate efficiently under dynamically changing environmental conditions and respond promptly to changing risk factors.

[0089] ② In terms of decision support, the present invention not only provides regular pollution intensity predictions but also can provide emergency response suggestions and decision support based on real-time data and prediction results. By combining warning information with information such as the spatial distribution of basin pollution sources and historical emergency response measures, the system can propose optimized emergency plans and pollution control measures. For example, in areas with a high warning level, the system will recommend that relevant departments strengthen the monitoring of pollution sources or propose suggestions such as enhancing the operation of water purification facilities and allocating emergency water quality monitoring resources.

[0090] In addition, the risk warning and decision support system can also predict the effects of different emergency plans through intelligent analysis, helping decision-makers evaluate the effectiveness of different management measures. By simulating the implementation effects of different policies, decision-makers can select the most suitable countermeasures for the current basin pollution situation and resource conditions, thereby improving the efficiency of pollution prevention and control and reducing potential ecological risks and public health hazards.

[0091] The present invention provides strong decision support for basin pollution management by combining a high-precision prediction model, real-time data monitoring, and weather forecasts. It not only improves the timeliness and accuracy of warnings but also provides scientific emergency response suggestions for environmental protection departments, thereby effectively reducing the risk of pollution incidents and ensuring the safety of water resources and the health of the public.

[0092] In summary, the present invention uses a hydrological water quality model to simulate hydrological elements and pollution loads at the outlet of a basin, making up for the lack of numerous basin data and providing big data support for machine learning models. Based on the simulation process combining the hydrological water quality model and the machine learning model, it can make up for the deficiency of data in traditional mechanism models and can predict the changes in pollution intensity during the flood season more quickly and accurately. In addition to realizing the above functions, the model can carry any common hydrological water quality model, with wide adaptability and flexibility. Specifically, the present invention supports the following common hydrological water quality models: SWAT model (Soil and Water Assessment Tool), HSPF model (Hydrological Simulation Program—Fortran), HEC-HMS model (Hydrologic Engineering Center-Hydrologic Modeling System), and MIKE11 model. The model can be integrated with different hydrological water quality models, make full use of the advantages of each model, and select the most suitable model for simulating the hydrological water quality process according to specific requirements. Through this integration, the model can adapt to the characteristics of different basins and provide more accurate prediction and early warning of pollution intensity during the flood season.

[0093] The present invention will be further described in detail below with reference to the accompanying drawings.

[0094] 1) Data processing and integration: The aim is to provide high-quality input data for the hydrological water quality model and ensure high compatibility and strong flexibility of the model. The system first collects and collates various data sources within the basin, including digital elevation data, land use data, soil data, meteorological data, pollution source data, and water quality monitoring data, etc. All data is standardized and preprocessed, including interpolation of missing data, correction of outliers, etc., to ensure the integrity and consistency of the data. In this process, through machine learning and data processing methods, the high quality of the original data is ensured, laying a foundation for subsequent model calculations and predictions. To ensure that the data can be compatible with a variety of common hydrological water quality models, the present invention performs format unification and standardization processing on all data, enabling it to seamlessly connect with different models, such as SWAT, HSPF, HEC-HMS, etc., to ensure efficient data transfer between various hydrological water quality models. Through this data integration method, all source data is fused into a unified data set, which can not only be directly input into various hydrological water quality models, but also provide high-quality training data for machine learning models, further improving the prediction accuracy. Through this data integration and processing process, the system has high compatibility with existing mainstream hydrological water quality models, enabling it to flexibly respond to changes in different basins and pollution situations, and enhancing the application universality and flexibility of the system.

[0095] 2) Fine accounting of pollution load emissions from point - area source superposition: Use a hydrological and water quality model to finely simulate the emissions of pollution control factors, and obtain the pollutant generation amounts of each control unit. The division results are as Figure 4 shown. Point sources include industrial pollution, urban domestic sewage, and large - scale livestock and poultry breeding pollution sources. Additional non - point sources include urban surface sources, rural domestic sewage, and aquaculture pollution sources. The calculated river - bound pollution load amounts of point - area sources in the basin are used as point sources and additional non - point sources to input into the hydrological and water quality mechanism model respectively. Among them, the point - source pollution load is statistically input corresponding to the point - source pollution interface of Point Source Discharges, and the non - point - source pollution load corresponds to the management measures of the hydrological and water quality mechanism model and is input into the model in the form of agricultural fertilizers.

[0096] 3) Flow - water quality simulation: Use a hydrological and water quality model to simulate the hydrological - water quality change process, and calibrate the model based on measured data to make the model simulation reach the best effect. For the simulation of flow and pollution load by the hydrological and water quality model, the Nash efficiency coefficient (NSE) and the coefficient of determination (R 2 ) are used to evaluate the calibration effect of the model. When R 2 > 0.60 and NSE> 0.50, the model flow simulation result is considered "satisfactory"; when R 2 > 0.40 and NSE> 0.35, the model pollution load simulation result is considered "satisfactory". Figure 5 Shows the simulated values and measured values of flow and water quality concentration obtained by the hydrological and water quality mechanism model.

[0097] 4) Construction of a flood - season pollution intensity prediction model based on the XGBoost machine learning model: It includes dataset construction and segmentation, model training, and model performance evaluation. The Julian date is corresponded to 6 meteorological data items (including daily cumulative rainfall, daily average radiation value, daily maximum temperature, daily minimum temperature, daily average relative humidity, daily average wind speed). To further expand the data volume of the machine learning model, linear interpolation is performed on the Julian date and 6 meteorological data items. An average of 3 new time points are inserted between two adjacent Julian dates. For each new time point, the linear interpolation method is used to obtain the 6 meteorological data values, permanganate index, and TP concentration. Specifically, for each pair of adjacent Julian dates and , as well as the corresponding meteorological data or water quality concentration and , the following formula is used to calculate the meteorological data or water quality concentration corresponding to the interpolated time point :

[0098] Among them, Indicates the time point after linear interpolation The corresponding meteorological data value Indicates the Julian date The corresponding meteorological data value Indicates the Julian date The corresponding meteorological data value

[0099] Add the interpolated time series data and its corresponding meteorological data and water quality concentration to the original data set to form a new data set as the prediction data set for permanganate index and TP concentration. Divide this data set into two parts, with 80% of the data as the training set and 20% of the data as the validation set. Secondly, when using the training set data to train the model, the best parameter combination is obtained by manually tuning the parameters. The best parameters are: the learning rate (Eta) is 0.1536, the regularization parameter (Gamma) is 0.001, the maximum depth of the tree (Max_Depth) is 6, the sample sampling rate of each tree (Subsample) is 0.91, the feature sampling rate (Colsample_Bytree) is 0.48, the number of boosting iterations (Nrounds) is 300, the number of early stopping iterations (Early_Stopping_Rounds) is 7, and the remaining parameters are set to the default values in the model; finally, based on the test set data, the performance of the permanganate index and TP concentration flood season pollution intensity prediction model is evaluated using linear regression, and the results are as Figure 6 shown

[0100] Based on the 6 meteorological data announced in the short-term future in the weather forecast, the values of permanganate index, total phosphorus concentration and pollution intensity at the control section during the flood season can be obtained from the verified XGBoost model

[0101] 5) Prediction and warning system: Based on the calibrated and verified XGBoost machine learning model, the present invention designs an automated prediction and warning system for accurately predicting water quality information and evaluating the pollution intensity during the flood season. The results are as Figure 7As shown in the figure. This system can not only predict the permanganate index and total phosphorus concentration, but also analyze and evaluate the changing trend of the water quality in the basin in real time, and then predict the dynamic changes of the pollution intensity. The system takes the historical monitoring data from users (such as the permanganate index and total phosphorus concentration) and combines the meteorological data provided by the weather forecast service (including key meteorological parameters such as precipitation, temperature, humidity, wind speed, etc.) as input variables, and uses the XGBoost machine learning model for high-precision prediction. During the operation of the system, based on the input real-time data and prediction results, the XGBoost machine learning model can dynamically evaluate the pollution intensity in the future flood season. When the system detects that the predicted pollution intensity value reaches or exceeds the preset safety threshold, it automatically triggers the warning mechanism. This mechanism can quickly identify potential pollution risks and send real-time warning notifications to relevant managers or decision-makers through the system interface, providing pollution intensity prediction data and response suggestions. This process ensures that targeted emergency measures can be taken in a timely manner when the pollution intensity exceeds the standard, thus effectively guaranteeing the safety of the water environment and the timeliness of water quality monitoring.

[0102] The core innovation of this system lies in achieving high-precision water quality prediction and high adaptability by integrating multi-source data (such as weather forecasts, historical water quality data, pollution source data, etc.) and machine learning technologies. Through the real-time analysis of the changing trend of water quality by machine learning algorithms, the system can flexibly respond to sudden pollution risks under different basins and different meteorological conditions, providing accurate prediction support and decision-making basis for water resource management.

[0103] Embodiment 2 As Figure 8 As shown in the figure, the present invention provides a prediction and warning system for the pollution intensity during the flood season of a river, characterized in that the prediction and warning system for the pollution intensity during the flood season of the river is used to execute the prediction and warning method for the pollution intensity during the flood season of the river described in Embodiment 1, including: The first processing module is used to obtain multi-source data within the basin and preprocess the multi-source data; The second processing module is used to construct a hydrological and water quality model for calculating the hydrological cycle process from the upstream to the downstream of the entire basin; The third processing module is used to calculate the point source and non-point source pollution loads by using the constructed hydrological and water quality model to obtain the water quality concentration; The fourth processing module is used to construct a prediction data set according to the water quality concentration and train the XGBoost machine learning model by using the prediction data set; The fifth processing module is used to use the trained XGBoost machine learning model to predict the pollutant concentration during the flood season under the influence of different meteorological conditions and issue an early warning for the exceeded concentration, so as to complete the prediction and warning of the pollution intensity during the flood season of the river.

[0104] AsFigure 8 As shown, the river flood season pollution intensity prediction and early warning system provided by the embodiment can execute the technical solutions shown in the river flood season pollution intensity prediction and early warning method of the above method embodiment. The implementation principle and beneficial effects are similar, and will not be elaborated here.

[0105] In this embodiment, the present application can divide functional units according to the river flood season pollution intensity prediction and early warning method. For example, each function can be divided into each functional unit, or two or more functions can be integrated into one processing unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit. It should be noted that the division of units in the present invention is illustrative, only a logical division, and there may be other division methods in actual implementation.

[0106] In this embodiment, in order to realize the principle and beneficial effects of the river flood season pollution intensity prediction and early warning method, the river flood season pollution intensity prediction and early warning system includes the corresponding hardware structure and / or software module for executing each function. Those skilled in the art should easily realize that, combined with the schematic units and algorithm steps described in the embodiments disclosed in the present invention, the present invention can be implemented in the form of hardware and / or a combination of hardware and computer software. Whether a certain function is executed in the way of hardware or computer software driving depends on the specific application and design constraint conditions of the technical solution. Different methods can be used for each specific application to implement the described function, but such implementation should not be considered to exceed the scope of this application.

Claims

1. A method for predicting and warning river pollution intensity during flood season, characterized in that: The following steps are involved: S1, obtain multi-source data in the basin and pre-process the multi-source data; S2. Construct a hydrological and water quality model to calculate the hydrological cycle process of the entire basin from upstream to downstream; S3. Using the constructed hydrological and water quality model, the pollution loads of point sources and non-point sources are calculated to obtain water quality concentrations; S4. Construct a prediction data set based on water quality concentration, and use the prediction data set to train the XGBoost machine learning model; S5. Use the trained XGBoost machine learning model to predict pollutant concentrations during the flood season under different meteorological conditions and issue warnings for concentrations exceeding the standard, thereby completing the prediction and warning of river pollution intensity during the flood season.

2. The method for predicting and warning river pollution intensity during flood season according to claim 1, characterized in that: The S1 is specifically: S101, obtaining multi-source data in the watershed, and cleaning and correcting the multi-source data; S102, identifying and removing outliers from the multi-source data processed by S101 by using statistical analysis methods, and filling in missing values; S103, according to the multi-source data processed by S102, in view of the discontinuous time series of the meteorological data and the water quality monitoring data, a number of new time points are inserted between two adjacent dates using linear interpolation method, and the meteorological data value and water quality concentration at the new time point are calculated using the trend of the existing data points; S104. Based on the data processed by S103, the multi-source data in the watershed are integrated to form a standard data set, wherein the data at each time point in the standard data set includes multi-dimensional information, thereby completing the preprocessing of the multi-source data.

3. The method for predicting and warning river pollution intensity during flood season according to claim 1, characterized in that: The hydrological cycle process in S2 includes atmospheric water process, soil water process, groundwater process and surface water process, wherein the soil water balance formula followed by the soil water process is as follows: in, Indicates the soil water storage on that day. Indicates the soil water storage on the previous day. Indicates the amount of irrigation water. represents the infiltration volume, represents soil evaporation, represents the evapotranspiration of vegetation, Indicates the amount of deep leakage, It means flow in soil; The groundwater process includes shallow groundwater process and deep groundwater process. The shallow groundwater process follows the shallow groundwater water balance formula as follows: in, Indicates the shallow groundwater storage volume on that day, represents the shallow groundwater storage volume on the previous day, represents the shallow recharge, represents the base flow generation, represents the amount of evaporation from diving, Indicates shallow usage; The deep groundwater balance formula followed by the deep groundwater process is as follows: in, Indicates the amount of deep groundwater storage on that day. represents the amount of deep groundwater storage on the previous day, Indicates the deep recharge amount, Indicates deep usage; The pond / wetland water balance equation for surface water storage is as follows: in, Indicates the amount of pond / wetland storage on that day. Indicates the amount of pond / wetland storage on the previous day. represents the inflow rate, Indicates the flow rate, Indicates rainfall, Indicates the evaporation rate, Indicates the amount of leakage; The main river water balance formula followed by the main river water volume evolution process in the surface water process is as follows: in, Indicates the storage volume of the main river channel on that day. Indicates the storage volume of the main river channel on the previous day. represents the upstream river inflow, represents the downstream outflow, represents the riverbank storage volume, Indicates irrigation and consumption, represents the evaporation of the river surface. Indicates the amount of leakage.

4. The method for predicting and warning river pollution intensity during flood season according to claim 1, characterized in that: The expression of the hydrological and water quality model in S2 is as follows: in, Represents soil water storage, Indicates time, Indicates the amount of precipitation, Indicates the evaporation rate, represents the infiltration volume, Represents surface runoff.

5. The method for predicting and warning river pollution intensity during flood season according to claim 1, characterized in that: The S3 is specifically: S301. Based on the hydrological response unit, the hydrological and water quality model is used to simulate the hydrological process, and the generation, transformation and migration process of water quality pollutants are considered. The generation and inflow of pollutants are calculated by the following formula: in, Indicates i The pollutant load of each control unit HRU, n Indicates the number of pollution sources in the control unit HRU, including point source pollution and surface source pollution. represents the area of ​​the control unit HRU, Indicates the concentration of pollutants; S302. According to the changes in pollutant concentration and inflow in each control unit HRU, the amount of pollutants entering the river in each control unit HRU is calculated using the hydrological and water quality model. For point source pollution, the pollutant discharge is calculated from the point source flow and concentration. For non-point source pollution, the amount of pollutants entering the river in each control unit HRU is calculated based on soil erosion, runoff and pollution load: in, It represents the total amount of pollutants entering the river in the entire basin. Represents the accumulation of all control units HRU; S303, according to the amount of pollutants entering the river in each control unit HRU, the water quality concentration is calculated to complete the calculation of point source and non-point source pollution loads: in, Indicates water concentration, represents the pollution load, represents the sewage flow rate, Indicates the time conversion factor.

6. The method for predicting and warning river pollution intensity during flood season according to claim 1, characterized in that: The S4 is specifically: S401, using six meteorological data items and Julian date of historical years as prediction variables of XGBoost machine learning model, and constructing the prediction data sets of permanganate index and total phosphorus concentration respectively; S402, performing linear interpolation processing on the six meteorological data and the Julian date, inserting three new time points on average between two adjacent Julian dates, and performing linear interpolation processing on each new time point; S403, adding the time series data obtained after the linear interpolation processing and its corresponding meteorological data and water quality concentration to the constructed permanganate index and total phosphorus concentration prediction data set to form a new data set, and using the new data set as the prediction data set; S404, dividing the prediction data set obtained in S403 into a training set and a validation set, and using the training set and the validation set to train and validate the XGBoost machine learning model; S405. Based on six meteorological data in the short term, the verified XGBoost machine learning model is used to obtain the permanganate index, total phosphorus concentration and pollution intensity values ​​of the control section during the flood season, and the training of the XGBoost machine learning model is completed.

7. The method for predicting and warning river pollution intensity during flood season according to claim 1, characterized in that: The loss function of the XGBoost machine learning model is expressed as follows: in, represents the loss function of the XGBoost machine learning model, i express i samples, n represents the total number of samples, Represents the predicted value and the true value The error between represents the regularization parameter, which is used to balance the loss function and the regularization term. k Indicates k A decision tree, K represents the total number of trees in the model, Represents XGBoost machine learning model parameters; The expression of the regularization term of the XGBoost machine learning model is as follows: in, represents the regularization penalty term of a single decision tree, Indicates the number of leaf nodes, represents the leaf node weight, j Represents the index of the leaf node, The penalty factor representing the number of leaf nodes.

8. The method for predicting and warning river pollution intensity during flood season according to claim 1, characterized in that: The S5 comprises the following steps: S501. Use the trained XGBoost machine learning model to predict pollutant concentrations during the flood season under different meteorological conditions. S502, according to the prediction result of S501, combined with the set warning threshold, triggering the warning mechanism for exceeding the safety standard; S503. According to the trigger warning mechanism of S502, emergency response measures and decision support are initiated to issue warnings for concentrations exceeding the standard and complete the prediction and warning of pollution intensity of rivers during the flood season.

9. A river flood season pollution intensity prediction and early warning system, characterized in that: The river flood season pollution intensity prediction and early warning system is used to execute the river flood season pollution intensity prediction and early warning method according to any one of claims 1 to 8, comprising: The first processing module is used to obtain multi-source data in the watershed and pre-process the multi-source data; The second processing module is used to construct a hydrological and water quality model that calculates the hydrological cycle process of the entire basin from upstream to downstream; The third processing module is used to calculate the pollution load of point sources and non-point sources by using the constructed hydrological and water quality model to obtain the water quality concentration; The fourth processing module is used to construct a prediction data set according to the water quality concentration, and use the prediction data set to train the XGBoost machine learning model; The fifth processing module is used to use the trained XGBoost machine learning model to predict the pollutant concentration during the flood season under different meteorological conditions and issue warnings for concentrations exceeding the standard, thereby completing the prediction and warning of the pollution intensity of rivers during the flood season.

Images