River basin nitrogen and phosphorus pollution prediction method based on rainfall runoff migration

Abstract

The invention discloses a drainage basin nitrogen and phosphorus pollution prediction method based on rainfall runoff migration, and relates to the technical field of water quality pollution prediction, and the method comprises the steps: accurately obtaining land utilization and water body distribution through a remote sensing technology, carrying out the space division of a pollution source through combining a nitrogen and phosphorus load coefficient, and constructing a rainfall runoff model. The runoff volume and time and space distribution under different rainfall events are reflected, and the spatial dynamic change of pollutants is captured through a hydrodynamic model, in combination with a convection diffusion equation and by introducing a conversion model of nitrogen and phosphorus in various forms. By integrating remote sensing data, rainfall runoff simulation, hydrodynamic force and nitrogen and phosphorus form transformation, the space-time migration and transformation process of nitrogen and phosphorus pollution in a drainage basin is described, so that the coupling relation between pollution source distribution and runoff power is revealed, the conveying and diffusion rule of pollutants in a river network is dynamically reflected, and the drainage effect is improved. And refining to transformation evolution of different nitrogen and phosphorus forms, and generating intuitive concentration distribution and thermodynamic diagrams through spatial interpolation.

Description

Technical Field

[0001] This invention relates to the field of water pollution prediction technology, specifically a watershed nitrogen and phosphorus pollution prediction method based on rainfall-runoff migration. Background Technology

[0002] In the process of monitoring and simulating early warning of nitrogen and phosphorus pollution in watersheds, rainfall not only triggers the formation and collection of surface runoff, but also significantly affects the forms of nitrogen and phosphorus in the watershed. Specifically, it causes nitrogen and phosphorus to transform between particulate and dissolved states. In addition, the rates of processes such as denitrification and adsorption sedimentation are also catalytically altered with changes in flow velocity. In other words, the migration characteristics of pollutants under rainfall conditions differ from those under non-rainfall conditions.

[0003] Specifically, under non-rainfall conditions, rivers flow in a near-steady-state manner, and pollutant migration mainly relies on diffusion and local flow transport. Therefore, pollutant concentration changes slowly, and spatial distribution is relatively stable. However, under rainfall conditions, nitrogen and phosphorus exist in water bodies in multiple forms, such as particulate and dissolved forms. The rainfall process has a significant impact on the proportion and conversion rate of these forms. Taking heavy rainfall as an example, the runoff caused by heavy rainfall will cause a large amount of particulate nitrogen and particulate phosphorus to enter the river, increasing the suspended particulate matter load. At the same time, dissolved nutrients carried in the runoff, such as nitrate nitrogen and phosphate phosphorus, diffuse rapidly with the water flow.

[0004] However, existing watershed water quality analysis methods, when applied to rainfall conditions, often focus only on the scouring and transport of pollution loads by rainfall runoff, neglecting the dynamic transformation of nitrogen and phosphorus speciation during rainfall and its coupling with hydrodynamic conditions. Because rainfall events cause rapid changes in flow velocity, flow rate, and water body dilution capacity, the speciation, migration rate, and pathways of nitrogen and phosphorus pollutants exhibit nonlinear characteristics over time. Therefore, ignoring the impact of rainfall on pollutant speciation and migration mechanisms can lead to excessively large errors in predicting the diffusion trends and concentration changes of nitrogen and phosphorus in watershed runoff, making it difficult to reflect the true dynamic evolution of pollutants under rainfall conditions. Summary of the Invention

[0005] 1) Technical problems to be solved This invention provides a watershed nitrogen and phosphorus pollution prediction method based on rainfall-runoff migration, which can reflect the transformation and migration trends and concentration changes of nitrogen and phosphorus pollution in the watershed under rainfall conditions.

[0006] (ii) Technical Solution To achieve the above objectives, the present invention provides the following technical solution: a watershed nitrogen and phosphorus pollution prediction method based on rainfall-runoff migration, comprising the following steps: Remote sensing imagery is used to acquire hydrological distribution information and surrounding land types of the target watershed, and nitrogen and phosphorus pollution sources within the target watershed are identified and located. The target watershed is divided into multiple spatial units according to land types, and each spatial unit is assigned a corresponding nitrogen and phosphorus load coefficient per unit area. Based on the land area of ​​each spatial unit and its nitrogen and phosphorus load coefficient per unit area, a static load distribution map of nitrogen and phosphorus pollution is established. A rainfall-runoff model is constructed by inputting the spatiotemporal distribution of rainfall in the target watershed and simulating the runoff volume, confluence path, and inflow time series of nitrogen and phosphorus pollutants generated by rainfall events. The static load distribution map is coupled with the scouring effect of rainfall to calculate the nitrogen and phosphorus flux carried by each runoff in each time step during rainfall, thus forming the inflow time series of nitrogen and phosphorus pollutants. Using the river network geometric features, cross-sectional morphology, and boundary conditions in the hydrological distribution information, the rainfall distribution is set, and the velocity field and flow field of each runoff in the river network during rainfall are calculated to establish a river network hydrodynamic model. The inflow time series of nitrogen and phosphorus pollutants into the river is used as the upstream input of the hydrodynamic model. Combined with the convection-diffusion equation, the migration process of nitrogen and phosphorus in the river network during rainfall is simulated to obtain the nitrogen and phosphorus concentration distribution at each time step. A nitrogen and phosphorus speciation transformation model is introduced. Based on the nitrogen and phosphorus concentration distribution at each time step, the interconversion and migration process between particulate and dissolved nitrogen and phosphorus under rainfall conditions, as well as the influence of rainfall velocity changes on denitrification and particle settling rate, are considered to calculate the spatiotemporal concentration distribution of different forms of nitrogen and phosphorus. Then, spatial interpolation is used to predict and generate nitrogen and phosphorus concentration distribution maps of different forms at each time step.

[0007] In feasible embodiments, remote sensing images and digital elevation maps of the target watershed are acquired, and after preprocessing them with radiometric correction, atmospheric correction and spatial geometric correction, supervised classification is used to identify the land use type of the area where the target watershed is located. The water distribution range of the target watershed is identified by calculating the spectral characteristics of the normalized water index in the remote sensing data. Based on the land type classification results and combined with the geographic information database, the location and type of nitrogen and phosphorus pollution sources are identified within the target watershed area.

[0008] In a feasible embodiment, the target watershed is divided into multiple spatial units based on the elevation and topographic features of the target watershed and the land use types identified and classified within the region; specifically, the topographic aspect, slope, and confluence path of the target watershed are calculated using the digital elevation map, and multiple sub-watersheds are initially divided based on the boundaries of the confluence paths. Within each sub-basin, regions with the same land use type and slope characteristics are defined as a spatial unit, taking into account the distribution of land use classifications and slope characteristics.

[0009] In a feasible embodiment, after the spatial unit division of the target watershed is completed, a nitrogen and phosphorus unit area load coefficient matching its land use type is assigned to each spatial unit according to the land use type corresponding to each spatial unit. Calculate the total nitrogen and phosphorus pollution load within each spatial unit by combining the actual area of ​​that unit; The total nitrogen and phosphorus pollution load calculated for each spatial unit is mapped to the space of the target watershed to form a static load distribution map of nitrogen and phosphorus pollution.

[0010] In a feasible embodiment, the spatiotemporal distribution data of rainfall during rainfall events in the target watershed are obtained; the data are input into the constructed rainfall-runoff model to simulate the flow distribution in each of the spatial units during rainfall; during the construction of the rainfall-runoff model, the runoff volume formed on the surface by rainfall and its flow process along the confluence path are calculated based on the slope and confluence direction of each spatial unit determined by the digital elevation map, so as to obtain the runoff in each spatial unit, the time delay of arrival at the river network and the confluence intensity; At each time step, the total nitrogen and phosphorus pollution load of each spatial unit in the static load distribution map of nitrogen and phosphorus pollution is coupled with the runoff corresponding to that spatial unit; wherein, the proportion of pollutants washed from the surface into the runoff is characterized by the rainfall scour coefficient, and the nitrogen and phosphorus output flux of each spatial unit at that time step is calculated. Along the confluence path, the nitrogen and phosphorus output fluxes of each spatial unit are accumulated with time delay to generate the nitrogen and phosphorus inflow time series of each river network confluence node in the target watershed.

[0011] In feasible embodiments, based on the digital elevation map and river geographic data of the target watershed, the spatial connectivity of the river network in the target watershed is extracted, the cross-sectional shape data and length information of each river segment are obtained, and a geometric model of the river network in the target watershed is constructed. Based on the rainfall-runoff model, the time series of nitrogen and phosphorus fluxes into the river corresponding to each inflow section are used as upstream input conditions, and the initial water level, flow velocity and downstream boundary water level of each river segment are set. A river network hydrodynamic model is established using numerical solution methods. Based on the shallow water equation, the velocity field and water depth distribution of the river network at each time step are calculated iteratively to form the dynamic hydrodynamic field of the target watershed. By combining the time series of nitrogen and phosphorus inflow loads into rivers calculated by the rainfall-runoff model, nitrogen and phosphorus pollutants are injected into the corresponding river sections as source terms. The migration and diffusion process of nitrogen and phosphorus in the river network under rainfall conditions is simulated using a two-dimensional convection-diffusion equation to obtain the nitrogen and phosphorus concentration distribution at each time step.

[0012] In feasible embodiments, a nitrogen and phosphorus speciation transformation model for nitrogen and phosphorus pollutants in water bodies is established based on the flow velocity, water depth, and flow distribution calculated by the hydrodynamic model of the target watershed river network. The nitrogen and phosphorus form conversion model simulates the mineralization, nitrification, and denitrification processes of nitrogen from organic to ammonia nitrogen, ammonia nitrogen to nitrate nitrogen, and nitrate nitrogen to nitrogen gas by coupling convection-diffusion equations with form conversion kinetic equations. It also simulates the adsorption, deposition, and resuspension processes of phosphorus between dissolved and particulate forms. The transport process is based on the flow rate and water volume in the hydrodynamic model to achieve convective transport and diffusion, while the conversion process is described by establishing a set of differential equations to describe the dynamic changes between the various forms, including the influence of key rate constants and environmental parameters.

[0013] In feasible embodiments, nitrogen and phosphorus concentration data for each river segment and each time step are calculated based on pollutant migration and transport models. Spatial interpolation algorithms are then used to calculate the nitrogen and phosphorus concentrations at each node of the target watershed river network, and spatial filling and continuity processing are performed.

[0014] In feasible embodiments, by interpolating the nitrogen and phosphorus concentration data at different time steps, a continuous spatial distribution surface of nitrogen and phosphorus concentration in the target watershed is generated. Combined with the time dimension, the concentration distribution surfaces at each time step are arranged in chronological order to form a time-series heatmap sequence of nitrogen and phosphorus concentration.

[0015] (iii) Beneficial effects: Compared with the prior art, this invention has the following beneficial effects: This invention utilizes remote sensing images to extract land use types, water body distribution, and topographic information. Based on the land use types along the watershed, nitrogen and phosphorus load coefficients are assigned per unit area to construct a static load distribution map, thereby realizing a quantitative expression of the potential pollution intensity in various areas of the watershed and providing a basis for subsequent dynamic transport calculations.

[0016] A rainfall-runoff model was established based on spatiotemporal rainfall distribution data to simulate the runoff, confluence paths, and inflow time delays of each spatial unit during rainfall events. The static load distribution was coupled with rainfall scouring to obtain a time-varying sequence of nitrogen and phosphorus inflow fluxes. Furthermore, a nitrogen and phosphorus speciation transformation model was introduced, comprehensively considering the impact of rainfall velocity variations on denitrification, particle deposition, and dissolution transformation. The concentration evolution of different forms of nitrogen and phosphorus in the watershed was dynamically calculated, reflecting the spatiotemporal evolution of pollutant concentrations under rainfall events. Attached Figure Description

[0017] Figure 1 A flowchart of a watershed nitrogen and phosphorus pollution prediction method based on rainfall-runoff migration provided in an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the process of obtaining a spatially interpolated concentration distribution map from remote sensing data in a watershed nitrogen and phosphorus pollution prediction method based on rainfall-runoff migration provided in an embodiment of the present invention. Figure 3 This is a schematic diagram of the process flow of each model in a watershed nitrogen and phosphorus pollution prediction method based on rainfall-runoff migration provided in an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0019] In the description of this invention, it should be understood that the terms "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0020] Furthermore, the terms "first" and "second" are used only to distinguish descriptions and should not be interpreted as indicating or implying relative importance.

[0021] It should be noted that, where there is no conflict, the features in the embodiments of the present invention can be combined with each other.

[0022] Combination Figures 1 to 3 This paper presents a watershed nitrogen and phosphorus pollution prediction method based on rainfall-runoff migration. By integrating remote sensing data, rainfall-runoff simulation, hydrodynamics, and nitrogen and phosphorus speciation transformation models, this method systematically describes the spatiotemporal migration and transformation processes of nitrogen and phosphorus pollution within the watershed. It reveals the coupling relationship between pollution source distribution and runoff dynamics, dynamically reflects the transport and diffusion patterns of pollutants in the river network, and refines the transformation and evolution of different nitrogen and phosphorus speciations. Finally, spatial interpolation generates intuitive concentration distribution and heat maps, facilitating pollution status monitoring and analysis, and providing strong technical support and decision-making basis for watershed environmental management and pollution prevention. Combination Figure 1 As shown, the first step is to acquire data on the target watershed using S10: using remote sensing imagery to obtain hydrological distribution information and the type of surrounding land in the target watershed, and to identify and locate nitrogen and phosphorus pollution sources within the target watershed; based on the land type, the target watershed is divided into multiple spatial units, and each spatial unit is assigned a corresponding nitrogen and phosphorus load coefficient per unit area.

[0023] In the above steps, the first step is to obtain information on land use types and water body distribution in the target watershed through remote sensing images, which is the basis for the entire identification and quantification of pollution sources.

[0024] Before being analyzed, remote sensing images should undergo necessary preprocessing, including radiometric calibration, atmospheric correction, geometric correction, and removal of clouds, fog, and shadows, to ensure that the spectral characteristics of the images accurately reflect the surface conditions. Subsequently, the differences in spectral response across the multispectral bands of the remote sensing images are utilized, along with auxiliary indicators such as vegetation indices and water indices. For example, NDVI is associated with vegetation and farmland, while MNDWI is associated with water body identification. Specifically, the highest NDVI (indicating abundant vegetation) is used during the growing season, and the lowest MNDWI (indicating small water volume) is used during the dry season.

[0025] By combining classification algorithms, such as support vector machines, random forests, and deep learning convolutional neural networks, a fine classification of land use types can be achieved, dividing the region into categories such as cultivated land, forest land, grassland, construction land, and water bodies, and extracting the spatial distribution range of water bodies such as rivers, lakes, and reservoirs.

[0026] In summary, this step first performs initial classification screening of remote sensing images by setting MNDWI and NDWI thresholds or water body class probability maps. Then, the acquired remote sensing images are corrected, such as by matching them with river network lines or lake / reservoir boundaries, and removing factors that are prone to misjudgment, such as shadows and dark-colored features. Finally, a volume mask (binary) is generated for subsequent source-water spatial relationship calculations.

[0027] After classification, spatial calibration is required by combining the DEM (Digital Elevation Map) of the target watershed obtained from remote sensing images with existing land use survey data to ensure that all types of data have a consistent coordinate system and resolution. Based on the above principles, a land use spatial distribution map and a water body distribution map covering the entire watershed are established, providing a foundation for subsequent spatial location of pollution sources.

[0028] After identifying land use types, the spatial distribution of nitrogen and phosphorus pollution sources within the watershed needs to be determined based on the correspondence between pollution source types and land use types. Typically, arable land is the main source of agricultural non-point source pollution, with nitrogen and phosphorus emissions influenced by fertilizer application, crop type, and farming practices. In livestock and poultry farming areas, pollution intensity is determined by the distribution and scale of farms. Urban and industrial areas correspond to non-point or point sources such as domestic sewage and industrial wastewater. To quantify the nitrogen and phosphorus output of each type of pollution source, a nitrogen and phosphorus load factor per unit area needs to be introduced. This factor is usually derived from measured watershed data or regional research findings and can be adjusted by incorporating local fertilizer application rates, livestock and poultry density, and drainage intensity to better reflect the characteristics of the watershed.

[0029] Specifically, after completing land use classification and determining the spatial distribution of nitrogen and phosphorus pollution sources, the entire target watershed is divided into multiple spatial units. In some embodiments of the present invention, this is typically done by combining a digital elevation map of the target watershed with regularized grid units.

[0030] The nitrogen and phosphorus pollution source intensities of each spatial unit in the target watershed are calculated. The core of this process is to associate different pollution source types with their nitrogen and phosphorus load coefficients per unit area. During the calculation, each spatial unit may contain multiple pollution sources, such as cultivated land, forest land, livestock and poultry breeding sites, and impermeable surfaces in urban areas. Therefore, it is necessary to decompose its area and then perform a weighted summation by combining various load coefficients.

[0031] Regarding the selection of load factors, you can refer to the measured data of the region or select the recorded typical values. For example, the nitrogen load factor of cultivated land (paddy fields) is about 15-30 kgN / ha·yr, and the phosphorus load factor is about 1-5 kgP / ha·yr; the nitrogen load of grassland is lower than that of cultivated land, generally 5-10 kgN / ha·yr; while in livestock and poultry breeding areas, it is necessary to convert the annual total amount of excrement into the nitrogen and phosphorus output intensity of the corresponding area.

[0032] After the above calculations, the source intensity of the spatial unit is mapped to the spatial grid or sub-basin to generate a nitrogen and phosphorus pollution source intensity distribution map, forming spatially explicit pollution source data, which provides source term input for subsequent simulation of coupled rainfall-runoff and hydrodynamic processes.

[0033] After acquiring and processing the corresponding data from the remote sensing images, S20 is performed: based on the land area of ​​each spatial unit and its nitrogen and phosphorus unit area load coefficients, a static load distribution map of nitrogen and phosphorus pollution is established.

[0034] In watershed nitrogen and phosphorus pollution prediction, the first step (S10) is to identify pollution sources, while the second step (S20) is to construct a static load distribution map of nitrogen and phosphorus pollution. The purpose of this step is to transform the emission capacity of different land types and pollution sources in the watershed into spatialized and quantitative data that can be used for model calculations, providing a basis for subsequent simulation of pollutant migration under rainfall-runoff conditions.

[0035] Considering that each spatial unit may correspond to a specific land use type, and that different land types have different nitrogen and phosphorus emission characteristics—for example, farmland may contain large amounts of nitrogen and phosphorus generated from fertilization, while urban surfaces may mainly contain pollutants from domestic sewage or stormwater runoff—assigning each spatial unit a nitrogen and phosphorus load coefficient matching its land type quantifies the "pollution potential" of the land type, providing basic parameters for subsequent simulations.

[0036] The total nitrogen and phosphorus load of a unit can be obtained by multiplying the load factor of each spatial unit by the area of ​​that unit. The calculation principle of this step is based on the linear superposition of emissions per unit area and land area, which converts local emission capacity into absolute total pollutant amount, ensuring that it can be distributed and accumulated in space.

[0037] Finally, the total nitrogen and phosphorus loads of each spatial unit are mapped onto the watershed space to form an overall static load distribution map. It's important to understand that "static" here refers to the initial state of the loads without considering rainfall erosion and runoff effects; it represents the baseline of the watershed's pollution potential. This distribution map does not describe the land's carrying capacity for pollutants, but rather the total potential nitrogen and phosphorus emissions from each spatial unit under natural conditions.

[0038] The static load distribution map constructed in this step clearly defines the potential spatial source locations and intensities of pollutants, serving as input conditions for subsequent simulations of rainfall runoff scouring, pollutant migration, and concentration distribution.

[0039] In summary, this step quantifies the watershed's land types and pollution emission capacity into a spatialized load map, which forms the basis and prerequisite for subsequent rainfall-runoff coupled simulations and nitrogen and phosphorus migration predictions. It provides the model with clear pollution source inputs, enabling the prediction of pollutant migration, transformation, and concentration under rainfall conditions to have a clear spatial and quantitative basis.

[0040] After constructing the static load distribution map, the impact of rainfall dynamics on the migration of nitrogen and phosphorus pollution is considered. Specifically, S30 is performed: a rainfall-runoff model is constructed, the spatiotemporal distribution of rainfall in the target watershed is input, and the runoff, confluence path, and time series of nitrogen and phosphorus pollutants entering the river generated by the rainfall event are simulated. Among them, the static load distribution map is coupled with the scouring effect of rainfall to calculate the nitrogen and phosphorus flux carried by each runoff in each time step during rainfall, forming the time series of nitrogen and phosphorus pollutants entering the river.

[0041] In the embodiments where this invention is applied, rainfall is the primary driver of nitrogen and phosphorus migration in the watershed. Rainfall not only generates surface runoff but also carries pollutants from the land into the river network through scouring. To simulate this dynamic process, it is necessary to input rainfall intensity and spatial distribution data during rainfall events so that the model can reflect rainfall conditions at different times and locations.

[0042] Regarding the acquisition of rainfall data, meteorological observation data or high-resolution rainfall radar data can be used to obtain the spatiotemporal distribution of rainfall. The rainfall-runoff model constructed in this step can employ calculation methods based on distributed hydrological models, such as the SCS-CN method, or watershed analysis methods based on digital elevation maps, to simulate the runoff generation in each spatial unit.

[0043] Specifically, before proceeding with the numerical simulation, the remote sensing and basic topographic data, along with the specific soil data, must first be mapped to the aforementioned hydrological calculation units, i.e., the spatial units defined above. After these steps, it can be understood that each spatial unit has acquired its defined area. Land use type, soil infiltration parameters, and calculated annual / event-scale source strength .

[0044] Given a single rainfall event or hourly rainfall sequence (mm / h), the runoff time series for each spatial unit is obtained by calculating from rainfall to infiltration and then to runoff according to the selected runoff generation model. ( / s or / time step).

[0045] In some embodiments of the present invention, commonly achievable flow generation calculations include event-based representations of SCS-CN or physics-based Green-Ampt equations. (usually taken) Then, the formula for calculating the production flow rate at a single time step is: ; It is important to note that consistent units and numerical stability must be ensured during calculations.

[0046] The surface runoff generated by the spatial unit is distributed to downstream river network nodes according to the confluence direction derived from the DEM. A time delay and the routing of the runoff confluence process need to be applied to each confluence path.

[0047] Regarding the routing algorithm, in some embodiments of the present invention, the routing can adopt the unit line method, that is, convolving the runoff sequence of each unit with the unit line to obtain the flow rate reaching the cross section, or it can adopt dynamic routing, the discretized form of which is used for numerical implementation.

[0048] If the transmission delay is represented by a unit line, then for a unit... To a certain section of the river The contributed traffic is: ; in, The symbol represents temporal convolution. This represents the unitline response from the unit to the river segment.

[0049] The discharge over the entire river segment within the target basin is calculated as the superimposed discharge of contributions from all upstream spatial units, using the following formula: ; in, For source traffic.

[0050] The static load distribution map obtained in the second step (S20) above provides the potential total nitrogen and phosphorus loads for each spatial unit. However, during rainfall events, not all loads immediately enter the river network. Rainfall scouring determines the proportion of pollutants washed into the runoff at each time step.

[0051] Therefore, in embodiments of the present invention, a rainfall scour factor needs to be redefined to characterize the efficiency of pollutants entering runoff from the surface. The nitrogen and phosphorus output fluxes for that time step are obtained by multiplying the static load of each spatial unit by the scour factor and combining it with the runoff volume of that unit. The calculation formula is as follows: ; in, For the first Each spatial unit at time step Output nitrogen and phosphorus fluxes To obtain the total static load of nitrogen and phosphorus within this unit, For this unit at time step The runoff volume is specifically obtained by multiplying the rainfall per unit area and the surface runoff coefficient for that unit. The rainfall scour coefficient is defined based on meteorological data and is usually taken as 0-1. It is used to characterize the proportion of pollutants washed into the runoff.

[0052] For each river network node, at each time step, the pollution fluxes of all spatial units flowing into that node are summed up. The result is the river inflow time series, which is the nitrogen and phosphorus fluxes of each river network node over time, reflecting the dynamic process of pollutants entering the river under rainfall events.

[0053] Regarding this inflow time series, it is important to understand that it not only involves the total amount but also reflects spatial heterogeneity, temporal dynamics, and the impact of rainfall erosion. It provides input for subsequent hydrodynamic models to simulate the changes in nitrogen and phosphorus concentrations in the river channel over time and space.

[0054] In embodiments of the present invention, it is considered that during rainfall events, pollutants entering the river network do not immediately distribute uniformly, but rather migrate, diffuse, and accumulate along the river channel depending on changes in river velocity, water depth, and flow rate. Therefore, to accurately simulate the dynamic distribution of nitrogen and phosphorus pollutants in the river network, it is necessary to establish a model that can describe the hydrodynamic field and pollutant transport.

[0055] Specifically, step S40 involves: utilizing the river network's geometric features, cross-sectional morphology, and boundary conditions from hydrological distribution information, setting the rainfall distribution, calculating the velocity and flow fields of each runoff in the river network during rainfall, and establishing a river network hydrodynamic model; using the aforementioned inflow time series of nitrogen and phosphorus pollutants as the upstream input to the hydrodynamic model, and combining it with convection-diffusion equations, simulating the migration process of nitrogen and phosphorus in the river network during rainfall, and obtaining the nitrogen and phosphorus concentration distribution at each time step. The goal of this step is to couple the dynamic input (inflow time series) formed by pollutants (nitrogen and phosphorus) entering the river under rainfall scouring with the hydrodynamic conditions of the river network, thereby simulating the migration behavior of pollutants in the flowing river channel and obtaining the nitrogen and phosphorus concentration distribution of each river segment at each time step.

[0056] In this step, a hydrodynamic model is first constructed based on the geometric characteristics, cross-sectional features, and boundary conditions of the target watershed river network. The river network geometry includes the length, width, slope, and cross-sectional shape of the river segments; these parameters are obtained through field measurements, remote sensing data processing, or existing hydrological geographic information systems. The boundary conditions encompass inflow, outflow, and water level, which are determined in this embodiment of the invention based on the results of the previous runoff simulation.

[0057] The data on the changes in nitrogen and phosphorus fluxes into the river over time for each river segment calculated in the previous steps constitute the river inflow time series. This series reflects the dynamic characteristics of pollutants contributed by different spatial units flowing into the river channel at different times under rainfall events. In this simulation step, this river inflow time series is input into the hydrodynamic model as a pollution source from the upstream or confluence node to achieve dynamic coupling between pollutants and flow.

[0058] Regarding the hydrodynamic model of the river network, the core of this model is based on the fundamental hydrodynamic equations, namely the shallow water equations (Saint-Venant equations), to calculate the instantaneous flow velocity, water depth, and discharge of the river segment. These equations are solved using numerical iterative methods, such as the finite difference or finite volume methods, to obtain the hydrodynamic field of the river network at each time step. Since the shallow water equations are general calculation formulas, no specific limitations are imposed on the calculations here.

[0059] Treating nitrogen and phosphorus pollutants as dissolved or particulate matter, in some embodiments of the present invention, a two-dimensional convection-diffusion equation is used to describe their migration and diffusion behavior in the river channel. The equation is as follows: + ; in, C Nitrogen and phosphorus concentrations and They are respectively and In terms of direction, the flow velocity calculated by the hydrodynamic model, and These are the horizontal diffusion coefficients in the corresponding directions. For pollutant source terms, the corresponding injection amount in the river section during the river inflow time series.

[0060] Based on the above diffusion equation, for each time step, the water depth and velocity field of each river segment are calculated using a hydrodynamic model. The inflow time series is then injected into the river segment at the corresponding nodes according to time, serving as the pollution source term. The convection-diffusion equation was used to simulate the migration, diffusion, and deposition of pollutants along the river channel, obtaining the nitrogen and phosphorus concentrations in each grid cell of the river section. Finally, the time steps of the entire rainfall event were iterated to form the spatiotemporal dynamic concentration distribution of the river network.

[0061] In summary, it can be understood that the river inflow time series used as input for this step reflects the temporal variation of pollutant input during rainfall events, rather than a simple average or total input. By combining the river inflow time series with the hydrodynamic field, this step can accurately simulate the propagation paths and concentration changes of pollutants in the river network.

[0062] In summary, this hydrodynamic and transport coupling step transforms the static land pollution load and rainfall scouring effect into a dynamic input of pollutants in the river network. Combined with hydrodynamic conditions, it obtains the actual spatiotemporal concentration distribution through convection-diffusion simulation. This step, which converts "inflow into the river" into "river network concentration," is the most crucial step in the entire nitrogen and phosphorus prediction process under rainfall conditions.

[0063] Addressing the initial issue, under rainfall conditions, the behavior of nitrogen and phosphorus pollutants in watersheds and river networks depends not only on hydrodynamic transport but also on transformations through chemical and biological processes. Different forms of nitrogen and phosphorus (such as ammonia nitrogen, nitrate nitrogen, organic nitrogen, particulate phosphorus, and dissolved phosphorus) exhibit varying rates of migration, sedimentation, adsorption, mineralization, and denitrification in water bodies. Therefore, simple hydrodynamic and transport models can only simulate the total migration of pollutants and cannot reflect the actual ecological water quality impact. In the embodiments applied in this invention, to predict the spatiotemporal distribution of different forms of nitrogen and phosphorus in watersheds under rainfall conditions, it is necessary to consider the coupling of form transformation dynamics and hydrodynamic processes.

[0064] It is important to understand that the simulation steps that consider nitrogen and phosphorus speciation are based on the existing nitrogen and phosphorus concentration distributions at each time step (obtained by hydrodynamic model simulation of inflow flux). Then, the speciation process of elements or combinations is introduced to calculate the spatial distribution of different forms of nitrogen and phosphorus in the river network over time, providing a basis for prediction and visualization.

[0065] Specifically, S50 is performed by introducing a nitrogen and phosphorus speciation transformation model. Based on the nitrogen and phosphorus concentration distribution at each time step, the model considers the interconversion and migration process between particulate and dissolved nitrogen and phosphorus under rainfall conditions, as well as the influence of rainfall velocity changes on denitrification and particle settling rate, to calculate the spatiotemporal concentration distribution of different forms of nitrogen and phosphorus. Then, spatial interpolation is used to predict and generate different forms of nitrogen and phosphorus concentration distribution maps at each time step.

[0066] The hydrodynamic simulation input for this step comes from the river network hydrodynamic model from the previous step, providing the velocity, flow rate, and depth distribution for each river segment at each time step. Understandably, these parameters are used for convective transport and diffusion calculations, enabling the morphological transformation model to accurately reflect the spatial effects of pollutant migration with the water flow.

[0067] The initial concentration of pollutants simulated in this step comes from the total nitrogen and phosphorus concentration distribution output by the hydrodynamic model in the previous step. This is used as the initial condition to define the pollutant concentration of each river segment at each time step. These concentrations are the input basis for the morphological transformation model.

[0068] The nitrogen transformation process involves the conversion of organic nitrogen to ammonia nitrogen (mineralization), ammonia nitrogen to nitrate nitrogen (nitrification), and then from nitrate nitrogen to nitrogen gas (denitrification), in a multi-stage cycle. In specific embodiments, the model needs to establish differential equations for the concentration changes of each form, expressing the interconversion between forms. That is, each process is described by a differential equation to describe the conversion rate, in the form of: ; in, This represents the concentration of nitrogen in its current form at the current time. This is the conversion rate constant. Considering that it is affected by factors such as temperature, dissolved oxygen, and flow rate, its value is specifically set according to the environmental factors that the calculation is adapted to, and no specific limitation is made here. This is the environmental factor correction coefficient, which is also specifically set according to the specific environmental factors, and is not specifically limited here.

[0069] The phosphorus conversion process typically involves the interconversion between dissolved and particulate phosphorus, which follows the law of conservation of mass. In the embodiments of this invention, the phosphorus conversion process considers the effects of sedimentation rate, flow velocity, particle concentration, and adsorption-desorption kinetics. The differential equation can be found in the following form: ; ; in, and These are coefficients set for flow velocity, water depth, and particle characteristics, and are generally based on measured or empirical coefficients. This represents the concentration of particulate phosphorus at the current time. The concentration of dissolved phosphorus at the current time is given. It is particularly important to note that the concentrations of particulate and dissolved phosphorus are not directly measured at each time step, but rather are dynamic state variables obtained through numerical simulation iterations. In some embodiments of this invention, initial concentrations of particulate and dissolved phosphorus are set for each river segment or grid cell before the simulation begins. These initial concentrations of their respective forms can be derived from measured water quality data, i.e., the results of water sample testing before the event can be used to proportionally allocate dissolved and particulate phosphorus. Alternatively, the initial proportions can be allocated based on the total amount of nitrogen and phosphorus entering the river and literature or empirical coefficients.

[0070] Based on the hydrodynamic model in the aforementioned steps, the concentrations of nitrogen and phosphorus diffuse laterally and longitudinally within the watershed are calculated. At the same time, the differential equations for the form transformation of nitrogen and phosphorus are calculated, and finally the concentrations of nitrogen and phosphorus in different forms at each time step and in each unit are obtained, forming a spatiotemporal distribution matrix for the entire rainfall event.

[0071] Furthermore, in other embodiments of the present invention, considering that the spatial distribution of the watershed simulation may be discontinuous due to the limited number of monitoring or simulation points, it is impossible to directly reflect the concentration field distribution within the entire watershed.

[0072] Therefore, in order to achieve the transformation from discrete points to a continuous surface, Kriging interpolation is commonly used in some embodiments of the present invention to utilize the spatial autocorrelation characteristics of concentration data. Specifically, the spatial variation patterns between data points are analyzed using a semi-variogram function, thereby predicting concentration values ​​at unknown locations.

[0073] During interpolation, spatial weight coefficients for each spatial unit can be set by determining its spatial importance or flow rate, and then interpolation calculations can be performed. This method not only considers distance factors but also reflects the spatial structure of concentration data, improving prediction accuracy.

[0074] For each time step, the concentrations of all sampling points (spatial cells), i.e., the nitrogen and phosphorus concentrations within that spatial cell calculated in the simulation above, are interpolated to form a two-dimensional concentration distribution map by calculating the nitrogen and phosphorus concentrations corresponding to the grid cells. It can be understood that these concentration maps reflect the spatial distribution pattern and concentration gradient of pollutants.

[0075] By organizing and visualizing the concentration map sequences of continuous time steps, a series of nitrogen and phosphorus pollution heat maps are generated, thereby dynamically displaying the diffusion paths, accumulation areas, and reduction trends of pollutants over time.

[0076] Furthermore, in some embodiments of the present invention, the heat map can express the concentration by varying the color intensity, so as to intuitively reflect the pollution hotspots and their spatiotemporal evolution.

[0077] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. The scope of patent protection of the present invention shall be determined by the claims. Similarly, any equivalent structural changes made based on the description and drawings of the present invention shall also be included within the scope of protection of the present invention.

Claims

1. A method for predicting nitrogen and phosphorus pollution in a watershed based on rainfall runoff migration, characterized in that, The method comprises the following steps: obtaining hydrological distribution information and types of surrounding land of a target basin by using remote sensing image technology, identifying and locating nitrogen and phosphorus pollution sources in the target basin; dividing the target basin into a plurality of spatial units according to the types of land, and assigning a nitrogen and phosphorus unit area load coefficient corresponding to the land to each spatial unit; based on the land area of each spatial unit and the nitrogen and phosphorus unit area load coefficient thereof, a static load distribution map of nitrogen and phosphorus pollution is established; a rainfall runoff model is constructed, and the spatio-temporal distribution of rainfall in the target basin is input to simulate the runoff, confluence path and time series of nitrogen and phosphorus pollutants entering a river generated by a rainfall event; wherein the static load distribution map is coupled with rainfall scouring to calculate the nitrogen and phosphorus flux carried by each runoff at each time step during rainfall, thereby forming the time series of nitrogen and phosphorus pollutants entering the river; by using the river network geometric characteristics, cross-section shape and boundary conditions in the hydrological distribution information, setting rainfall distribution, calculating the flow velocity field and flow field of each runoff in the river network during rainfall, and establishing a river network hydrodynamic model; the time series of nitrogen and phosphorus pollutants entering the river is taken as the upstream input of the hydrodynamic model, and the migration process of nitrogen and phosphorus in the river network during rainfall is simulated by combining the convection-diffusion equation, thereby obtaining the nitrogen and phosphorus concentration distribution at each time step; a nitrogen and phosphorus form conversion model is introduced, and on the basis of the nitrogen and phosphorus concentration distribution at each time step, the mutual transformation and migration process between the particulate state and the dissolved state of nitrogen and phosphorus under rainfall conditions, and the influence of rainfall flow velocity variation on denitrification and particle settling rate are considered, thereby calculating the spatio-temporal concentration distribution of different forms of nitrogen and phosphorus; then a spatial interpolation method is used to predict and generate the nitrogen and phosphorus concentration distribution map of different forms at each time step. 2.The method of claim 1, wherein, After pre-processing of radiation correction, atmospheric correction and spatial geometric correction on the remote sensing image and digital elevation map of the target basin, the land use type of the region where the target basin is located is identified by using supervised classification, and the water body distribution range of the target basin is identified by calculating the spectral characteristics of the normalized water body index in the remote sensing data; based on the classification result of the land type, in combination with the geographic information database, the position and type of the nitrogen and phosphorus pollution source in the target basin region are located and identified. 3.The method of claim 2, wherein, According to the elevation topographic features of the target basin and the land use types identified and classified in the region, the target basin is divided into a plurality of spatial units; specifically, the terrain slope direction, slope and confluence path of the target basin are calculated by using the digital elevation map, and a plurality of sub-basins are preliminarily divided according to the boundaries of the confluence path; in each of the sub-basins, in combination with the distribution of land use classification and the slope and slope features, the regions with the same land use type and slope and slope features are defined as a spatial unit.

4. The method according to claim 3, wherein, After completing the spatial unit division of the target basin, a nitrogen and phosphorus unit area load coefficient matched with the land use type of each spatial unit is assigned to each spatial unit according to the land use type corresponding to each spatial unit; in combination with the actual area of each spatial unit, the total load of nitrogen and phosphorus pollution in the spatial unit is calculated; The total load of nitrogen and phosphorus pollution in each spatial unit is mapped to the space of the target watershed to form a static load distribution map of nitrogen and phosphorus pollution.

5. The method of claim 3, wherein the method is characterized by, Obtain the spatio-temporal distribution data of rainfall during the rainfall event of the target watershed; input the constructed rainfall runoff model to simulate the flow distribution of each spatial unit during rainfall; in the process of constructing the rainfall runoff model, according to the slope and confluence direction of each spatial unit determined according to the digital elevation map, calculate the runoff formed on the ground surface by rainfall and its flow process along the confluence path to obtain the runoff in each spatial unit, the time delay and confluence intensity of the runoff reaching the river network; At each time step, the total load of nitrogen and phosphorus pollution in each spatial unit in the static load distribution map of nitrogen and phosphorus pollution is coupled with the corresponding runoff of the spatial unit; wherein the proportion of pollutants washed into runoff from the ground surface is represented by a rainfall scouring coefficient, and the nitrogen and phosphorus output flux of each spatial unit at the time step is calculated. Along the confluence path, the nitrogen and phosphorus output flux of each spatial unit is accumulated according to the time delay to generate a nitrogen and phosphorus inflow time series of each river network confluence node in the river network of the target watershed.

6. The method of claim 3, wherein the method is characterized by, Based on the digital elevation map and river channel geographic data of the target watershed, the spatial connectivity of the river network of the target watershed is extracted, the cross-sectional shape data and length information of each river section are obtained, and a river network geometric model of the target watershed is constructed; Based on the rainfall runoff model, the nitrogen and phosphorus inflow flux time series of each inflow cross section is taken as the upstream input condition, and the initial water level, flow velocity and downstream boundary water level of each river section are set. A numerical solution method is used to establish a river network hydrodynamic model, and based on the shallow water equation, the flow velocity field and water depth distribution of the river network at each time step are iteratively calculated to form a dynamic hydrodynamic field of the target watershed. Based on the nitrogen and phosphorus inflow load time series calculated by the rainfall runoff model, nitrogen and phosphorus pollutants are injected into the corresponding river section position as a source term, and the two-dimensional convection-diffusion equation is used to simulate the migration and diffusion process of nitrogen and phosphorus in the river network under rainfall conditions to obtain the nitrogen and phosphorus concentration distribution at each time step.

7. The method according to claim 6, wherein, Based on the flow velocity, water depth and flow distribution calculated by the hydrodynamic model of the river network of the target watershed, a nitrogen and phosphorus form transformation model of nitrogen and phosphorus pollutants in water is established; The nitrogen and phosphorus form transformation model couples the convection-diffusion equation with the form transformation kinetics equation to simulate the mineralization, nitrification and denitrification processes of organic nitrogen to ammonia nitrogen, ammonia nitrogen to nitrate nitrogen and nitrate nitrogen to nitrogen gas, as well as the adsorption, deposition and resuspension processes between dissolved phosphorus and particulate phosphorus; wherein the flow transport and diffusion are realized based on the flow velocity and water volume in the hydrodynamic model, and the dynamic changes between forms are described by establishing a system of differential equations, including key rate constants and environmental parameter influences. 8.The method of claim 1, wherein, Based on the nitrogen and phosphorus concentration data of each river section and each time step calculated by the pollutant migration and transport model, a spatial interpolation algorithm is used to calculate the nitrogen and phosphorus concentration of each node of the river network of the target watershed, and spatial filling and continuous processing are performed. 9.The method of claim 8, wherein, The continuous spatial distribution surfaces of the nitrogen and phosphorus concentrations in the target basin are generated by respectively interpolating the nitrogen and phosphorus concentration data at different time steps, and the concentration distribution surfaces at different time steps are arranged in time sequence to form a time sequence thermodynamic map sequence of the nitrogen and phosphorus concentrations by combining the time dimension.

Images