Calibration and optimization methods of spatiotemporal dynamic parameters of water environment numerical models

Abstract

The present invention discloses a method for calibrating and optimizing the spatiotemporal dynamic parameters of a water environment numerical model. The method comprises the following steps: S1: constructing a water environment numerical model for a target area; S2: determining key dynamic parameters; S3: dividing the target area into multiple spatial ranges based on environmental characteristics and parameter adjustment efficiency; S4: defining parameter value equations and iteratively modifying sequence values ​​of different parameters with different spatial distributions; S5: performing parameter adjustment until the simulated values ​​and the measured values ​​are close to each other, i.e., obtaining a parameter adjustment result with a good fit; and S6: stopping parameter adjustment to prevent overfitting when the simulated values ​​and the measured values ​​are close and the fit evaluation index between the simulated values ​​and the measured values ​​is no longer significantly improved, and determining the parameters at this point as the final optimized parameters. The present invention improves the operational efficiency of the parameter calibration and optimization model and simplifies the model structure.

Description

Technical Field

[0001] The present invention belongs to the technical field of computer data simulation and data tuning and processing, and in particular relates to a parameter tuning and parameter optimization method for a water environment numerical model. Background Art

[0002] Numerical simulation uses mathematical equations to depict the real world. With the rapid development of digital twins, the accuracy of water environment numerical simulation results has become crucial. The complex mechanisms and numerous state variables and parameters of water environment numerical models make accurate parameter calibration extremely difficult. Therefore, there is an urgent need for an accurate, efficient, and easy-to-implement method to optimize water environment numerical model parameters and improve simulation accuracy.

[0003] When calibrating numerical water environment models, it's often assumed that these parameters remain constant across the entire region and throughout the simulation period to simplify the calibration process. However, the parameters of various material circulation processes in real water bodies vary both temporally and spatially, and fixed parameters cannot adequately characterize these processes, often resulting in poor accuracy in numerical water environment models.

[0004] Time-varying parameters assume that parameters vary over time, while partitioned parameters assume that parameters vary over space. These two methods improve model accuracy to a certain extent and are relatively easy to implement, but they only account for variations in one aspect of time or space. This falls short when more refined spatiotemporal dynamic parameters are required to further improve model accuracy. The application of spatiotemporal dynamic parameters also presents some difficulties: first, the complex process of creating spatiotemporal dynamic parameter files significantly reduces the efficiency of model parameter calibration; second, spatiotemporal dynamic parameter files occupy a large amount of memory, making the model bloated and reducing computational efficiency. Spatiotemporal dynamics refers to the characteristic parameter variations across time and space within a given range.

[0005] like Figure 1 As shown in , usually the parameters are fixed at all times in the entire target area and the entire simulation period. The target area is divided according to the changes or characteristics of the terrain. During the simulation period, the model is calculated step by step according to the step size. For example, if the simulation period is 3 hours, the step size is set to 1 hour. Figure 1 As shown in (a), the entire target area 10 and the multiple grids 20 (a, b, c, d, e, f, g, hi, j, k) corresponding to the minimum calculation unit of the water environment numerical model divided in the target area. Figure 1As shown in (b), a spatial range corresponds to grid values ​​of two parameter types: fixed parameters and time-varying parameters. Assume that the parameter value is x. In the case of fixed parameters, the value of each grid at different times, including time 1, time 2, and time 3, is a fixed value x. In the case of time-varying parameters, the value of each grid at different times, including time 1, time 2, and time 3, is x, y, and z, that is, the value of each grid changes over time. Figure 1 As shown in (c), the grid values ​​of the partition parameters corresponding to multiple spatial ranges are as follows: the value of each grid in range 1 at different times 1, 2, and 3 is a fixed value x; the value of each grid in range 2 at different times 1, 2, and 3 is a fixed value y. Figure 1 As shown in (d), for time-varying parameters within two spatial ranges, each grid within spatial range 1 takes on a value of x at different times, including time 1, y at time 2, and z at time 3; each grid within spatial range 2 takes on a value of x' at different times, including time 1, y' at time 2, and z' at time 3. As shown in (e), the file containing the parameter values ​​for the entire region at all times is called the spatiotemporal dynamic file. Specifically, at time 1, the grid values ​​for spatial range 1 take on a value of x, and the grid values ​​for spatial range 2 take on a value of x'; at time 2, the grid values ​​for spatial range 1 take on a value of y, and the grid values ​​for spatial range 2 take on a value of y'; at time 3, the grid values ​​for spatial range 1 take on a value of z, and the grid values ​​for spatial range 2 take on a value of z'. The grid values ​​for the entire region at time 1, time 2, and time 3 together constitute the spatiotemporal dynamic parameter file. The tuning process for each spatiotemporal dynamic parameter involves first adjusting the sequence values, repeatedly creating spatiotemporal dynamic files, and then calling these files when running the model. This parameter tuning method is very time-consuming.

[0006] In summary, the current water environment numerical model parameter calibration method cannot meet the requirements of high-precision and high-efficiency calibration model, and a new technical solution needs to be proposed to improve the above technical problems. Summary of the Invention

[0007] In order to solve the above problems, the present invention proposes a method for calibrating and optimizing the spatiotemporal dynamic parameters of a water environment numerical model. The parameter space partition file is used to associate the key dynamic parameters of the time variable with sequence values. The sequence values ​​associated with the key dynamic parameters are modified to achieve convenient parameter adjustment, providing an effective adjustment solution for parameter optimization.

[0008] The present invention is achieved by adopting the following technical solutions:

[0009] A method for calibrating spatiotemporal dynamic parameters of a water environment numerical model of the present invention is characterized by comprising:

[0010] S1: Construct a numerical model of the water environment in the target area, i.e., obtain the topography, meteorology, water quality and flow according to the characteristics of the target area as the first part, and obtain the state variables of the water environment simulation, the mechanism process and parameter composition of each state variable according to the research purpose as the second part, and construct the initial numerical model of the water environment in the target area from the first and second parts;

[0011] S2: Determine the parameters of the state variable parameters that are highly correlated with the sensitivity analysis or trial operation result analysis of the target water environment as key dynamic parameters;

[0012] S3: Divide the target area into multiple spatial ranges according to environmental characteristics and parameter adjustment efficiency indicators. Each spatial range is composed of multiple grids. The grid is the minimum calculation unit of the water environment numerical model. The grid attribute value is set to correspond to the sequence value of the spatial range.

[0013] S4: Define parameter value equations: associate the sequence values ​​of different key dynamic parameters in different spatial ranges, including determining the spatial range n to which the grid belongs, making a parameter space partitioning file, and assigning sequence values ​​to the spatial ranges to which all grids belong; add a new partitioning instruction function zoning in the water environment numerical model to realize the key dynamic parameter calibration process in the target area, that is, for the i-th key dynamic parameter Ri, iteratively modify the sequence values ​​of different key dynamic parameters in different spatial ranges. When the partitioning instruction function Zoning is equal to 1, the value of Ri at each moment is expressed as the sequence value Ri_1. Otherwise, the sequence value is modified from the current sequence value Ri_1 to the sequence values ​​of other spatial ranges Ri_2 to Ri_n, and parameter adjustment is achieved by modifying the sequence value.

[0014] S5: Iteratively modify the sequence values ​​of different key dynamic parameters in different spatial ranges until the simulated values ​​are close to the measured values, and the parameter adjustment results are obtained.

[0015] In some embodiments, the iterative modification in S5 of the key dynamic parameters according to the degree of agreement between the simulated value and the measured value is performed according to the parameter value equation, and the sequence values ​​associated with different spatial ranges are dynamically modified according to the attribute value through the parameter space division instruction.

[0016] A method for optimizing spatiotemporal dynamic parameters of a water environment numerical model according to the present invention is characterized by comprising:

[0017] S1: Construct a numerical model of the water environment in the target area, i.e., obtain the topography, meteorology, water quality and flow according to the characteristics of the target area as the first part, and obtain the state variables of the water environment simulation, the mechanism process and parameter composition of each state variable according to the research purpose as the second part, and construct the initial numerical model of the water environment in the target area from the first and second parts;

[0018] S2: Determine the parameters of the state variable parameters that are highly correlated with the sensitivity analysis or trial operation result analysis of the target water environment as key dynamic parameters;

[0019] S3: Divide the target area into multiple spatial ranges according to environmental characteristics and parameter adjustment efficiency indicators. Each spatial range is composed of multiple grids. The grid is the minimum calculation unit of the water environment numerical model. The grid attribute value is set to correspond to the sequence value of the spatial range.

[0020] S4: Define parameter value equations: associate the sequence values ​​of different key dynamic parameters in different spatial ranges, including determining the spatial range to which the grid belongs, making a parameter space partitioning file, and assigning sequence values ​​to the spatial range to which all grids belong; add a new partitioning instruction function zoning in the water environment numerical model to realize the key dynamic parameter calibration process in the target area, that is, for the i-th key dynamic parameter Ri, iteratively modify the sequence values ​​of different key dynamic parameters in different spatial ranges. When the partitioning instruction function Zoning is equal to 1, the value of Ri at each moment is expressed as the sequence value Ri_1. Otherwise, the sequence value is modified from the current sequence value Ri_1 to the sequence value of other spatial ranges Ri_2 to Ri_n, and the parameter adjustment is achieved by modifying the sequence value;

[0021] S5: Iteratively modify the sequence values ​​of different key dynamic parameters in different spatial ranges until the simulated values ​​and the measured values ​​are close to each other, that is, a parameter adjustment result with a better fit is obtained;

[0022] S6: When the simulated value is close to the measured value and the evaluation index of the simulated value and the measured value is no longer significantly improved, the parameter adjustment is stopped to prevent overfitting, and the parameters at this time are determined as the final optimization parameters.

[0023] In some embodiments, the iterative modification in S5 of the key dynamic parameters according to the degree of agreement between the simulated value and the measured value is performed according to the parameter value equation, and the sequence values ​​associated with different spatial ranges are dynamically modified according to the attribute value through the parameter space division instruction.

[0024] In some embodiments, the parameters pre-adjusted to a point where the simulated value and the measured value are at the same level are used as the initial parameters.

[0025] Compared with the prior art, the present invention has the following significant advantages:

[0026] 1) Taking into account the characteristics of water environment numerical model parameters and the requirements for model accuracy, the accuracy of the model is greatly improved by dynamically optimizing the key parameters in time and space;

[0027] 2) By defining parameter spatial changes and parameter value equations, the parameter value equations are used to dynamically modify the sequence values ​​of key dynamic parameters in different spatial ranges according to the degree of fit between the simulated values ​​and the measured values, and key dynamic parameters are adjusted and optimized. This can solve the changes in one aspect of time or space and seek more refined spatiotemporal dynamic parameter adjustment and optimization, without the need to repeatedly create spatiotemporal dynamic files, thereby simplifying the parameter spatiotemporal dynamic optimization process, greatly improving the parameter adjustment efficiency, and better ensuring the parameter adjustment accuracy.

[0028] 3) Through the design and process optimization of the present invention, only relatively easy-to-produce sequence files need to be produced, and there is no need to produce spatiotemporal dynamic parameter files that occupy a large amount of memory, which simplifies the model structure and improves the efficiency of model parameter calibration and optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is an example diagram of the numerical parameter adjustment process of the existing technology;

[0030] Figure 2 This is an overall flow chart of the spatiotemporal dynamic parameter calibration and optimization method of the water environment numerical model of the present invention;

[0031] Figure 3 Schematic diagram of the temporal and spatial dynamic parameter calibration process of the water environment numerical model of the present invention;

[0032] Figure 4 The technical roadmap of the water environment numerical model spatiotemporal dynamic parameter adjustment and optimization method of the present invention;

[0033] Figure 5 A spatial division diagram of underwater terrain parameters of a certain region according to an embodiment of the present invention;

[0034] Figure 6 This is a mapping relationship diagram between spatial range sequences and key dynamic parameters of the present invention, where i represents the total number of key dynamic parameters and n represents the total number of spatial range sequences;

[0035] Figure 7 A comparison diagram of simulated values ​​and actual values ​​of measured points at a rate period according to an embodiment of the present invention;

[0036] Figure 8 Comparison chart between simulated values ​​and measured values ​​of measured points during the verification period of an embodiment of the present invention. DETAILED DESCRIPTION

[0037] The present invention will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0038] like Figure 2 As shown, a method for calibrating spatiotemporal dynamic parameters of a water environment numerical model of the present invention specifically includes the following steps:

[0039] Step S1: Construct a numerical model of the water environment of the target area. The first part collects data such as topography, meteorology, water quality and flow according to the characteristics of the target area, and the second part determines the state variables of the water environment simulation according to the research purpose, the mechanism process (including physical process, chemical process, biological process) and parameters of each state variable. The first part and the second part are used to construct an initial numerical model of the water environment of the target area; a specific description is given in conjunction with the underwater topography of a certain area in the embodiment: for example, based on the collected underwater topography, water quality and flow of rivers entering the lake and meteorological data of the area and according to the research purpose, ammonia nitrogen NH4, nitrate nitrogen NO3, orthophosphate PO4, total nitrogen TN, total phosphorus TP and algae biomass ALGAE are determined as state variables of the water environment simulation; considering BOD degradation and release, nitrification and denitrification, plant absorption, bacterial absorption, sediment release and algae growth, death and deposition, etc., 33 parameters of each state variable of the water environment simulation are involved, as shown in Table 1 as a parameter list, including a complete parameter list and key dynamic parameters; the expression of the mechanism process of each state variable of the present embodiment is as follows:

[0040] The process expression of oxygen demand BOD degradation to release ammonia nitrogen NH4 and orthophosphate PO4 is as follows:

[0041]

[0042] Where: BOD is the biochemical oxygen demand for 5 days, DO is the dissolved oxygen concentration, HS is the half-saturation concentration of oxygen, y X is the release rate of NH4 / PO4 in oxygen demand BOD degradation, k1 is the oxygen demand BOD degradation coefficient, θ1 is the oxygen demand BOD degradation temperature coefficient, T is the water temperature, decayX is the BOD degradation and release process of ammonia nitrogen NH4 and orthophosphate PO4, and X is used to summarize the above two processes.

[0043] The process expressions of nitrification and denitrification are as follows:

[0044]

[0045] Where: Nitrification is the nitrification process, Denitrification is the denitrification process, NH4 is the ammonia nitrogen concentration, NO3 is the nitrate nitrogen concentration, k2 is the nitrification reaction rate, θ2 is the nitrification temperature coefficient, k3 is the denitrification rate, and θ3 is the denitrification temperature coefficient.

[0046] The process expression of plant absorption of ammonia nitrogen NH4 and orthophosphate PO4 is as follows:

[0047]

[0048] in:

[0049] PSN=P max / H·F(P)

[0050] RESP=R s / H·F(R)

[0051]

[0052] Where: PU x is the rate of NH4 / PO4 absorption by plants, P max is the maximum oxygen production rate at noon, H is the water depth, PSN is photosynthesis, F(P) is the photosynthesis limiting function, assuming that the photosynthesis intensity during the day is a cosine function, θ4 is the photosynthesis temperature coefficient, τ is the actual time relative to noon, obtained by latitude calculation, and f i Correction, α is the actual relative day length, obtained by latitude calculation, RESP is respiration, R s is the respiratory rate, F(R) is the respiratory limiting function, θ5 is the respiration temperature coefficient, F(N) is the nutrient limitation function, IN is the inorganic nitrogen concentration, IN = NH4 + NO3, PO4 is the orthophosphate concentration, KSN is the half-saturation concentration of nitrogen, KSP is the half-saturation concentration of phosphorus. The difference between the expressions of plant absorption of ammonia nitrogen and orthophosphate is only the plant absorption rate PUx, so PlantUptakenX is used to express the two processes of plant absorption of ammonia nitrogen and orthophosphate, PU x It represents two parameters of ammonia nitrogen and orthophosphate absorbed by plants.

[0053] The process expression of bacteria absorbing NH4 and PO4 is as follows:

[0054]

[0055] Where: BU x It is the rate at which bacteria absorb NH4 / PO4. The rest of the meanings are the same as above.

[0056] The process expression of sediment releasing NH4, NO3 and PO4 is as follows:

[0057] SediRelease X =R x / H (6)

[0058] Where: R x It is the rate of NH4 / NO3 / PO4 release from sediment, and the rest of the meanings are the same as above;

[0059] The process of algae growth, death and deposition is expressed as follows:

[0060] ProdALGAE=(k4-k5)·F(P)·F(N)·ALGAE (7)

[0061] DeathALGAE=k6·ALGAE (8)

[0062] SediALGAE=k7 / H·ALGAE (9)

[0063] Where: ALGAE is the algae biomass, k4 is the algae proliferation rate, k5 is the algae basal metabolic rate, k6 is the algae mortality rate, k7 is the algae deposition rate, and the rest have the same meanings as above.

[0064] Total nitrogen TN and total phosphorus TP are determined by measuring the ratio of inorganic nitrogen to total nitrogen K N and the ratio of orthophosphate to total phosphorus K P Calculate inversely;

[0065] The initial values ​​of the state variable parameters of the water environment simulation in step S1 are obtained by, for example, reference literature, model manuals, or parameter pre-adjustment. Specifically, the parameters are pre-adjusted to the point where the simulated values ​​and the measured values ​​are at the same level as the initial parameters;

[0066] Step S2: Determine the parameters with high correlation with the sensitivity analysis or trial operation result analysis of the target water environment among the state variable parameters as key dynamic parameters; to avoid the difficulty of parameter adjustment caused by dynamic optimization of too many parameters, which increases the risk of different parameters having the same effect; the key dynamic parameters are time dynamic parameters. Through the trial operation, this embodiment determines 8 key dynamic parameters, as shown in Table 1, which is a parameter list including a complete parameter list and key dynamic parameters;

[0067] Step S3: Divide the target area into multiple spatial ranges according to environmental characteristics and parameter adjustment efficiency, such as Figure 3 As shown, for example, the target area is divided into spatial range sequence 1 and spatial range sequence 2, wherein the attribute value of all grids in spatial range sequence 1 is defined as 1, and the attribute value of all grids in spatial range sequence 2 is defined as 2;

[0068] Step S4: Define the parameter value equation: associate the sequence values ​​of different key dynamic parameters in different spatial ranges, including determining the spatial range to which the grid belongs, making a parameter space partitioning file, and assigning sequence values ​​to the spatial range to which all grids belong; add a new partitioning instruction function zoning in the water environment numerical model to realize the key dynamic parameter calibration process in the target area, that is, for the i-th key dynamic parameter Ri, iteratively modify the sequence values ​​of different key dynamic parameters in different spatial ranges. When the partitioning instruction function Zoning is equal to 1, the value of Ri at each moment is expressed as the sequence value Ri_1. Otherwise, the sequence value is modified from the current sequence value Ri_1 to the sequence value of other spatial ranges Ri_2 to Ri_n. Parameter adjustment is achieved by modifying the sequence value. Here, parameter adjustment for spatial range 1 and spatial range 2 is used as an example: Therefore, only sequences such as Ri_1 and Ri_2 need to be modified to achieve parameter adjustment, which greatly improves the parameter adjustment efficiency. The expression of the parameter adjustment process is as follows:

[0069] Ri(t)=If(Zoning=1, Sequence Ri-1, Sequence Ri-2)

[0070] Among them, Ri(t) represents the value of the i-th key dynamic parameter Ri at each moment;

[0071] Specific approach: A new zoning parameter, zoning, is added to the water environment numerical model, which is divided into two spatial ranges. The grid value in spatial range 1 is 1, and the grid value in spatial range 2 is 2.

[0072] For example, when the partition instruction function zoning takes a value equal to 1, the value of the key dynamic parameter R1 at each moment is assigned using the sequence R1-1. When the partition instruction function zoning takes a value different from 1, the value of the key dynamic parameter R1 at each moment is assigned using the sequence R1-2. Sequence R1-1 includes time 1, time 2, and time 3, corresponding to grid values ​​x, y, and z, respectively. Sequence R1-2 includes time 1, time 2, and time 3, corresponding to grid values ​​x', y', and z', respectively. The parameter adjustment process is shown in the following formula:

[0073] R1(t)=If(Zoning=1,Sequence R1-1,Sequence R1-2)

[0074] Among them, R1(t) represents the value of the key dynamic parameter R1 at each moment;

[0075] Specifically, the sequence values ​​of different key dynamic parameters in different spatial ranges are associated. The mapping relationship diagram between the spatial range sequence and the key dynamic parameters of the present invention is shown in FIG. Figure 6 As shown, i represents the total number of key dynamic parameters, and n represents the total number of spatial range sequences.

[0076] For example, Figure 5 In the illustrated embodiment, all key dynamic parameters are spatially divided into three regions: the north, the middle, and the south, based on the actual characteristics of the region and research needs. In the prepared parameter space division file, the north, the middle, and the south are defined as spatial ranges 1, 2, and 3; all sequence values ​​in spatial range 1 (corresponding to the north) are 1; all sequence values ​​in range 2 (corresponding to the middle) are 2, and all sequence values ​​in range 3 (corresponding to the south) are 3. All grids are assigned values ​​using if statements, and the following example is given using the key dynamic parameter, the sediment NH4 release rate (R1):

[0077] R1(t)=If(Zoning==1,R1(t)_NC,if(Zoning==2,R1(t)_C,R1(t)_SC))

[0078] Where Zoning is the partitioning instruction function, and R1(t)_NC, R1(t)_C, and R1(t)_SC are the sequence values ​​of the key dynamic parameters in different spatial ranges. When the partitioning instruction function Zoning is equal to 1, the R1(t)_NC sequence is assigned to spatial range 1; when the partitioning instruction function Zoning is equal to 2, the R1(t)_C sequence is assigned to spatial range 2; when the partitioning instruction function Zoning is not equal to 1 or 2, the R1(t)_SC sequence is assigned to spatial range 3. Parameter adjustment can be achieved by simply modifying the sequence values ​​of the R1(t)_NC, R1(t)_C, and R1(t)_SC spatial ranges.

[0079] Step S5: Iteratively modify the sequence values ​​of different parameters with different spatial distributions until the simulated values ​​are well matched with the measured values, i.e., a well-fitting parameter adjustment result is obtained; the degree of agreement between the simulated values ​​and the measured values ​​can be evaluated using indicators such as Nash efficiency coefficient (NSE) and mean relative error (MRE);

[0080] Furthermore, based on the above-mentioned S1 to S5 processes, the following step S6 is further included to implement a method for optimizing the spatiotemporal dynamic parameters of a water environment numerical model based on the spatiotemporal dynamic parameter calibration method of the present invention:

[0081] Step S6: Prevent overfitting and determine the key dynamic parameters for final optimization; when the simulated values ​​are close to the measured values ​​and the evaluation indicators such as NSE and MRE are no longer significantly improved, the parameter adjustment process of the key dynamic parameters is stopped to prevent the water environment numerical model from overfitting, and the parameters at this time are determined as the final optimization parameters.

[0082] In summary, generally, the time-space dynamic parameters are prepared once and then run on the model. The present invention is equivalent to separating the two processes of time and space dynamics, and then using the connection equation to correspond the time dynamics and space dynamics.

[0083] Specific example: The simulation time is 2 years (June 1, 2021-June 1, 2023), and the simulation area is about 245 km 2 The size of a spatiotemporal dynamic file for a single parameter is approximately 7.3MB. After optimization using this method, only a single 546KB sequence file is required, significantly compressing the model structure. In this example, there are eight key dynamic parameters, and the compression effect is significant (8*7.3MB → 8*546KB), improving the model's operational efficiency. Furthermore, the original "create sequence file - create spatiotemporal dynamic file - run model" process is simplified to "create sequence file - run model." During repeated adjustments to model parameters, only the relatively easy-to-produce sequence file needs to be created, greatly improving parameter adjustment efficiency.

[0084] Table 1

[0085]

[0086]

[0087] Note: The values ​​with ※ are key dynamic parameters

[0088] Table 2

[0089]

[0090]

[0091] It should be noted that although the present invention has been shown and described with reference to specific exemplary embodiments of the present invention, those skilled in the art should understand that the present invention is not limited to the above-mentioned embodiments and all kinds of changes to the present invention fall within the scope of protection of the present invention.

Claims

1. A method for calibrating spatiotemporal dynamic parameters of a water environment numerical model, characterized in that: include: S1: Construct a numerical model of the water environment in the target area, i.e., obtain the topography, meteorology, water quality and flow according to the characteristics of the target area as the first part, and obtain the state variables of the water environment simulation, the mechanism process and parameter composition of each state variable according to the research purpose as the second part, and construct the initial numerical model of the water environment in the target area from the first and second parts; S2: Determine the parameters of the state variable parameters that are highly correlated with the sensitivity analysis or trial operation result analysis of the target water environment as key dynamic parameters; S3: Divide the target area into multiple spatial ranges according to environmental characteristics and parameter adjustment efficiency indicators. Each spatial range is composed of multiple grids. The grid is the minimum calculation unit of the water environment numerical model. The grid attribute value is set to correspond to the sequence value of the spatial range. S4: Define parameter value equations: associate the sequence values ​​of different key dynamic parameters in different spatial ranges, including determining the spatial range n to which the grid belongs, making a parameter space partitioning file, and assigning sequence values ​​to the spatial ranges to all grids; A new zoning instruction function, zoning, is added to the water environment numerical model to realize the key dynamic parameter calibration process in the target area. That is, for the i-th key dynamic parameter Ri, the sequence values ​​of different key dynamic parameters in different spatial ranges are iteratively modified. When the zoning instruction function is equal to 1, the value of Ri at each moment is represented as the sequence value Ri_1. Otherwise, the sequence value is modified from the current sequence value Ri_1 to the sequence values ​​of other spatial ranges Ri_2 to Ri_n. The parameter adjustment is achieved by modifying the sequence value. S5: until the simulated value and the measured value are close to each other, and the fitting parameter adjustment result is obtained.

2. A method for calibrating spatiotemporal dynamic parameters of a water environment numerical model according to claim 1, characterized in that: The iterative modification in S5 of the key dynamic parameters according to the degree of agreement between the simulated value and the measured value based on the parameter value equation is performed dynamically according to the sequence values ​​associated with different spatial ranges through the parameter space division instruction according to the attribute value.

3. A method for optimizing the spatiotemporal dynamic parameters of a water environment numerical model implemented by a spatiotemporal dynamic parameter calibration method of a water environment numerical model according to any one of claims 1 to 2, characterized in that: include: S1: Construct a numerical model of the water environment in the target area, i.e., obtain the topography, meteorology, water quality and flow according to the characteristics of the target area as the first part, and obtain the state variables of the water environment simulation, the mechanism process and parameter composition of each state variable according to the research purpose as the second part, and construct the initial numerical model of the water environment in the target area from the first and second parts; S2: Determine the parameters of the state variable parameters that are highly correlated with the sensitivity analysis or trial operation result analysis of the target water environment as key dynamic parameters; S3: Divide the target area into multiple spatial ranges according to environmental characteristics and parameter adjustment efficiency indicators. Each spatial range is composed of multiple grids. The grid is the minimum calculation unit of the water environment numerical model. The grid attribute value is set to correspond to the sequence value of the spatial range. S4: Define parameter value equations: associate the sequence values ​​of different key dynamic parameters in different spatial ranges, including determining the spatial range n to which the grid belongs, making a parameter space partitioning file, and assigning sequence values ​​to the spatial ranges to all grids; A new zoning instruction function, zoning, is added to the water environment numerical model to realize the key dynamic parameter calibration process in the target area: that is, for the i-th key dynamic parameter Ri, the sequence values ​​of different key dynamic parameters in different spatial ranges are iteratively modified. When the zoning instruction function is equal to 1, the value of Ri at each moment is represented as the sequence value Ri_1. Otherwise, the sequence value is modified from the current sequence value Ri_1 to the sequence values ​​of other spatial ranges Ri_2 to Ri_n. The parameter adjustment is achieved by modifying the sequence value. S5: until the simulated value and the measured value are close to each other, and the fitting parameter adjustment result is obtained; S6: When the simulated value is close to the measured value and the evaluation index of the simulated value and the measured value is no longer significantly improved, the parameter adjustment is stopped to prevent overfitting, and the parameters at this time are determined as the final optimization parameters.

4. A method for optimizing spatiotemporal dynamic parameters of a water environment numerical model according to claim 3, characterized in that: In S5, the key dynamic parameters are dynamically modified according to the degree of agreement between the simulated values ​​and the measured values ​​based on the parameter value equation. The sequence values ​​associated with different spatial ranges are dynamically modified according to the attribute values ​​through the parameter space division instruction.

5. The method for optimizing spatiotemporal dynamic parameters of a water environment numerical model according to claim 3, characterized in that: The parameters pre-adjusted to the point where the simulated value and the measured value are at the same level are used as the initial parameters.

Images