Grid scale wetland flooding state judgment method, device, medium and program

Abstract

The invention relates to a grid scale wetland flooding state judgment method, device, medium and program, a linear relation model of subgrid underground water level and topographic index is constructed, the subgrid underground water level is 0 to represent a critical flooding state, the topographic index value at the moment is obtained based on the linear relation model, and the subgrid underground water level is calculated according to the topographic index value. The terrain index is used as a terrain index threshold value of sub-grid flooding; calculating a flooding area proportion for each grid based on the flooding state of the sub-grids; the grid is composed of a plurality of sub-grids; constructing a fitting model of the grid average groundwater level and the proportion of the flooding area; acquiring a wetland image and a grid average underground water level in the image; and obtaining the grid flooding area proportion of the wetland based on the fitting model. According to the method, the underground water level change and the subgrid terrain heterogeneity characteristics can be coupled on the grid scale, a new technical scheme for fine judgment of the wetland flooding state is achieved, and the defect of an existing hydrological model in the aspect of wetland flooding space expression ability is overcome.

Description

Technical Field

[0001] This invention belongs to the field of hydrological data processing technology, specifically relating to methods, equipment, media, and programs for determining wetland flooding status at the grid scale. Background Technology

[0002] Wetlands are typical hydrological-ecological coupled systems, and their flooding status is jointly controlled by multiple factors, including precipitation, topography, soil properties, groundwater level, and evapotranspiration. Whether wetlands flood and the spatial distribution pattern of flooding not only directly affect regional water cycle processes but also have significant impacts on carbon cycle, biological habitats, and ecosystem service functions. Therefore, accurately determining the flooding status of wetlands is one of the fundamental technical issues in hydrological simulation and ecological assessment.

[0003] Currently, distributed or semi-distributed hydrological models, such as SWAT, TOPMODEL, and various land surface process models, are widely used in wetland hydrological process simulation research. These models typically use grid cells or sub-basin cells as the basic computational units, and groundwater level, soil moisture content, or runoff as key hydrological state variables to numerically simulate surface-groundwater hydrological processes. However, the output results of these models are mostly average groundwater level or average water content at the scale of a single grid cell or sub-basin, which essentially belongs to "spatial average hydrological state description".

[0004] For wetlands, a geomorphic unit with strong spatial heterogeneity, flooding often occurs first in low-lying catchment areas, while areas with slightly higher elevations or steeper slopes may not experience flooding even under the same average groundwater level. In other words, within the same hydrological model grid, groundwater level and flooding status exhibit significant sub-grid scale differences. However, existing hydrological models typically lack the ability to characterize micro-topographical differences within grid cells, making it difficult to address the crucial question of whether different locations within the same grid are flooded.

[0005] In existing technologies, some studies have attempted to use groundwater level threshold methods or soil saturation determination methods, directly using the groundwater level or water content output by hydrological models as the basis for judging whether wetlands are flooded. However, these methods typically use whether the grid average groundwater level exceeds a certain empirical threshold as the criterion, thus simply dividing the entire grid into "fully flooded" or "not flooded" states. This method ignores the topographic relief and runoff differences within the grid, easily leading to overgeneralization or underestimation of wetland flooding areas, and failing to reflect the spatial distribution characteristics of real wetland flooded patches.

[0006] In addition, although some distributed hydrological models have incorporated topographic factors such as topographic indices and catchment areas to characterize runoff generation and saturation zone distribution, most related studies focus on runoff formation mechanisms rather than directly aiming at "wetland flooding status determination". Moreover, their results are often output in the form of source area proportion or statistical saturation zone proportion, making it difficult to achieve a refined quantitative expression of the wetland flooding area proportion at the grid scale.

[0007] Therefore, existing wetland flooding assessment techniques based on hydrological models generally have the following shortcomings:

[0008] First, it can only obtain the average hydrological status at the grid or sub-basin scale, which is difficult to reflect the local flooding differences at the sub-grid scale;

[0009] Second, the determination of wetland flooding relies heavily on empirical thresholds and lacks a quantitative constraint mechanism that couples groundwater level with topography.

[0010] Third, it is difficult to directly output the "raster-scale flooded area ratio" result that corresponds one-to-one with the remote sensing flooded area, thus limiting the accuracy verification and application of the model results. Summary of the Invention

[0011] The primary objective of this invention is to overcome the deficiencies of the prior art and provide a method for determining wetland flooding status at the grid scale.

[0012] To achieve the above objectives, the present invention adopts the following technical solution:

[0013] A method for determining wetland flooding status at a grid scale, the method comprising:

[0014] Constructing subgrid groundwater level and topographic index The linear relationship model is used to characterize the critical flooding state with the groundwater level of the subgrid as 0. The topographic index value at this time is obtained based on the linear relationship model and used as the topographic index threshold for subgrid flooding.

[0015] Based on the flooding status of the subgrids, the flooded area percentage is calculated for each grid; the grid consists of multiple subgrids.

[0016] Construct a fitting model between the grid average groundwater level and the proportion of the flooded area;

[0017] Wetland topographic spatial data of the study area are obtained, and the average groundwater level of the grid is calculated; the proportion of wetland grid flooded area is obtained based on the fitted model.

[0018] In some embodiments of the present invention, the flooding state of the subgrid is determined based on the following method:

[0019] Set a minimum flooding threshold for subgrids;

[0020] For each subgrid, the maximum value between the terrain index threshold and the minimum flooding threshold of the subgrid is taken as the judgment threshold. If the terrain index of the subgrid is not less than the judgment threshold, the subgrid is judged to be in a flooded state.

[0021] In some embodiments of the present invention, the fitting model takes the following form:

[0022]

[0023] in, These are the fitting parameters; The average groundwater level is the grid average. The percentage of flooded area in the fitted grid.

[0024] In some embodiments of the present invention, the method further includes setting a minimum flooding threshold for the subgrid;

[0025] For each subgrid, if the terrain index of the subgrid is not less than the minimum flooding threshold, the subgrid is determined to be in a flooded state; the flooded area percentage of each grid at this time is calculated as the maximum potential flooded area percentage.

[0026] The flooded area percentage is obtained based on the fitted model. The smaller value between the flooded area percentage obtained by the model and the maximum potential flooded area percentage is taken as the actual flooded area percentage of the grid.

[0027] In some embodiments of the present invention, the fitting parameters of the fitting model are obtained based on the following method:

[0028] A minimum flooding threshold is set for the subgrid, and parameter ranges are set for the minimum flooding threshold and soil parameters. Multiple parameter combinations are constructed within the parameter range. The soil parameter is selected as the exponential decay coefficient of soil saturated water conductivity with depth.

[0029] For each set of parameters, assume a series of average groundwater levels for the grid, and obtain the corresponding percentage of flooded area in the grid under the assumed water level to form multiple sets of water level-area percentage sample points. Based on the sample points, a fitting model is obtained.

[0030] The measured daily groundwater level sequence is substituted into the fitting model to obtain the daily wetland flooding area ratio, and the fitting parameters are evaluated using the true values ​​to obtain the optimal combination of fitting parameters.

[0031] In some embodiments of the present invention, the method further includes obtaining the annual grid flooded area percentage based on daily-scale grid flooded area percentage data, including:

[0032] Obtain the time series of the daily percentage of wetland flooded area;

[0033] The time series is slidably sampled based on an N-day sliding window, with the minimum percentage of wetland flooded area within the N-day sliding window used as the smoothing value of the sliding window; N is the preset sliding window size, preferably N=30;

[0034] The maximum value of the smoothed value series throughout the year is taken as the percentage of flooded wetland area in the annual wetland grid.

[0035] In some embodiments of the present invention, the sliding value is obtained in segments, as follows:

[0036]

[0037] In the formula, t is the minimum value after N days; t is the time sequence number of the current date in the total number of days to be calculated; T is the total number of days to be calculated.

[0038] The present invention further provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the above method.

[0039] The present invention further provides a computer-readable storage medium having a computer program / instructions stored thereon, which, when executed by a processor, implement the steps of the above-described method.

[0040] The present invention further provides a computer program product, including a computer program / instruction that, when executed by a processor, implements the steps of the above-described method.

[0041] The method of this invention can couple groundwater level changes with subgrid topographic heterogeneity features at the grid scale, realizing a new technical solution for refined determination of wetland flooding status, and making up for the shortcomings of existing hydrological models in terms of spatial representation capabilities of wetland flooding. Attached Figure Description

[0042] Figure 1 It is a flowchart of the model parameter fitting process in a single grid.

[0043] Figure 2 It is a flowchart for parameter selection in a single raster.

[0044] Figure 3 This is a comparison between the sliding window minimum method and the traditional average method. Detailed Implementation

[0045] The technical solution of the present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0046] This invention uses topographic spatial data to determine flooding status at the raster scale (i.e., the grid described in the embodiment). Topographic spatial data is used to construct a sub-grid topographic index system and calculate the potential flooded area; in this embodiment, DEM data is used. Additionally, the wetland flooded area percentage sequence contained in remote sensing observations or publicly available wetland flooding dynamic products is used as ground truth data for final comparison and parameter calibration. The remote sensing observation data must have the same resolution as the DEM data, or be resampled to match the DEM data.

[0047] Example 1

[0048] The method for determining wetland flooding status at the grid scale according to the present invention includes the following steps:

[0049] I. Obtaining the threshold for the critical terrain index;

[0050] Existing studies have proposed distributed hydrological models (such as TOPMODEL) that establish the relationship between the grid-averaged groundwater level and the saturated area fraction within the subgrid. The Compound Topographic Index (CTI) of the subgrid reflects the local topographic water catchment characteristics; the higher the value, the higher the probability that the subgrid will be flooded.

[0051] Distributed hydrological models determine the location and area of ​​saturated zones (source zones) based on soil moisture content, and their core is derived from the continuity equation and Darcy's law.

[0052] (1)

[0053] in, For precipitation intensity, For the flow intensity, For coefficients, This is the distance along the direction of maximum slope. This is the water shortage, which is the difference between the saturated water content and the soil moisture content.

[0054] Assuming the hydraulic gradient over the saturated area approximates the surface topographic slope. Furthermore, soil permeability decreases exponentially with depth, from which the underground runoff can be obtained:

[0055] (2)

[0056] in, The conductivity when the soil is just saturated. is a parameter representing the exponential decay coefficient of soil saturated water conductivity with depth.

[0057] The relationship between local water shortage and runoff can be obtained from equation (2):

[0058] (3)

[0059] The average water shortage can be obtained by integrating over the entire catchment area:

[0060] (4)

[0061] in, The total area of ​​the grid. This represents the subgrid area.

[0062] Assuming that the flow rate (i.e., runoff) per unit area is spatially consistent, and that precipitation replenishment... , as well as Keeping the water shortage constant, we can obtain the distribution equation for the water shortage:

[0063] (5)

[0064] in, This is the average topographic index of the watershed.

[0065] This allows for the establishment of groundwater levels. With topographic index Linear relationship:

[0066] (6)

[0067] in, For the subgrid groundwater level, The average groundwater level in the grid. For subgrid topographic indices, This represents the average terrain index for the grid.

[0068] when When the subgrid is in a flooded state, the critical topographic index threshold (the threshold when the subgrid is exactly in a flooded state) can be obtained:

[0069] (7)

[0070] II. Calculation of the proportion of flooded area in the grid

[0071] For each grid cell, its flooded area percentage Defined as:

[0072] (8)

[0073] in, The threshold for the flooding critical topographic index in the subgrid. This is the minimum flood threshold for the subgrid, used to exclude areas that are unlikely to be flooded, such as steep slopes or high ground. For the first Terrain index of each subgrid.

[0074] Percentage of the largest potential flooded area This represents the theoretical maximum flooded area percentage under extreme saturation conditions:

[0075] (9)

[0076] III. Empirical Fitting of the Relationship between Groundwater Level and Flooded Area

[0077] By fitting the relationship between the average groundwater level of the grid and the proportion of the grid flooded area, an empirical formula can be obtained:

[0078] (10)

[0079] in, These are empirical parameters, which can be obtained through least squares fitting. The final actual flooded area percentage is:

[0080] (11)

[0081] IV. Parameter Optimization and Model Fitting

[0082] In parameter space It internally constructs 135 parameter combinations. For each parameter combination... A series of assumptions were set regarding the average groundwater level of the grid. And at each assumed water level, the proportion of flooded area of ​​the corresponding grid is calculated based on formula (8). Obtain discrete sample points By fitting, we can obtain With continuous flood response function ( ).

[0083] Subsequently, the daily groundwater level sequence was derived based on the regional surface water level conditions. ,Will Substituting the values ​​into the flood response function and calculating the daily percentage of flooded wetland area, we obtain... .

[0084] Ultimately, The proportion of flooded area observed by remote sensing A grid-by-grid comparison is performed to evaluate the applicability of the current parameter combination and to select the optimal parameter combination. The parameter calculation process for a single grid is as follows: Figure 1 As shown.

[0085] The parameter set with the smallest root mean square error (RMSE) is selected as the optimal solution.

[0086] (12)

[0087] in, This represents the percentage of flooded area observed by remote sensing in the nth grid. The model simulates the percentage of flooded area for the nth grid. This represents the number of grid cells.

[0088] The optimal parameter selection process is as follows: Figure 2 As shown.

[0089] V. Calculation Methods for Wetland Flooded Area from Daily to Annual Scales

[0090] In wetland dynamic monitoring, remote sensing and model simulation typically output the percentage of flooded area on a daily scale. However, the actual hydrological state of wetlands is influenced by the combined effects of climate, topography, vegetation, and hydrological processes, exhibiting significant short-term fluctuations and seasonal characteristics. Calculating annual values ​​directly from daily-scale results is highly susceptible to extreme rainfall events, short-term droughts, or observational errors, leading to inaccurate sudden increases or decreases in wetland area over time. This not only affects the determination of wetland evolution trends but also distorts inter-regional comparisons of wetland dynamics.

[0091] Therefore, this invention proposes a time-series aggregation method that combines the sliding window minimum method with the annual maximum value extraction in the extraction of annual wetland area. Its core idea is: (1) By using the sliding minimum method, the continuous flooding state of each day within a certain time window (such as 30 days) before and after it is identified, effectively eliminating instantaneous water accumulation or observation noise interference; (2) Then, the maximum flooded area of ​​wetland throughout the year is extracted by the annual maximum value, reflecting the maximum effective distribution range of wetland in that year.

[0092] Compared with traditional daily or monthly averaging methods, this method has significant advantages: it effectively filters out false signals caused by short-term fluctuations by operating on the minimum value within the time window, making the interannual variation stable and robust; the extracted continuous water accumulation characteristics are consistent with the physical meaning of long-term water accumulation in wetland ecosystems; the use of a unified time window can ensure comparability between different climatic regions and effectively suppress abnormal factors such as remote sensing shadows, sensor noise, or short-term heavy rainfall, thereby improving the reliability and consistency of wetland identification. Figure 3 It shows a comparison of different methods for converting the percentage of flooded area from daily to annual scale on a single grid.

[0093] The calculation method is as follows:

[0094] Let the daily percentage of wetland flooded area be... ,in A sliding minimum method with a 30-day window is used to smooth the value for each day. The definition is as follows:

[0095] (13)

[0096] Subsequently, the annual wetland area can be extracted from the maximum value of the annual sliding minimum sequence, i.e.:

[0097] (14)

[0098] in, It represents the percentage of the wetland's maximum effective distribution area in a hydrological sense in that year.

[0099] This method can accurately reflect the maximum continuous flooded area of ​​wetlands at the annual scale, while preserving short-term dynamic characteristics and effectively avoiding interference from extreme events and noise on annual statistics.

Claims

1. A method for determining wetland flooding status at the grid scale, characterized in that, The method includes: Constructing subgrid groundwater level and topographic index The linear relationship model is used to characterize the critical flooding state with the groundwater level of the subgrid as 0. The topographic index value at this time is obtained based on the linear relationship model and used as the topographic index threshold for subgrid flooding. Based on the flooding status of the subgrids, the flooded area percentage is calculated for each grid; the grid consists of multiple subgrids. Construct a fitting model between the grid average groundwater level and the proportion of the flooded area; Wetland topographic spatial data of the study area are obtained, and the average groundwater level of the grid is calculated; the proportion of wetland grid flooded area is obtained based on the fitted model.

2. The method according to claim 1, characterized in that, The flooding status of the subgrid is determined based on the following method: Set a minimum flooding threshold for subgrids; For each subgrid, the maximum value between the terrain index threshold and the minimum flooding threshold of the subgrid is taken as the judgment threshold. If the terrain index of the subgrid is not less than the judgment threshold, the subgrid is judged to be in a flooded state.

3. The method according to claim 1, characterized in that, The fitting model is in the following form: ； in, These are the fitting parameters; The average groundwater level is the grid average. The percentage of flooded area in the fitted grid.

4. The method according to claim 1, characterized in that, The method also includes setting a minimum flooding threshold for the subgrid; For each subgrid, if the terrain index of the subgrid is not less than the minimum flooding threshold, the subgrid is determined to be in a flooded state; the flooded area percentage of each grid at this time is calculated as the maximum potential flooded area percentage. The flooded area percentage is obtained based on the fitted model. The smaller value between the flooded area percentage obtained by the model and the maximum potential flooded area percentage is taken as the actual flooded area percentage of the grid.

5. The method according to claim 1 or 3, characterized in that, The fitting parameters of the fitting model are obtained based on the following method: A minimum flooding threshold is set for the subgrid, and parameter ranges are set for the minimum flooding threshold and soil parameters. Multiple parameter combinations are constructed within the parameter range. The soil parameter is selected as the exponential decay coefficient of soil saturated water conductivity with depth. For each set of parameters, assume a series of average groundwater levels for the grid, and obtain the corresponding percentage of flooded area in the grid under the assumed water level to form multiple sets of water level-area percentage sample points. Based on the sample points, a fitting model is obtained. The measured daily groundwater level sequence is substituted into the fitting model to obtain the daily wetland flooding area ratio, and the fitting parameters are evaluated using the true values ​​to obtain the optimal combination of fitting parameters.

6. The method according to claim 1, characterized in that, The method further includes obtaining the annual grid flooded area percentage based on daily-scale grid flooded area percentage data, including: Obtain the time series of the daily percentage of wetland flooded area; The time series is slidably sampled based on an N-day sliding window, with the minimum percentage of wetland flooded area within the N-day sliding window used as the smoothing value of the sliding window; N is the preset sliding window size, preferably N=30; The maximum value of the smoothed value series throughout the year is taken as the percentage of flooded wetland area in the annual wetland grid.

7. The method according to claim 6, characterized in that, The sliding value retrieval method is segmented value retrieval, as follows: ； In the formula, This represents the minimum value over N days. t is the time sequence number of the current date in the total number of days to be calculated; T is the total number of days to be calculated.

8. A computer device, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program / instructions stored thereon, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method according to any one of claims 1 to 7.

10. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method according to any one of claims 1 to 7.

Images