A method of identifying a flooding wetland compound heat event

Abstract

This invention discloses a method for identifying combined dry-heat events in floodplain wetlands, comprising: acquiring topographic, hydrological, meteorological, land use, and vegetation index data of the study area, determining the extent of the sandbars and the critical exposure water level, and constructing a standardized analysis dataset; simulating the wetland inundation process using a hydrodynamic model, and calculating the dry and thermal indices for each grid; calculating the vegetation degradation rate based on the vegetation index, and determining the dry and thermal thresholds by combining a random forest model and joint probability distribution analysis; determining whether the dry and thermal threshold conditions are simultaneously met for each grid, and statistically analyzing the grid proportion of dry-heat events; determining the regional event judgment threshold based on the historical grid proportion sequence using Pearson's three-type curve, and then determining whether it is a regional-level combined dry-heat event year; this invention achieves multi-level accurate identification of combined dry-heat events in floodplain wetlands, providing effective technical support for wetland ecological protection and disaster early warning.

Description

Technical Field

[0001] This invention relates to the field of hydrology and wetland ecological monitoring technology, specifically a method for identifying combined dry and hot events in floodplain wetlands. Background Technology

[0002] Floodplains, possessing characteristics of both terrestrial and aquatic ecosystems, are key areas for global biodiversity conservation and play important ecological roles such as carbon sequestration, water purification, and water conservation. Their health directly impacts regional ecological security and sustainable development. In recent years, under the dual influence of global climate change and human activities, extreme weather events have become more frequent. Among them, combined dry-heat events, a type of disaster caused by the synergistic effects of drought and high temperatures, cause continuous and cumulative damage to wetland ecosystems, seriously threatening the stability and recovery of key ecological components such as wetland vegetation and benthic organisms. Therefore, accurately identifying combined dry-heat events has become a crucial prerequisite for wetland ecological protection, disaster early warning, and restoration.

[0003] Currently, research on complex events in water systems such as lakes has been preliminarily initiated. These events are typically defined as a combination of single extreme events that occur concurrently or continuously in time or space, and it is noted that their probability of occurrence and impact are significantly enhanced under climate warming, water level fluctuations, and human disturbance. Existing methods mainly focus on complex processes directly related to water bodies, such as heat waves-hypoxia and drought-water level decline, and the assessment objects are mostly concentrated on the ecological responses within lakes, such as fish mortality and algal blooms. However, floodplains have unique sandbar ecological units with elevations between low and high water levels, relying on periodic changes in water level to achieve the alternation of aquatic and terrestrial states. Traditional methods have failed to study this unit, ignoring the drought response mechanism of long-term exposure of sandbars due to drought, and failing to incorporate the sustained high temperatures after exposure into the complex event identification system. Therefore, they cannot accurately characterize the complexity and ecological damage characteristics of dry-heat events in floodplains.

[0004] Furthermore, existing methods, when assessing the impact of complex events, often focus on the ecological processes of lake water bodies, lacking quantitative analysis of key wetland ecological responses such as degradation of beach vegetation and loss of benthic organisms. This leads to a disconnect between the identification results and the actual ecological damage to wetlands, making it difficult to support targeted protection and restoration decisions.

[0005] Therefore, there is an urgent need for a composite dry-heat event identification method that combines the characteristics of beach ecological units, integrates multi-source data and ecological response models, in order to solve the problems of existing technologies that are divorced from the actual ecological conditions of beach areas, ignore the composite amplification effect of dry-heat, and have threshold settings that are disconnected from ecological damage. This would enable accurate and scientific identification of composite dry-heat events in floodplain wetlands and provide reliable technical support for wetland ecosystem protection and adaptive management. Summary of the Invention

[0006] To address the shortcomings of existing technologies, the purpose of this invention is to provide a method for identifying combined dry and hot events in floodplain wetlands, thereby solving the problems of traditional methods being divorced from the ecological characteristics of sandbars, ignoring the combined amplification effect of drought and high temperature, and having a disconnect between threshold settings and ecological responses.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for identifying combined dry-heat events in floodplain wetlands, comprising the following steps:

[0008] Step S1: Obtain research data within the study area, including topographic data, land use data, and year-end normalized vegetation index data;

[0009] Step S2: Determine the extent of the sandbar based on topographic and land use data; construct a hydrodynamic model based on topographic and sandbar extent; determine the critical water level for sandbar exposure using the dry and wet water depth parameters of the hydrodynamic model;

[0010] Step S3: Calculate the vegetation degradation rate based on the year-end normalized vegetation index data; standardize the dry and thermal indices of the hydrodynamic model and construct a gridded analysis dataset;

[0011] Step S4: Use the hydrodynamic model to identify and analyze the dry and wet states of each grid in the dataset; count the number of consecutive dry days for each grid throughout the year as a dryness index, and count the number of high-temperature days during the dry period as a heat index;

[0012] Step S5: Using standardized dry and heat indices as inputs and vegetation degradation rate as output, construct a random forest regression model; extract ecologically acceptable dry and heat constraint boundaries through the random forest regression model to obtain ecological weighted constraint conditions; simultaneously establish a joint probability distribution model of dry and heat indices, and combine it with preset extreme event probability thresholds to obtain joint probability constraint conditions; take the intersection of ecological weighted constraint conditions and joint probability constraint conditions to determine the dry threshold and heat threshold.

[0013] Step S6: For each grid in the analysis dataset of the target year, determine whether the dry index is not lower than the dry threshold and the thermal index is not lower than the thermal threshold at the same time. If both are met, determine that a dry-thermal event has occurred in a single grid of the target year, and calculate the proportion of grids with dry-thermal events in the target year.

[0014] Step S7: Based on the grid proportion of dry and hot events in historical years, determine the regional proportion threshold through frequency analysis; if the grid proportion of dry and hot events in the target year is not lower than the regional proportion threshold, then the target year is determined to be a regional composite dry and hot event year.

[0015] Furthermore, in step S1, research data within the study area is acquired. This research data includes topographic data, land use data, and year-end normalized vegetation index data; the research data also includes hydrological data and meteorological data; specifically:

[0016] Step S11: Select the floodplain wetland as the study area and divide the study area into grid cells with a resolution of 30m×30m;

[0017] Step S12: Collect topographic data, hydrological data, meteorological data, land use data, and year-end vegetation index data of the floodplain wetland;

[0018] Step S13: The terrain data is acquired through the National Geomatics Center of China to obtain a digital elevation model with a resolution of not less than 30m.

[0019] Step S14: Hydrological data are acquired through the hydrological station network to obtain daily water level and flow data of the study area and upstream control stations;

[0020] Step S15: Meteorological data is acquired through the National Meteorological Science Data Center to obtain daily maximum temperature data of meteorological stations within the study area and a ten-kilometer radius around it.

[0021] Step S16: Land use data is acquired through the national geographic conditions monitoring cloud platform to obtain an annual land use map with a resolution of not less than 30m. The land use map is classified into cultivated land, forest land, grassland, construction land, water area, and unused land.

[0022] Step S17: The year-end vegetation index data is extracted using the normalized vegetation index product of the medium resolution imaging spectrometer to obtain the vegetation index data for December of each year.

[0023] Furthermore, in step S2, the extent of the sandbar is determined based on topographic and land use data; a hydrodynamic model is constructed based on the topographic data and the extent of the sandbar; and the critical water level for sandbar exposure is determined using the wet and dry water depth parameters of the hydrodynamic model; specifically:

[0024] Step S21: Based on the digital elevation model determined by topographic data and the multi-year average low water level and flood level determined by hydrological data, determine the elevation range of the sandbar, and overlay the land use map to remove permanent water bodies and building land to obtain the sandbar range.

[0025] Step S22: Using two-dimensional simulation tools for physical, chemical, or biological processes in water bodies, a hydrodynamic model of the study area is constructed based on the digital elevation model and the extent of the sandbar.

[0026] Step S23, determine the critical exposure water level using the dry water depth parameters in the hydrodynamic model: directly use the dry and wet water depth parameters built into the hydrodynamic model to set the critical exposure water level.

[0027] Furthermore, in step S3, the vegetation degradation rate is calculated based on the year-end normalized vegetation index data; the dry and thermal indices of the hydrodynamic model are standardized to construct a gridded analysis dataset; specifically:

[0028] Step S31: Based on the year-end normalized vegetation index data, calculate the annual vegetation degradation rate for each grid in the analysis dataset, using the following formula:

[0029] ;

[0030] in, This represents the vegetation degradation rate in grid j of year i. This represents the maximum vegetation index at the end of the past 10 years for the j-th grid. This represents the year-end vegetation index value of the j-th grid in the i-th year;

[0031] Step S32: Standardize the dry and thermal indices of the hydrodynamic model using standard fractions to obtain the standardized dry and thermal indices, as shown in the following formulas:

[0032] ;

[0033] in, This represents the standardized dry and thermal properties. This represents the raw values ​​of the dryness and heat indices. This represents the mean of the dryness index and the heat index. This represents the standard deviation of the dry and thermal indices;

[0034] Step S33: Construct an analysis dataset for the past 10 years using grids as units. The input features are standardized dry and thermal indicators, and the output variable is the vegetation degradation rate, resulting in a total of N×10 samples, where N represents the total number of grids on the beach.

[0035] Furthermore, in step S4, the hydrodynamic model is used to identify and analyze the dry and wet states of each grid in the dataset; the number of consecutive dry days for each grid throughout the year is counted as a dryness index, and the number of high-temperature days during the dry period is counted as a heat index; specifically:

[0036] Step S41: Use a hydrodynamic model to simulate the wetland inundation process and identify the dry and wet states of each grid in the dataset.

[0037] Step S42: Run the hydrodynamic model on an annual basis with a daily step size, and output the daily water level data and dry or wet state determination results for each grid of the analysis dataset throughout the year; for each grid of the analysis dataset, count the cumulative number of days in the year that the hydrodynamic model determines to be dry, which is the dry index;

[0038] Step S43: Directly call the dry days output by the hydrodynamic model. Based on the daily dry or wet records output by the hydrodynamic model, determine all the periods of continuous dryness in each grid throughout the year. Perform spatiotemporal matching between the grid-level daily maximum temperature data and the above-mentioned dry periods, and count the cumulative number of days with daily maximum temperature ≥36℃ within the dry periods, which is the heat index.

[0039] Furthermore, in step S5, a random forest regression model is constructed using standardized dry and heat indices as inputs and vegetation degradation rate as output. The ecologically acceptable dry and heat constraint boundaries are extracted through the random forest regression model to obtain ecologically weighted constraints. Simultaneously, a joint probability distribution model of the dry and heat indices is established, and combined with preset extreme event probability thresholds, joint probability constraints are obtained. The intersection of the ecologically weighted constraints and the joint probability constraints is used to determine the dry and heat thresholds. Specifically:

[0040] Step S51: Using standardized dry and thermal indices as input features and vegetation degradation rate as output variable, construct a random forest regression model.

[0041] Step S52: Train the random forest regression model. The coefficient of determination and root mean square error are used to evaluate the fitting effect of the random forest regression model, and the trained random forest model is obtained, as shown in the following formula:

[0042] ; ;

[0043] in, The coefficient of determination is represented by the coefficient of determination. This indicates the number of test set samples in the analysis dataset. This represents the actual vegetation degradation rate in year i. This represents the predicted vegetation degradation rate for year i. This represents the average actual value of the vegetation degradation rate. Indicates the root mean square error;

[0044] Step S53: Output the feature importance ratio of dry and hot indices to vegetation degradation rate through the trained random forest model.

[0045] Step S54: Set an ecologically acceptable degradation threshold. Using the trained random forest model, generate a two-dimensional partial dependency graph of dry and hot indices. From the two-dimensional partial dependency graph, read the maximum values ​​of the dry and hot indices that satisfy the predicted vegetation degradation rate being less than or equal to the ecologically acceptable degradation threshold. Combine the maximum values ​​of the dry and hot indices with the feature importance ratio to obtain the ecological weighted constraint conditions, as shown in the following formula:

[0046] ;

[0047] in, This indicates the maximum value of the dry index. This indicates the proportion of the characteristic importance of dry indicators to the vegetation degradation rate. This indicates the maximum value of the thermal index. This indicates the percentage of the characteristic importance of thermal indicators to vegetation degradation rate;

[0048] Step S55: Assuming that the dryness index and the heat index follow a two-dimensional log-normal distribution, construct the joint probability density function as follows:

[0049] ;

[0050] in, Denotes the joint probability density function. Represents pi (π). This represents the standard deviation of the logarithmic series of the dry index. The standard deviation of the logarithmic series of heat indicators. This represents the correlation coefficient between the logarithmic series of dry and heat indicators. The square of, Represents an exponential function. Indicators The natural logarithm, This represents the mean of the logarithmic sequence of the dry index. The natural logarithm of the heat index T. This represents the mean of the logarithmic series of the heat index;

[0051] Step S56: Set the extreme joint event probability threshold P0=5%. Solve the constraint condition set by integrating the extreme joint event probability that satisfies the co-occurrence of the dry index and the thermal index in the joint probability density function to obtain the joint probability constraint condition, as shown in the following formula:

[0052] ;

[0053] in, Indicators Not less than the threshold And thermal index Not less than the threshold The probability of this extreme combined event occurring, The threshold representing the dry index, The threshold representing the thermal index, Indicates in and Within the research area, the joint probability density function Perform double integrals;

[0054] Step S57: Take the intersection of the ecological weighted constraint and the joint probability constraint, that is, the dry index critical value and the thermal index critical value that simultaneously satisfy the ecological weighted constraint and the joint probability constraint. The dry index critical value and the thermal index critical value are determined as the dry threshold and the thermal threshold, respectively.

[0055] Furthermore, in step S6, for each grid in the analysis dataset of the target year, it is determined whether both the dryness index and the heat index are simultaneously satisfied. If both are satisfied, it is determined that a dry-heat event has occurred in a single grid of the target year, and the percentage of grids experiencing dry-heat events in the target year is calculated; specifically:

[0056] Step S61: For each grid in the analysis dataset of the target year, call the number of dry days output by the hydrodynamic model and the number of high-temperature days. If both the dry index and the thermal index are greater than or equal to the dry threshold and the thermal index is greater than or equal to the thermal threshold, then it is determined that a dry-hot event has occurred in the target year for a single grid.

[0057] Step S62: Count the number of beach grids where hot and dry events occurred within the target year, and calculate the percentage of grids involved in hot and dry events, using the following formula:

[0058] ;

[0059] in, Indicates the grid percentage of dry and hot events. Indicates the number of grid cells on the shoal. This represents the total number of beach grids within the study area.

[0060] Furthermore, in step S7, based on the grid proportion of dry and hot events in historical years, a regional proportion threshold is determined through frequency analysis; if the grid proportion of dry and hot events in the target year is not lower than the regional proportion threshold, then the target year is determined to be a regional-level composite dry and hot event year; specifically:

[0061] Step S71: Based on the grid proportion of dry and hot events in a single target year, calculate the grid proportion of dry and hot events for each of the past 10 years to obtain the grid proportion of dry and hot events in historical years, using the following formula:

[0062] ;

[0063] in, The grid represents the percentage of dry and hot events in historical years. This indicates the percentage of dry and hot events in the grid during the first year. This indicates the percentage of dry and hot events in the grid in the second year. This indicates the percentage of grid-related dry and hot events in the tenth year;

[0064] Step S72: Frequency analysis is performed on the grid proportion of dry and hot events in historical years using Pearson type III curves to calculate the mean, coefficient of variation, and skewness coefficient, as shown in the following formulas:

[0065] , where n=10;

[0066] ;

[0067] ;

[0068] in, This represents the mean. Indicates the year. Represents the coefficient of variation. Indicates the skewness coefficient;

[0069] Step S73: Combining the extremes of the dry threshold and the thermal threshold, the mean and coefficient of variation obtained by calculating the frequency of the Pearson type III curve, and the set guarantee rate, the regional proportion threshold is obtained, as shown in the following formula:

[0070] ;

[0071] in, Indicates the threshold for the proportion of a region. This represents the deviation coefficient from the mean of the Pearson type III curve;

[0072] Step S74: If the grid proportion of dry and hot events in the target year is not lower than the regional proportion threshold, it is determined to be a regional composite dry and hot event year.

[0073] Compared with the prior art, the present invention has the following beneficial effects:

[0074] This invention is the first to explicitly define a complex dry-heat event in floodplain wetlands, with drought leading to long-term exposure of sandbars and high temperatures as the core. It breaks through the traditional method's limitation to the identification framework of lake water bodies and accurately focuses the identification object on the ecological units of sandbars with land-water alternation characteristics, which is more in line with the actual ecological structure and response mechanism of floodplain wetlands.

[0075] This invention constructs a three-level identification system consisting of grid-level threshold identification, regional-level proportion determination, and ecological verification feedback. This system achieves full-chain technical coverage from local to overall and from event identification to ecological impact assessment, thereby improving the systematicness and operability of the method.

[0076] This invention achieves dual constraints on the dry heat threshold by integrating hydrodynamic simulation, random forest ecological response modeling, and joint probability distribution analysis. This avoids the problem of threshold setting relying on experience and being out of touch with ecological reality in traditional methods, and significantly improves the objectivity and ecological relevance of the identification results.

[0077] This invention enables refined spatiotemporal identification of hot and dry events based on standardized multi-source data and high-resolution grid computing. By combining historical sequence frequency analysis and regional proportion determination, it can scientifically identify regional compound hot and dry event years, providing reliable data support and decision-making basis for wetland disaster early warning, ecological protection and restoration planning.

[0078] This method is applicable to floodplain wetlands of different types and regions. It can be adapted by adjusting model parameters and data sources, and has good promotional value. It provides an effective technical tool for the identification and response to extreme wetland events under the background of global change. Attached Figure Description

[0079] Figure 1 This is a flowchart of the method steps for identifying combined dry and hot events in floodplain wetlands according to the present invention. Detailed Implementation

[0080] like Figure 1 As shown, this embodiment proposes the following technical solution: a method for identifying combined dry-heat events in floodplain wetlands, the method comprising the following steps:

[0081] Step S1: Obtain research data within the study area, including topographic data, land use data, and year-end normalized vegetation index data;

[0082] Step S2: Determine the extent of the sandbar based on topographic and land use data; construct a hydrodynamic model based on topographic and sandbar extent; determine the critical water level for sandbar exposure using the dry and wet water depth parameters of the hydrodynamic model;

[0083] Step S3: Calculate the vegetation degradation rate based on the year-end normalized vegetation index data; standardize the dry and thermal indices of the hydrodynamic model and construct a gridded analysis dataset;

[0084] Step S4: Use the hydrodynamic model to identify and analyze the dry and wet states of each grid in the dataset; count the number of consecutive dry days for each grid throughout the year as a dryness index, and count the number of high-temperature days during the dry period as a heat index;

[0085] Step S5: Using standardized dry and heat indices as inputs and vegetation degradation rate as output, construct a random forest regression model; extract ecologically acceptable dry and heat constraint boundaries through the random forest regression model to obtain ecological weighted constraint conditions; simultaneously establish a joint probability distribution model of dry and heat indices, and combine it with preset extreme event probability thresholds to obtain joint probability constraint conditions; take the intersection of ecological weighted constraint conditions and joint probability constraint conditions to determine the dry threshold and heat threshold.

[0086] Step S6: For each grid in the analysis dataset of the target year, determine whether the dry index is not lower than the dry threshold and the thermal index is not lower than the thermal threshold at the same time. If both are met, determine that a dry-thermal event has occurred in a single grid of the target year, and calculate the proportion of grids with dry-thermal events in the target year.

[0087] Step S7: Based on the grid proportion of dry and hot events in historical years, determine the regional proportion threshold through frequency analysis; if the grid proportion of dry and hot events in the target year is not lower than the regional proportion threshold, then the target year is determined to be a regional composite dry and hot event year.

[0088] Furthermore, in step S1, research data within the study area is acquired. This research data includes topographic data, land use data, and year-end normalized vegetation index data; the research data also includes hydrological data and meteorological data; specifically:

[0089] Step S11: Select the floodplain wetland as the study area and divide the study area into grid cells with a resolution of 30m×30m;

[0090] Step S12: Collect topographic data, hydrological data, meteorological data, land use data, and year-end vegetation index data of the floodplain wetland;

[0091] Step S13: The terrain data is acquired through the National Geomatics Center of China to obtain a digital elevation model with a resolution of not less than 30m.

[0092] Step S14: Hydrological data are acquired through the hydrological station network to obtain daily water level and flow data of the study area and upstream control stations;

[0093] Step S15: Meteorological data is acquired through the National Meteorological Science Data Center to obtain daily maximum temperature data of meteorological stations within the study area and a ten-kilometer radius around it.

[0094] Step S16: Land use data is acquired through the national geographic conditions monitoring cloud platform to obtain an annual land use map with a resolution of not less than 30m. The land use map is classified into cultivated land, forest land, grassland, construction land, water area, and unused land.

[0095] Step S17: The year-end vegetation index data is extracted using the normalized vegetation index product of the medium resolution imaging spectrometer to obtain the vegetation index data for December of each year.

[0096] Furthermore, in step S2, the extent of the sandbar is determined based on topographic and land use data; a hydrodynamic model is constructed based on the topographic data and the extent of the sandbar; and the critical water level for sandbar exposure is determined using the wet and dry water depth parameters of the hydrodynamic model; specifically:

[0097] Step S21: Based on the digital elevation model determined by topographic data and the multi-year average low water level and flood level determined by hydrological data, determine the elevation range of the sandbar, and overlay the land use map to remove permanent water bodies and building land to obtain the sandbar range.

[0098] Step S22: Using a two-dimensional simulation tool for physical, chemical, or biological processes in water, a hydrodynamic model of the study area is constructed based on the digital elevation model and the area of ​​the sandbar. The grid resolution of the hydrodynamic model is consistent with the 30m×30m grid divided in the previous study area.

[0099] The conventional roughness parameters for the hydrodynamic model include the Manning coefficient n = 0.03-0.05, 0.04-0.05 for shoal areas, and 0.03-0.04 for water areas; the dry water depth for the hydrodynamic model is set to -0.05m to -0.1m, and the wet water depth for the hydrodynamic model is set to 0.05m.

[0100] The hydrodynamic model was calibrated using hydrological data: the daily flow rate of the upstream control station was used as the inflow boundary, the daily water level at the outlet of the study area was used as the outflow boundary, and the initial water level was set as the multi-year average initial water level.

[0101] The parameters of the hydrodynamic model were calibrated: the first 5 years of data were selected for the periodic calibration, and the conventional roughness coefficient and dry and wet water depth thresholds were adjusted so that the Nash efficiency coefficient of the simulated water level and the measured water level were ≥0.75 and the correlation coefficient was ≥0.8; the last 5 years of data were used for the validation period, and the accuracy of the hydrodynamic model was verified according to the same standard.

[0102] Step S23, determine the critical exposure water level using the dry water depth parameters in the hydrodynamic model: directly use the dry and wet water depth parameters built into the hydrodynamic model to set the critical exposure water level;

[0103] In the hydrodynamic model, "dry water depth" is defined as the critical water level at which the grid transitions from "wet" to "dry". When the simulated water level is less than or equal to the dry water depth, the grid is determined to be exposed, meaning the critical water level for exposure is equal to the dry water depth value set by the hydrodynamic model.

[0104] Furthermore, in step S3, the vegetation degradation rate is calculated based on the year-end normalized vegetation index data; the dry and thermal indices of the hydrodynamic model are standardized to construct a gridded analysis dataset; specifically:

[0105] Step S31: Based on the year-end normalized vegetation index data, calculate the annual vegetation degradation rate for each grid in the analysis dataset, using the following formula:

[0106] ;

[0107] in, This represents the vegetation degradation rate in grid j of year i. This represents the maximum vegetation index at the end of the past 10 years for the j-th grid. This represents the year-end vegetation index value of the j-th grid in the i-th year;

[0108] Step S32: Standardize the dry and thermal indices of the hydrodynamic model using standard fractions to obtain the standardized dry and thermal indices, as shown in the following formulas:

[0109] ;

[0110] in, This represents the standardized dry and thermal properties. This represents the raw values ​​of the dryness and heat indices. This represents the mean of the dryness index and the heat index. This represents the standard deviation of the dry and thermal indices;

[0111] Step S33: Construct an analysis dataset for the past 10 years using grids as units. The input features are standardized dry and thermal indicators, and the output variable is the vegetation degradation rate, resulting in a total of N×10 samples, where N represents the total number of grids on the beach.

[0112] Furthermore, in step S4, the hydrodynamic model is used to identify and analyze the dry and wet states of each grid in the dataset; the number of consecutive dry days for each grid throughout the year is counted as a dryness index, and the number of high-temperature days during the dry period is counted as a heat index; specifically:

[0113] Step S41: Use a hydrodynamic model to simulate the wetland inundation process and identify the dry and wet states of each grid in the dataset.

[0114] Step S42: Run the hydrodynamic model on an annual basis with a daily step size, and output the daily water level data and dry or wet state determination results for each grid of the analysis dataset throughout the year; for each grid of the analysis dataset, count the cumulative number of days in the year that the hydrodynamic model determines to be dry, which is the dry index;

[0115] Step S43: Directly call the dry days output by the hydrodynamic model. Based on the daily dry or wet records output by the hydrodynamic model, determine all the periods of continuous dryness in each grid throughout the year. Perform spatiotemporal matching between the grid-level daily maximum temperature data and the above-mentioned dry periods, and count the cumulative number of days with daily maximum temperature ≥36℃ within the dry periods, which is the heat index.

[0116] Furthermore, in step S5, a random forest regression model is constructed using standardized dry and heat indices as inputs and vegetation degradation rate as output. The ecologically acceptable dry and heat constraint boundaries are extracted through the random forest regression model to obtain ecologically weighted constraints. Simultaneously, a joint probability distribution model of the dry and heat indices is established, and combined with preset extreme event probability thresholds, joint probability constraints are obtained. The intersection of the ecologically weighted constraints and the joint probability constraints is used to determine the dry and heat thresholds. Specifically:

[0117] Step S51: Using standardized dryness and heat indices as input features and vegetation degradation rate as the output variable, construct a random forest regression model. The core parameters of the random forest regression model are set as follows:

[0118] The number of decision trees is 100; the maximum tree depth is 5; the minimum number of sample splits is 5; the formula is as follows:

[0119] ;

[0120] in, The predicted value representing the rate of vegetation degradation. Indicates the total number of decision trees. Let D represent the prediction function of the k-th decision tree, and T represent the dry index and T represent the hot index. The parameters represent the k-th tree;

[0121] The analysis dataset of the past 10 years constructed in step S33 is divided into a training set and a test set in a 7:3 ratio.

[0122] Step S52: Train the random forest regression model. The coefficient of determination and root mean square error are used to evaluate the fitting effect of the random forest regression model. The test set R² ≥ 0.7 and RMSE ≤ 0.08 are required to obtain the trained random forest model, as shown in the following formula:

[0123] ; ;

[0124] in, The coefficient of determination is represented by the coefficient of determination. This indicates the number of test set samples in the analysis dataset. This represents the actual vegetation degradation rate in year i. This represents the predicted vegetation degradation rate for year i. This represents the average actual value of the vegetation degradation rate. Indicates the root mean square error;

[0125] Step S53: Output the feature importance ratio of dry and hot indices to vegetation degradation rate through the trained random forest model.

[0126] Step S54: Set an ecologically acceptable degradation threshold. Using the trained random forest model, generate a two-dimensional partial dependency graph of dry and hot indices. From the two-dimensional partial dependency graph, read the maximum values ​​of the dry and hot indices that satisfy the predicted vegetation degradation rate being less than or equal to the ecologically acceptable degradation threshold. Combine the maximum values ​​of the dry and hot indices with the feature importance ratio to obtain the ecological weighted constraint conditions, as shown in the following formula:

[0127] ;

[0128] in, This indicates the maximum value of the dry index. This indicates the proportion of the characteristic importance of dry indicators to the vegetation degradation rate. This indicates the maximum value of the thermal index. This indicates the percentage of the characteristic importance of thermal indicators to vegetation degradation rate;

[0129] Step S55: Assuming that the dryness index and the heat index follow a two-dimensional log-normal distribution, construct the joint probability density function as follows:

[0130] ;

[0131] in, Denotes the joint probability density function. Represents pi (π). This represents the standard deviation of the logarithmic series of the dry index. The standard deviation of the logarithmic series of heat indicators. This represents the correlation coefficient between the logarithmic series of dry and heat indicators. The square of, Represents an exponential function. Indicators The natural logarithm, This represents the mean of the logarithmic sequence of the dry index. The natural logarithm of the heat index T. This represents the mean of the logarithmic series of the heat index;

[0132] Step S56: Set the extreme joint event probability threshold P0=5%. Solve the constraint condition set by integrating the extreme joint event probability that satisfies the co-occurrence of the dry index and the thermal index in the joint probability density function to obtain the joint probability constraint condition, as shown in the following formula:

[0133] ;

[0134] in, Indicators Not less than the threshold And thermal index Not less than the threshold The probability of this extreme combined event occurring, The threshold representing the dry index, The threshold representing the thermal index, Indicates in and Within the research area, the joint probability density function Perform double integrals;

[0135] Step S57: Take the intersection of the ecological weighted constraint and the joint probability constraint, that is, the dry index critical value and the thermal index critical value that simultaneously satisfy the ecological weighted constraint and the joint probability constraint. The dry index critical value and the thermal index critical value are determined as the dry threshold and the thermal threshold, respectively.

[0136] Furthermore, in step S6, for each grid in the analysis dataset of the target year, it is determined whether both the dryness index and the heat index are simultaneously satisfied. If both are satisfied, it is determined that a dry-heat event has occurred in a single grid of the target year, and the percentage of grids experiencing dry-heat events in the target year is calculated; specifically:

[0137] Step S61: For each grid in the analysis dataset of the target year, call the number of dry days output by the hydrodynamic model and the number of high-temperature days. If both the dry index and the thermal index are greater than or equal to the dry threshold and the thermal index is greater than or equal to the thermal threshold, then it is determined that a dry-hot event has occurred in the target year for a single grid.

[0138] Step S62: Count the number of beach grids where hot and dry events occurred within the target year, and calculate the percentage of grids involved in hot and dry events, using the following formula:

[0139] ;

[0140] in, Indicates the grid percentage of dry and hot events. Indicates the number of grid cells on the shoal. This represents the total number of beach grids within the study area.

[0141] Furthermore, in step S7, based on the grid proportion of dry and hot events in historical years, a regional proportion threshold is determined through frequency analysis; if the grid proportion of dry and hot events in the target year is not lower than the regional proportion threshold, then the target year is determined to be a regional-level composite dry and hot event year; specifically:

[0142] Step S71: Based on the grid proportion of dry and hot events in a single target year, calculate the grid proportion of dry and hot events for each of the past 10 years to obtain the grid proportion of dry and hot events in historical years, using the following formula:

[0143] ;

[0144] in, The grid represents the percentage of dry and hot events in historical years. This indicates the percentage of dry and hot events in the grid during the first year. This indicates the percentage of dry and hot events in the grid in the second year. This indicates the percentage of grid-related dry and hot events in the tenth year;

[0145] Step S72: Frequency analysis is performed on the grid proportion of dry and hot events in historical years using Pearson type III curves to calculate the mean, coefficient of variation, and skewness coefficient, as shown in the following formulas:

[0146] , where n=10;

[0147] ;

[0148] ;

[0149] in, This represents the mean. Indicates the year. Represents the coefficient of variation. Indicates the skewness coefficient;

[0150] Step S73: Considering the extremes of the dry and hot thresholds, a guarantee rate of 95% is set. The mean and coefficient of variation obtained through frequency calculations using the Pearson type III curves are combined with the set guarantee rate to obtain the regional proportion threshold, as shown in the following formula:

[0151] ;

[0152] in, Indicates the threshold for the proportion of a region. This represents the deviation coefficient from the mean of the Pearson type III curve;

[0153] Step S74: If the grid proportion of dry and hot events in the target year is not lower than the regional proportion threshold, it is determined to be a regional composite dry and hot event year.

[0154] Example effect verification: The study area was set as follows: longitude 115°49′-116°46′E, latitude 28°24′-29°46′N, covering the main island and shoal areas of Poyang Lake; the time range was set as 2014 to 2023.

[0155] Taking 2022 as an example, the above method yielded a dry threshold of 45 days and a hot threshold of 12 days. The calculated grid proportion of dry and hot events was 18.5%, and the historical sequence analysis yielded a regional proportion threshold of 15.0%. Since the grid proportion of dry and hot events is greater than or equal to the regional proportion threshold, 2022 was determined to be a year of combined dry and hot events in the Poyang Lake floodplain wetlands. This result is highly consistent with the actual vegetation degradation and ecological disaster events recorded that year, verifying the accuracy and practicality of this method.

[0156] The embodiments described above are merely illustrative of implementation methods of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the protection scope of the present invention.

Claims

1. A method for identifying combined hot and dry events in floodplain wetlands, characterized in that, Includes the following steps: Step S1: Obtain research data within the study area, including topographic data, land use data, and year-end normalized vegetation index data; Step S2: Determine the extent of the sandbar based on topographic and land use data; construct a hydrodynamic model based on topographic and sandbar extent; determine the critical water level for sandbar exposure using the wet and dry water depth parameters of the hydrodynamic model; Step S3: Calculate the vegetation degradation rate based on the year-end normalized vegetation index data; standardize the dry and thermal indices of the hydrodynamic model and construct a gridded analysis dataset; Step S4: Use the hydrodynamic model to identify and analyze the dry and wet states of each grid in the dataset; count the number of consecutive dry days for each grid throughout the year as a dryness index, and count the number of high-temperature days during the dry period as a heat index; Step S5: Using standardized dry and heat indices as inputs and vegetation degradation rate as output, construct a random forest regression model; extract ecologically acceptable dry and heat constraint boundaries through the random forest regression model to obtain ecological weighted constraint conditions; simultaneously establish a joint probability distribution model of dry and heat indices, and combine it with preset extreme event probability thresholds to obtain joint probability constraint conditions; take the intersection of ecological weighted constraint conditions and joint probability constraint conditions to determine the dry threshold and heat threshold. Step S6: For each grid in the analysis dataset of the target year, determine whether the dry index is not lower than the dry threshold and the thermal index is not lower than the thermal threshold at the same time. If both are met, it is determined that a dry-thermal event has occurred in a single grid of the target year, and the proportion of grids with dry-thermal events in the target year is calculated. Step S7: Based on the grid proportion of dry and hot events in historical years, determine the proportion threshold within the region through frequency analysis; If the proportion of dry and hot events in the grid in the target year is not lower than the regional proportion threshold, then the target year is determined to be a regional-level compound dry and hot event year. In step S4, the hydrodynamic model is used to identify the dry and wet states of each grid in the dataset; the number of consecutive dry days for each grid throughout the year is counted as a dryness index, and the number of high-temperature days during the dry period is counted as a heat index. Specifically: Step S41: Use a hydrodynamic model to simulate the wetland inundation process and identify the dry and wet states of each grid in the dataset. Step S42: Run the hydrodynamic model on an annual basis with daily steps, and output the daily water level data for each grid of the analysis dataset throughout the year and the dry or wet state determination results. For each grid in the analysis dataset, the cumulative number of days that the hydrodynamic model classifies as dry throughout the year is counted, which is the dryness index; Step S43: Directly call the dry days output by the hydrodynamic model. Based on the daily dry or wet records output by the hydrodynamic model, determine all the periods of continuous dryness in each grid throughout the year. Perform spatiotemporal matching between the grid-level daily maximum temperature data and the above-mentioned dry periods, and count the cumulative number of days with daily maximum temperature ≥36℃ within the dry periods, which is the heat index.

2. The method for identifying combined hot and dry events in floodplain wetlands according to claim 1, characterized in that: In step S1, research data within the study area is acquired. This data includes topographic data, land use data, and year-end normalized vegetation index (NDI) data; the research data also includes hydrological and meteorological data. Specifically: Step S11: Select the floodplain wetland as the study area and divide the study area into grid cells with a resolution of 30m×30m; Step S12: Collect topographic data, hydrological data, meteorological data, land use data, and year-end vegetation index data of the floodplain wetland; Step S13: The terrain data is acquired through the National Geomatics Center of China to obtain a digital elevation model with a resolution of not less than 30m. Step S14: Hydrological data are acquired through the hydrological station network to obtain daily water level and flow data of the study area and upstream control stations; Step S15: Meteorological data is acquired through the National Meteorological Science Data Center to obtain daily maximum temperature data of meteorological stations within the study area and a ten-kilometer radius around it. Step S16: Land use data is acquired through the national geographic conditions monitoring cloud platform to obtain an annual land use map with a resolution of not less than 30m. The land use map is classified into cultivated land, forest land, grassland, construction land, water area, and unused land. Step S17: The year-end vegetation index data is extracted using the normalized vegetation index product of the medium resolution imaging spectrometer to obtain the vegetation index data for December of each year.

3. The method for identifying combined hot and dry events in floodplain wetlands according to claim 2, characterized in that: In step S2, the extent of the sandbar is determined based on topographic and land use data; a hydrodynamic model is constructed based on the topographic data and the extent of the sandbar; and the critical water level for sandbar exposure is determined using the wet and dry water depth parameters of the hydrodynamic model. Specifically: Step S21: Based on the digital elevation model determined by topographic data and the multi-year average low water level and flood level determined by hydrological data, determine the elevation range of the sandbar, and overlay the land use map to remove permanent water bodies and building land to obtain the sandbar range. Step S22: Using a two-dimensional simulation tool for physical, chemical, or biological processes in water bodies, a hydrodynamic model of the study area is constructed based on the digital elevation model and the extent of the sandbar. Step S23, determine the critical exposure water level using the dry water depth parameters in the hydrodynamic model: directly use the dry and wet water depth parameters built into the hydrodynamic model to set the critical exposure water level.

4. The method for identifying combined hot and dry events in floodplain wetlands according to claim 3, characterized in that: In step S3, the vegetation degradation rate is calculated based on the year-end normalized vegetation index data; the dry and thermal indices of the hydrodynamic model are standardized to construct a gridded analysis dataset; specifically: Step S31: Based on the year-end normalized vegetation index data, calculate the annual vegetation degradation rate for each grid in the analysis dataset, using the following formula: ； in, This represents the vegetation degradation rate in grid j of year i. This represents the maximum vegetation index at the end of the past 10 years for the j-th grid. This represents the year-end vegetation index value of the j-th grid in the i-th year; Step S32: Standardize the dry and thermal indices of the hydrodynamic model using standard fractions to obtain the standardized dry and thermal indices, as shown in the following formulas: ； in, This represents the standardized dry and thermal properties. This represents the raw values ​​of the dryness and heat indices. This represents the mean of the dryness index and the heat index. Indicates the standard deviation of dry and thermal indicators; Step S33: Construct an analysis dataset for the past 10 years using grids as units. The input features are standardized dry and thermal indicators, and the output variable is the vegetation degradation rate, resulting in a total of N×10 samples, where N represents the total number of grids on the beach.

5. The method for identifying combined hot and dry events in floodplain wetlands according to claim 4, characterized in that: In step S5, a random forest regression model is constructed using standardized dry and thermal indices as inputs and vegetation degradation rate as output. Ecologically acceptable dry and thermal constraint boundaries are extracted through the random forest regression model to obtain ecologically weighted constraints. Simultaneously, a joint probability distribution model of the dry and thermal indices is established, combined with preset extreme event probability thresholds, to obtain joint probability constraints. The intersection of the ecologically weighted constraints and the joint probability constraints is used to determine the dry and thermal thresholds. Specifically: Step S51: Using standardized dry and thermal indices as input features and vegetation degradation rate as output variable, construct a random forest regression model. Step S52: Train the random forest regression model. The coefficient of determination and root mean square error are used to evaluate the fitting effect of the random forest regression model, and the trained random forest model is obtained, as shown in the following formula: ； ； in, The coefficient of determination is represented by the coefficient of determination. This indicates the number of test set samples in the analysis dataset. This represents the actual vegetation degradation rate in year i. This represents the predicted vegetation degradation rate for year i. This represents the average actual value of the vegetation degradation rate. Indicates the root mean square error; Step S53: Output the feature importance ratio of dry and hot indices to vegetation degradation rate through the trained random forest model. Step S54: Set an ecologically acceptable degradation threshold. Using the trained random forest model, generate a two-dimensional partial dependency graph of dry and hot indices. From the two-dimensional partial dependency graph, read the maximum values ​​of the dry and hot indices that satisfy the predicted vegetation degradation rate being less than or equal to the ecologically acceptable degradation threshold. Combine the maximum values ​​of the dry and hot indices with the feature importance ratio to obtain the ecological weighted constraint conditions, as shown in the following formula: ； in, This indicates the maximum value of the dry index. This indicates the proportion of the characteristic importance of dry indicators to the vegetation degradation rate. This indicates the maximum value of the thermal index. This indicates the percentage of the characteristic importance of thermal indicators to vegetation degradation rate; Step S55: Assuming that the dryness index and the heat index follow a two-dimensional log-normal distribution, construct the joint probability density function as follows: ； in, Denotes the joint probability density function. Represents pi (π). Indicates the total number of decision trees. This represents the standard deviation of the logarithmic series of the dry index. The standard deviation of the logarithmic series of heat indicators. This represents the correlation coefficient between the logarithmic series of dry and heat indicators. The square of, Represents an exponential function. Indicators The natural logarithm, This represents the mean of the logarithmic sequence of the dry index. The natural logarithm of the heat index T. This represents the mean of the logarithmic series of the heat index; Step S56: Set the extreme joint event probability threshold P0=5%. Solve the constraint condition set by integrating the extreme joint event probability that satisfies the co-occurrence of the dry index and the thermal index in the joint probability density function to obtain the joint probability constraint condition, as shown in the following formula: ； in, Indicators Not less than the threshold And thermal index Not less than the threshold The probability of this extreme combined event occurring, The threshold representing the dry index, The threshold representing the thermal index, Indicates in and Within the research area, the joint probability density function Perform double integrals; Step S57: Take the intersection of the ecological weighted constraint and the joint probability constraint, that is, the dry index critical value and the thermal index critical value that simultaneously satisfy the ecological weighted constraint and the joint probability constraint. The dry index critical value and the thermal index critical value are determined as the dry threshold and the thermal threshold, respectively.

6. The method for identifying combined hot and dry events in floodplain wetlands according to claim 5, characterized in that: In step S6, for each grid in the analysis dataset of the target year, it is determined whether both the dryness index and the heat index are simultaneously satisfied. If both are satisfied, a dry-heat event is determined to have occurred in a single grid of the target year, and the percentage of grids experiencing dry-heat events in the target year is calculated. Specifically: Step S61: For each grid in the analysis dataset of the target year, call the number of dry days output by the hydrodynamic model and the number of high-temperature days. If both the dry index and the thermal index are greater than or equal to the dry threshold and the thermal index is greater than or equal to the thermal threshold, then it is determined that a dry-hot event has occurred in the target year for a single grid. Step S62: Count the number of beach grids where hot and dry events occurred within the target year, and calculate the percentage of grids involved in hot and dry events, using the following formula: ； in, Indicates the grid percentage of dry and hot events. Indicates the number of grid cells on the shoal. This represents the total number of beach grids within the study area.

7. The method for identifying combined hot and dry events in floodplain wetlands according to claim 6, characterized in that: In step S7, based on the grid proportion of dry and hot events in historical years, a regional proportion threshold is determined through frequency analysis; if the grid proportion of dry and hot events in the target year is not lower than the regional proportion threshold, then the target year is determined to be a regional-level complex dry and hot event year; specifically: Step S71: Based on the grid proportion of dry and hot events in a single target year, calculate the grid proportion of dry and hot events for each of the past 10 years to obtain the grid proportion of dry and hot events in historical years, using the following formula: ； in, The grid represents the percentage of dry and hot events in historical years. This indicates the percentage of dry and hot events in the grid during the first year. This indicates the percentage of dry and hot events in the grid in the second year. This indicates the percentage of grid-related dry and hot events in the tenth year; Step S72: Frequency analysis is performed on the grid proportion of dry and hot events in historical years using Pearson type III curves to calculate the mean, coefficient of variation, and skewness coefficient, as shown in the following formulas: where n=10; ； ； in, This represents the mean. Indicates the year. Represents the coefficient of variation. Indicates the skewness coefficient; Step S73: Combining the extremes of the dry threshold and the thermal threshold, the mean and coefficient of variation obtained by calculating the frequency of the Pearson type III curve, and the set guarantee rate, the regional proportion threshold is obtained, as shown in the following formula: ； in, Indicates the threshold for the proportion of a region. This represents the deviation coefficient from the mean of the Pearson type III curve; Step S74: If the grid proportion of dry and hot events in the target year is not lower than the regional proportion threshold, it is determined to be a regional composite dry and hot event year.

Images