A river-atmosphere coupled heat wave assessment method and system

Abstract

The present invention discloses a river-atmosphere coupled heat wave assessment method and system, including: obtaining continuous daily river surface water temperature data of multiple basins in a large range within a certain period of time; calculating the heat wave temperature threshold of the study area, calculating the basic characteristic indicators of river heat wave events and atmospheric heat wave events; calculating the relevant time characteristic indicators and coupling rate of river-atmosphere coupled heat wave events; evaluating the annual change trend of the basic characteristic indicators of river heat wave events and atmospheric heat wave events, the time characteristic indicators of river-atmosphere coupled heat wave events and the coupling rate, and evaluating the correlation between the basic characteristic indicators of river heat wave events and atmospheric heat wave events, the time characteristic indicators of river-atmosphere coupled heat wave events and the coupling rate. The present invention comprehensively evaluates the feedback response relationship between river heat wave and heat wave, more comprehensively understands and predicts the interaction between river and atmospheric heat wave, and provides an important reference for responding to climate change.

Description

Technical Field

[0001] The present invention belongs to the technical field of extreme weather and climate assessment, and in particular relates to a river-atmosphere coupled heat wave assessment method and system. Background Art

[0002] In recent years, as high temperature records have been constantly broken, the world has experienced frequent extreme weather and climate events. Heat waves - one of the dangerous consequences of climate change - are increasing in frequency and intensity at an alarming rate around the world. At present, relevant research on atmospheric heat waves and the chain reactions they cause has been widely carried out around the world, but there is no unified definition of atmospheric heat waves. Relevant research on water heat waves has also gradually begun. However, due to the scarcity and discontinuity of river water temperature data on a large scale, there is currently a lack of large-scale research on river heat waves.

[0003] The study of concurrent extreme climate events has received great attention from the academic community and the public. Using the joint probability distribution function, the study found that the probability of simultaneous drought and heat waves in the United States showed a significant increasing trend; the RU and UD surface observation meteorological values ​​and the CMIP5 simulation values ​​were used to compare and analyze the changing trends of simultaneous extreme precipitation and extreme high temperature events around the world. Compared with single atmospheric heat waves or river heat wave events, river-atmosphere coupled heat wave events are more complex and dangerous; rivers exchange heat with the atmosphere through processes such as evaporation, radiation and convection, thereby buffering the drastic changes in temperature. However, the occurrence of atmospheric heat waves may further trigger river heat waves. The increase in river water temperature will enhance evaporation and increase atmospheric humidity, thereby intensifying the intensity and duration of atmospheric heat waves, forming a vicious cycle. The combination of high temperature and high humidity caused by river-atmosphere coupled heat wave events may increase the incidence of heat-related diseases and pose a serious threat to human health. The focus of studying river-atmosphere coupled heat waves is to reveal the key role of rivers in regulating regional climate and how changes in river heat wave events respond to and feedback atmospheric heat wave events. In-depth research on the coupling mechanism between rivers and atmospheric heat waves is crucial for predicting and mitigating the impacts of climate change. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the purpose of the present invention is to provide a river-atmosphere coupled heat wave assessment method and system to solve the problem of scarce and discontinuous river water temperature data at a large regional scale in the prior art, and the problem of coupling of atmospheric heat wave events and river heat wave events under different extreme event definitions.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] A river-atmosphere coupled heat wave assessment method of the present invention comprises the following steps:

[0007] 1) Obtain daily continuous river surface water temperature data for multiple river basins over a large area within a set time;

[0008] 2) Calculate the heat wave temperature threshold in the study area, define river heat wave events and atmospheric heat wave events in the study area, and calculate the basic characteristic indicators of river heat wave events and atmospheric heat wave events;

[0009] 3) Couple river heat wave events with atmospheric heat wave events to define river-atmosphere coupled heat wave events; calculate the relevant temporal characteristic indicators and coupling rate of river-atmosphere coupled heat wave events;

[0010] 4) Evaluate the annual variation trends of the basic characteristic indicators of river heat wave events and atmospheric heat wave events in step 2) and the temporal characteristic indicators and coupling rates of river-atmosphere coupled heat wave events in step 3), and evaluate the correlation between the basic characteristic indicators of river heat wave events and atmospheric heat wave events in step 2) and the temporal characteristic indicators and coupling rates of river-atmosphere coupled heat wave events in step 3).

[0011] Furthermore, the step 1) specifically includes:

[0012] 11) Obtain meteorological and hydrological driving data, river surface water temperature observation data and basin attribute data for multiple basins in a large range within a set time;

[0013] 12) Preprocessing the data obtained in step 11), including deleting duplicate time series data, deleting erroneous records, deleting outliers and standardizing the data, to obtain preprocessed meteorological and hydrological driving data, river surface water temperature observation data and watershed attribute data; then converting the date of the data into time characteristic data; dividing the processed meteorological and hydrological driving data, river surface water temperature observation data, watershed attribute data and time characteristic data into a training set and a test set according to a certain ratio;

[0014] 13) Set the hyperparameters of the LSTM model, input the training set data processed in step 12) into the LSTM model for training, and update the model parameters. The goal is to minimize the loss function L;

[0015] 14) Input the test set data processed in step 12) into the LSTM model trained in step 13), and use the trained LSTM model to calculate and output the river surface water temperature prediction data layer by layer; after denormalization, the predicted data of river surface water temperature and the river surface water temperature observation data are processed to evaluate the model performance and obtain the daily continuous river surface water temperature data.

[0016] Furthermore, the meteorological and hydrological driving data in step 11) is time series data, including precipitation, snow water equivalent, surface air pressure, maximum temperature, minimum temperature, daily average temperature and wind speed component.

[0017] Furthermore, the river surface water temperature observation data is non-daily continuous time series data.

[0018] Furthermore, the watershed attribute data are classified into six main groups, including topography, climate, hydrology, land use conditions, soil and geological properties; specific data features include: the coordinate position of the watershed, altitude, watershed slope, annual snow cover ratio, average annual precipitation, average annual and monthly temperature, as well as flow and base flow index.

[0019] Furthermore, the data standardization process in step 12) specifically includes: performing Z-score standardization on the data, so that the data is converted into a distribution with a mean of 0 and a standard deviation of 1, and ensuring that the data has a similar scale and distribution; the calculation formula is as follows:

[0020] ;

[0021] Among them, x is the original data, μ is the average value of the original data, is the standard deviation of the original data, is the normalized data.

[0022] Furthermore, the step 13) specifically includes:

[0023] Set the hyperparameters of the LSTM model, including defining the total number of training cycles, model save frequency, batch size, time step, number of units in the LSTM hidden layer, and dropout rate;

[0024] The training set data processed in step 12) is input into the LSTM model for training, including the pre-processed meteorological and hydrological driving data, river surface water temperature observation data and watershed attribute data. During the training process, the model generates river surface water temperature simulation data according to the LSTM network forward propagation;

[0025] The model parameters are updated through the back-propagation algorithm, the goal is to minimize the loss function L, the expression of the loss function L is as follows:

[0026] ;

[0027] in, represents the simulated data of river surface water temperature, represents the river surface water temperature observation data, represents the variance of the river surface water temperature observations, Take 1×10−8 , n is the number of river surface water temperature observation data samples;

[0028] Calculate the gradient of the loss function L with respect to the model parameters θ ; The parameter θ includes all weight matrices and bias terms in the LSTM model;

[0029] Use the gradient descent method to update the model parameters. The update rule is:

[0030] ;

[0031] Where k represents the current number of iterations, θ(k) is the parameter value at the kth iteration, η is the learning rate, is the gradient of the loss function with respect to the parameter θ, and θ(k+1) is the updated parameter value;

[0032] The model parameters are gradually optimized through multiple iterations until the loss function converges to the minimum value or the predetermined number of training cycles is reached, completing the model training.

[0033] Furthermore, the step 14) specifically includes:

[0034] The pre-processed meteorological and hydrological driving data and watershed attribute data in the test set are input into the trained LSTM model; the trained LSTM model is used to calculate and output the river surface water temperature prediction data layer by layer; the specific process is as follows:

[0035] At each time step t, the trained LSTM model receives the input data , the forget gate is calculated to obtain the river surface water temperature forecast information of the previous time step that the model decides needs to retain under the current meteorological and hydrological conditions;

[0036] The calculation formula of the forget gate is as follows:

[0037] ;

[0038] in, represents the water temperature information retention rate, express function, represents the weight matrix of the forget gate, represents the predicted river surface water temperature at the previous time step, Represents the meteorological and hydrological driving data and watershed attribute data input at the current time step, Represents the bias term of the forget gate;

[0039] The input gate calculates the impact ratio of the meteorological and hydrological driving data and basin attribute data input at the current time step on the update of the river surface water temperature forecast state And the candidate unit status The value of

[0040] The input gate calculation formula is as follows:

[0041] ;

[0042] ;

[0043] in, represents the meteorological input impact rate, represents the weight matrix of the input gate, represents the bias vector of the input gate, Indicates the candidate unit status, represents the hyperbolic tangent activation function, The weight matrix representing the candidate unit state, A bias vector representing the candidate cell state;

[0044] Update the cumulative water temperature prediction information C of the current time step t , C t Including the meteorological and hydrological input of the current time step and the river surface water temperature forecast information of the previous time step, the calculation formula is as follows:

[0045] ;

[0046] in, and Respectively represent the cumulative water temperature prediction information of the current time step and the previous time step, represents element-wise multiplication;

[0047] By calculating the activation value of the output gate To control the information output ratio of the accumulated water temperature prediction information of the current time step, calculate the river surface water temperature prediction information of the current time step As output information, the calculation formula is as follows:

[0048] ;

[0049] ;

[0050] in, represents the activation value of the output gate, represents the river surface water temperature forecast information at the current time step, represents the weight matrix of the output gate, represents the bias vector of the output gate;

[0051] Prediction information of river surface water temperature at the last time step As the feature representation of the entire time series, it is passed to the output layer to output the river surface water temperature prediction data. The expression is as follows:

[0052] ;

[0053] in, represents the weight matrix of the output layer, represents the bias term, Represents the river surface water temperature prediction data output by the model.

[0054] Furthermore, in step 14), the predicted data of the river surface water temperature and the observed data of the river surface water temperature are denormalized to evaluate the model performance and obtain the daily continuous river surface water temperature data, which specifically includes the following steps:

[0055] The calculation formula for denormalization is as follows:

[0056] ;

[0057] Among them, Y represents the water temperature data after denormalization (water temperature restored to the real scale), represents the normalized water temperature data, represents the standard deviation of the water temperature data, It represents the average value of water temperature data, which includes river surface water temperature prediction data and river surface water temperature observation data;

[0058] The model was evaluated using the denormalized river surface temperature prediction data and river surface temperature observation data as follows:

[0059] ;

[0060] ;

[0061] ;

[0062] ;

[0063] In the formula, represents the predicted river surface water temperature data, represents the river surface water temperature observation data, n represents the total number of paired river surface water temperature prediction data and observation data during the model testing period, represents the average value of the predicted river surface water temperature data, represents the average value of the river surface water temperature observation data, represents the percentage deviation, represents the Nash efficiency coefficient, represents the root mean square error, Represents the Pearson correlation coefficient; after evaluating the reliability of the model output results based on the four evaluation parameters of Pbias, NSE, RMSE and Pcorr, the denormalized river surface water temperature prediction data is the daily continuous river surface water temperature data obtained.

[0064] Furthermore, the step 2) specifically includes:

[0065] 21) Calculate the heat wave temperature threshold for the study area based on the river surface water temperature data obtained in step 1) and the daily maximum temperature data used;

[0066] The daily temperature ensemble of the five days before and after the d-day of all years in the study area during the reference period is Sort by ascending order, The expression is:

[0067] ;

[0068] in, represents the daily temperature on the i-th day of the y-th year; represents the union of sets, represents the daily temperature set of 5 days before and after the d-th day in the year 1 to m;

[0069] The heat wave temperature threshold on day d is the daily temperature set The daily temperature corresponding to the 90th percentile after ascending order, the river heat wave temperature threshold and the atmospheric heat wave temperature threshold were calculated using the river surface water temperature and the daily maximum temperature as the daily temperature;

[0070] 22) Define river heat wave events and atmospheric heat wave events in the study area, and set basic parameters for heat wave identification, including: the maximum interval length allowed for atmospheric heat wave events and river heat wave events is 2 days, the shortest atmospheric heat wave duration is 3 days, and the shortest river heat wave duration is 5 days;

[0071] 23) Calculate the basic characteristic indicators of river heat wave events and atmospheric heat wave events, including heat wave frequency, heat wave duration and average heat wave intensity.

[0072] Furthermore, the identification condition for an atmospheric heat wave event in step 22) is: when the maximum temperature of the day is greater than the atmospheric heat wave temperature threshold for at least three consecutive days, it is recorded as an atmospheric heat wave event.

[0073] Furthermore, the identification condition for river heat wave events in step 22) is: when the river surface water temperature is greater than the river heat wave temperature threshold for at least five consecutive days, it is recorded as a river heat wave event.

[0074] Furthermore, in step 23), the heat wave frequency The total number of river heat waves or atmospheric heat wave events that occurred in each river basin within a set time period; for each river basin, retrieve the events that meet the heat wave event identification conditions within a given time period, sort the heat wave events in chronological order and mark them with serial numbers;

[0075] Duration of heat wave The number of days of river heat wave events or atmospheric heat wave events that occurred in each basin within a set period of time is calculated as follows:

[0076] ;

[0077] in, and represent the end date and start date of the ith heat wave event respectively;

[0078] Average intensity of heat waves avg The calculation method is:

[0079] ;

[0080] ;

[0081] HD is the duration of the heat wave, and HD is the duration of the river heat wave. , duration of atmospheric heat waves , It represents the abnormal value between the maximum temperature of each heat wave day and the 50% quantile threshold, HI cum is the cumulative intensity of heat wave events.

[0082] Furthermore, the step 3) specifically includes:

[0083] 31) Obtain the start and end dates of each river heat wave event and atmospheric heat wave event, create a search window for each river heat wave event based on the start and end dates of the river heat wave event, search for overlapping atmospheric heat wave events and river heat wave events on the search window and couple them, and define the matched coupled events as river-atmosphere coupled heat wave events;

[0084] 32) Based on the basic characteristic indicators of river heat wave events and atmospheric heat wave events calculated in step 2), calculate the relevant time characteristic indicators and coupling rate of river-atmosphere coupled heat wave events.

[0085] Furthermore, in step 31), the detect_event function is used to determine the start date tRhw,start and the end date tRhw,end of each river heat wave event;

[0086] A search window is created for each river heat wave event, and overlapping atmospheric heat wave events and river heat wave events are searched and coupled on the search window;

[0087] Define the start time tswb and end time tswe of the search window. The start time of the search window is defined as 120 hours (i.e. 5 days) before the start date of the river heat wave event. Specifically:

[0088] ;

[0089] The end time of the search window is defined as the end date of the river heat wave event, specifically:

[0090] ;

[0091] Compare whether the start date tAhw,start and the end date tAhw,end of the atmospheric heat wave event are within the search window [tswb,tswe] to determine whether the atmospheric heat wave event is coupled with the river heat wave event. The coupling condition is:

[0092] ;

[0093] or, .

[0094] All matching coupled events are recorded and stored in the event list as river-atmosphere coupled heat wave events; if no overlapping atmospheric heat wave events are found, the river heat wave events are defined as independent events.

[0095] Furthermore, the relevant time characteristic indicators of the river-atmosphere coupled heat wave event in step 32) include:

[0096] Coupled river heatwave duration The sum of the duration of river heat waves in all river-atmosphere coupled heat wave events in each basin within a set time period is expressed as:

[0097] ;

[0098] Where N is the total number of river-atmosphere coupled heat wave events occurring within a set time, tro,i represents the start time of the river heat wave in the i-th coupled event, and tre,i represents the end time of the river heat wave in the i-th coupled event;

[0099] Coupled atmospheric heatwave duration is the sum of the duration of atmospheric heatwaves in all river-atmosphere coupled heatwave events within a set period for each basin;

[0100] When a river heat wave event in the search window is coupled with multiple atmospheric heat wave events, the coupled atmospheric heat wave time is the sum of all coupled atmospheric heat wave events in the window, expressed as:

[0101] ;

[0102] Where M represents the number of atmospheric heat wave events coupled to the ith river-atmosphere coupled heat wave event, tao,ij represents the start time of the jth atmospheric heat wave event coupled to the ith river-atmosphere coupled heat wave event, and tae,ij represents the end time of the jth atmospheric heat wave event coupled to the ith river-atmosphere coupled heat wave event. represents the sum of the durations of all atmospheric heat wave events in the i-th river-atmosphere coupled heat wave event;

[0103] Coupled heat wave response lag time The sum of the time distances from the start time of the atmospheric heat wave to the start time of the river heat wave in each river-atmosphere coupled heat wave event within a set time for each basin;

[0104] When a river heat wave event is coupled with multiple atmospheric heat wave events, the start time of the first coupled atmospheric heat wave event is selected for calculation, and the expression is:

[0105] ;

[0106] in, represents the starting time of the first atmospheric heat wave event coupled in the ith river-atmosphere coupled heat wave event;

[0107] Coupled heat wave time difference The sum of the time difference between the duration of coupled river heat waves and the duration of coupled atmospheric heat waves in each river-atmosphere coupled heat wave event in each basin within a set time period is expressed as:

[0108] .

[0109] Furthermore, the coupling rate of the river-atmosphere coupled heat wave event is the ratio of the number of river-atmosphere coupled heat wave events occurring in each basin within a set time to the number of river heat wave events. If no river heat wave event occurs in the basin within the set time, the coupling rate is calculated as NA. The expression is:

[0110] ;

[0111] Among them, CR represents the coupling rate of the coupling event in the basin within the set time, represents the number of river-atmosphere coupled heatwave events that occurred in the basin within a set time period. represents the number of riverine heatwave events that occurred in the basin over a set period of time; When , the coupling ratio CR is NA.

[0112] Furthermore, in step 4), the basic characteristic indicators of the river heat wave events and the atmospheric heat wave events in step 2) and the annual variation trend of the time indicators and coupling rate of the river-atmosphere coupled heat wave events in step 3) are evaluated based on linear regression and Theil Sens slope method. The annual variation trend is expressed in absolute value and percentage value respectively, and the percentage trend change formula is:

[0113] ;

[0114] Among them, C End and C Start Respectively represent the average of the end date and the start date, Nyears represents the number of years of the time series span, Indicates the annual percentage change.

[0115] Furthermore, in step 4), the Pearson correlation coefficient is used to evaluate the correlation between the basic characteristic indicators of the river heat wave event and the atmospheric heat wave event in step 2) and the temporal characteristic indicators and coupling rate of the river-atmosphere coupled heat wave event in step 3);

[0116] The Pearson correlation coefficient indicates the strength and direction of the linear relationship between two variables. The Pearson correlation coefficient calculation formula is as follows:

[0117] ;

[0118] Where r is the Pearson correlation coefficient, and are the observed values ​​of two variables, and is the mean of the two variables, and the Pearson correlation coefficient ranges from -1 to 1, where -1 indicates a perfect negative correlation, 1 indicates a perfect positive correlation, and 0 indicates no correlation.

[0119] The present invention also provides a river-atmosphere coupled heat wave assessment system, comprising: a data acquisition module, a first calculation module, an event coupling module, a second calculation module, a first assessment module and a second assessment module;

[0120] A data acquisition module is used to obtain river water temperature data of multiple river basins in a large range within a set time;

[0121] The first calculation module is used to calculate the heat wave temperature threshold of the study area, define the river heat wave events and atmospheric heat wave events in the study area, and calculate the basic characteristic indicators of the river heat wave events and atmospheric heat wave events;

[0122] An event coupling module is used to couple river heat wave events with atmospheric heat wave events to define river-atmosphere coupled heat wave events;

[0123] The second calculation module is used to calculate the relevant time characteristic indicators and coupling rate of river-atmosphere coupled heat wave events;

[0124] The first evaluation module is used to evaluate the basic characteristic indicators of river heat wave events and atmospheric heat wave events, the temporal characteristic indicators of river-atmosphere coupled heat wave events, and the annual variation trend of the coupling rate;

[0125] The second evaluation module is used to evaluate the correlation between the basic characteristic indicators of river heat wave events and atmospheric heat wave events, the temporal characteristic indicators of river-atmosphere coupled heat wave events, and the coupling rate.

[0126] Beneficial effects of the present invention:

[0127] The present invention solves the problem of scarcity and discontinuity of river water temperature data through the LSTM model, making data acquisition and processing more efficient and accurate.

[0128] Through river-atmosphere heat wave coupling, the present invention quantifies the changes in the temporal characteristic indicators of coupled events and uncoupled events, analyzes the dynamic trends of the temporal characteristic indicators of coupled events and the coupling rate, and reveals the correlation between these trends and the characteristic indicators of a single heat wave event.

[0129] The present invention comprehensively evaluates the feedback response relationship between river heat waves and heat waves, which can more comprehensively understand and predict the interaction between rivers and atmospheric heat waves, and provide an important reference for addressing climate change. BRIEF DESCRIPTION OF THE DRAWINGS

[0130] Figure 1 Schematic diagram of the method of the present invention. DETAILED DESCRIPTION

[0131] In order to facilitate the understanding of those skilled in the art, the present invention is further described below in conjunction with embodiments and drawings. The contents mentioned in the implementation modes are not intended to limit the present invention.

[0132] Reference Figure 1 As shown, a river-atmosphere coupled heat wave assessment method of the present invention comprises the following steps:

[0133] 1) Obtain daily continuous river surface water temperature data for multiple river basins over a large area within a set time; specifically including:

[0134] 11) Obtain meteorological and hydrological driving data, river surface water temperature observation data and basin attribute data for multiple basins in a large range within a set time;

[0135] 12) Preprocessing the data obtained in step 11), including deleting duplicate time series data, deleting erroneous records, deleting outliers and standardizing the data, to obtain preprocessed meteorological and hydrological driving data, river surface water temperature observation data and watershed attribute data; then converting the date of the data into time characteristic data; dividing the processed meteorological and hydrological driving data, river surface water temperature observation data, watershed attribute data and time characteristic data into a training set and a test set according to a certain ratio;

[0136] 13) Set the hyperparameters of the LSTM model, input the training set data processed in step 12) into the LSTM model for training, and update the model parameters. The goal is to minimize the loss function L;

[0137] 14) Input the test set data processed in step 12) into the LSTM model trained in step 13), and use the trained LSTM model to calculate and output the river surface water temperature prediction data layer by layer; after denormalization, the predicted data of river surface water temperature and the river surface water temperature observation data are processed to evaluate the model performance and obtain the daily continuous river surface water temperature data.

[0138] The meteorological and hydrological driving data in step 11) are time series data, including precipitation, snow water equivalent, surface air pressure, maximum temperature, minimum temperature, daily average temperature and wind speed component.

[0139] The river surface water temperature observation data is non-daily continuous time series data.

[0140] The watershed attribute data are classified into six main groups, including topography, climate, hydrology, land use conditions, soil and geological properties; specific data features include: the coordinate position of the watershed, altitude, watershed slope, annual snow cover ratio, average annual precipitation, average annual and monthly temperature, as well as flow and base flow index.

[0141] The data standardization process in step 12) specifically includes: performing Z-score standardization on the data, so that the data is converted into a distribution with a mean of 0 and a standard deviation of 1, and ensuring that the data has a similar scale and distribution; the calculation formula is as follows:

[0142] ;

[0143] Among them, x is the original data, μ is the average value of the original data, is the standard deviation of the original data, is the normalized data.

[0144] Wherein, the step 13) specifically includes:

[0145] Set the hyperparameters of the LSTM model, including defining the total number of training cycles, model save frequency, batch size, time step, number of units in the LSTM hidden layer, and dropout rate;

[0146] The training set data processed in step 12) is input into the LSTM model for training, including the pre-processed meteorological and hydrological driving data (such as precipitation, snow water equivalent, average temperature, maximum temperature, etc.), river surface water temperature observation data and watershed attribute data. During the training process, the model generates river surface water temperature simulation data according to the LSTM network forward propagation;

[0147] The model parameters are updated through the back-propagation algorithm, the goal is to minimize the loss function L, the expression of the loss function L is as follows:

[0148] ;

[0149] in, represents the simulated data of river surface water temperature, represents the river surface water temperature observation data, represents the variance of the river surface water temperature observations, Take 1×10 −8 , n is the number of river surface water temperature observation data samples;

[0150] Calculate the gradient of the loss function L with respect to the model parameters θ , the parameter θ includes all weight matrices and bias terms in the LSTM model;

[0151] Use the gradient descent method to update the model parameters. The update rule is:

[0152] ;

[0153] Where k represents the current number of iterations, θ(k) is the parameter value at the kth iteration, η is the learning rate, is the gradient of the loss function with respect to the parameter θ, and θ(k+1) is the updated parameter value;

[0154] The model parameters are gradually optimized through multiple iterations until the loss function converges to the minimum value or the predetermined number of training cycles is reached, completing the model training.

[0155] Wherein, the step 14) specifically includes:

[0156] The pre-processed meteorological and hydrological driving data and watershed attribute data in the test set are input into the trained LSTM model; the trained LSTM model is used to calculate and output the river surface water temperature prediction data layer by layer; the specific process is as follows:

[0157] At each time step t, the trained LSTM model receives the input data , the forget gate is calculated to obtain the river surface water temperature forecast information of the previous time step that the model decides needs to retain under the current meteorological and hydrological conditions;

[0158] The calculation formula of the forget gate is as follows:

[0159] ;

[0160] in, represents the water temperature information retention rate, express function, represents the weight matrix of the forget gate, represents the predicted river surface water temperature at the previous time step, Represents the meteorological and hydrological driving data and watershed attribute data input at the current time step, Represents the bias term of the forget gate;

[0161] The input gate calculates the impact ratio of the meteorological and hydrological driving data and basin attribute data input at the current time step on the update of the river surface water temperature forecast state And the candidate unit status The value of

[0162] The input gate calculation formula is as follows:

[0163] ;

[0164] ;

[0165] in, represents the meteorological input impact rate, represents the weight matrix of the input gate, represents the bias vector of the input gate, Indicates the candidate unit status, represents the hyperbolic tangent activation function, The weight matrix representing the candidate unit state, A bias vector representing the candidate cell state;

[0166] Update the cumulative water temperature prediction information C of the current time step t , C t Including the meteorological and hydrological input of the current time step and the river surface water temperature forecast information of the previous time step, the calculation formula is as follows:

[0167] ;

[0168] in, and Respectively represent the cumulative water temperature prediction information of the current time step and the previous time step, represents element-wise multiplication;

[0169] By calculating the activation value of the output gate To control the information output ratio of the accumulated water temperature prediction information of the current time step, calculate the river surface water temperature prediction information of the current time step As output information, the calculation formula is as follows:

[0170] ;

[0171] ;

[0172] in, represents the activation value of the output gate, represents the river surface water temperature forecast information at the current time step, represents the weight matrix of the output gate, represents the bias vector of the output gate;

[0173] Prediction information of river surface water temperature at the last time step As the feature representation of the entire time series, it is passed to the output layer to output the river surface water temperature prediction data. The expression is as follows:

[0174] ;

[0175] in, represents the weight matrix of the output layer, represents the bias term, Represents the river surface water temperature prediction data output by the model.

[0176] In the step 14), the predicted data of the river surface water temperature and the observed data of the river surface water temperature are denormalized to evaluate the model performance and obtain the continuous daily river surface water temperature data, which specifically includes the following steps:

[0177] The calculation formula for denormalization is as follows:

[0178] ;

[0179] Among them, Y represents the water temperature data after denormalization (water temperature restored to the real scale), represents the normalized water temperature data, represents the standard deviation of the water temperature data, It represents the average value of water temperature data, which includes river surface water temperature prediction data and river surface water temperature observation data;

[0180] The model was evaluated using the denormalized river surface temperature prediction data and river surface temperature observation data as follows:

[0181] ;

[0182] ;

[0183] ;

[0184] ;

[0185] In the formula, represents the predicted river surface water temperature data, represents the river surface water temperature observation data, n represents the total number of paired river surface water temperature prediction data and observation data during the model testing period, represents the average value of the predicted river surface water temperature data, represents the average value of the river surface water temperature observation data, represents the percentage deviation, represents the Nash efficiency coefficient, represents the root mean square error, Represents the Pearson correlation coefficient; after evaluating the reliability of the model output results based on the four evaluation parameters of Pbias, NSE, RMSE and Pcorr, the denormalized river surface water temperature prediction data is the daily continuous river surface water temperature data obtained.

[0186] 2) Calculate the heat wave temperature threshold of the study area, define the river heat wave events and atmospheric heat wave events in the study area, and calculate the basic characteristic indicators of river heat wave events and atmospheric heat wave events; specifically include:

[0187] 21) Calculate the heat wave temperature threshold for the study area based on the river surface water temperature data obtained in step 1) and the daily maximum temperature data used;

[0188] The daily temperature ensemble of the five days before and after the d-day of all years in the study area during the reference period is Sort by ascending order, The expression is:

[0189] ;

[0190] in, represents the daily temperature on the i-th day of the y-th year; represents the union of sets, represents the daily temperature set of 5 days before and after the d-th day in the year 1 to m;

[0191] The heat wave temperature threshold on day d is the daily temperature set The daily temperature corresponding to the 90th percentile after ascending order, the river heat wave temperature threshold and the atmospheric heat wave temperature threshold were calculated using the river surface water temperature and the daily maximum temperature as the daily temperature;

[0192] 22) Define river heat wave events and atmospheric heat wave events in the study area, and set basic parameters for heat wave identification, including: the maximum interval length allowed for atmospheric heat wave events and river heat wave events is 2 days, the shortest atmospheric heat wave duration is 3 days, and the shortest river heat wave duration is 5 days;

[0193] 23) Calculate the basic characteristic indicators of river heat wave events and atmospheric heat wave events, including heat wave frequency, heat wave duration and average heat wave intensity.

[0194] Among them, the identification condition for atmospheric heat wave events in step 22) is: when the maximum temperature of the day is greater than the atmospheric heat wave temperature threshold for at least three consecutive days, it is recorded as an atmospheric heat wave event.

[0195] The identification condition for river heat wave events in step 22) is: when the river surface water temperature is greater than the river heat wave temperature threshold for at least five consecutive days, it is recorded as a river heat wave event.

[0196] Wherein, in step 23), the heat wave frequency The total number of river heat waves or atmospheric heat wave events that occurred in each river basin within a set time period; for each river basin, retrieve the events that meet the heat wave event identification conditions within a given time period, sort the heat wave events in chronological order and mark them with serial numbers;

[0197] Heat wave duration The number of days of river heat wave events or atmospheric heat wave events that occurred in each basin within a set period of time is calculated as follows:

[0198] ;

[0199] in, and represent the end date and start date of the ith heat wave event respectively;

[0200] Average intensity of heat waves avg The calculation method is:

[0201] ;

[0202] ;

[0203] HD is the duration of the heat wave, and HD is the duration of the river heat wave. , duration of atmospheric heat waves , It represents the abnormal value between the maximum temperature of each heat wave day and the 50% quantile threshold, HI cum is the cumulative intensity of heat wave events.

[0204] 3) Couple river heat wave events with atmospheric heat wave events to define river-atmosphere coupled heat wave events; calculate relevant temporal characteristic indicators and coupling rates of river-atmosphere coupled heat wave events; specifically include:

[0205] 31) Obtain the start and end dates of each river heat wave event and atmospheric heat wave event, create a search window for each river heat wave event based on the start and end dates of the river heat wave event, search for overlapping atmospheric heat wave events and river heat wave events on the search window and couple them, and define the matched coupled events as river-atmosphere coupled heat wave events;

[0206] 32) Based on the basic characteristic indicators of river heat wave events and atmospheric heat wave events calculated in step 2), calculate the relevant time characteristic indicators and coupling rate of river-atmosphere coupled heat wave events.

[0207] Wherein, in said step 31), the detect_event function is used to determine the start date tRhw,start and the end date tRhw,end of each river heat wave event;

[0208] A search window is created for each river heat wave event, and overlapping atmospheric heat wave events and river heat wave events are searched and coupled in the search window;

[0209] Define the start time tswb and end time tswe of the search window. The start time of the search window is defined as 120 hours (i.e. 5 days) before the start date of the river heat wave event. Specifically:

[0210] ;

[0211] The end time of the search window is defined as the end date of the river heat wave event, specifically:

[0212] ;

[0213] Compare whether the start date tAhw,start and the end date tAhw,end of the atmospheric heat wave event are within the search window [tswb,tswe] to determine whether the atmospheric heat wave event is coupled with the river heat wave event. The coupling condition is:

[0214] ;

[0215] or, .

[0216] All matching coupled events are recorded and stored in the event list as river-atmosphere coupled heat wave events; if no overlapping atmospheric heat wave events are found, the river heat wave events are defined as independent events.

[0217] The relevant time characteristic indicators of the river-atmosphere coupled heat wave event in step 32) include:

[0218] Coupled river heatwave duration The sum of the duration of river heat waves in all river-atmosphere coupled heat wave events in each basin within a set time period is expressed as:

[0219] ;

[0220] Where N is the total number of river-atmosphere coupled heat wave events occurring within a set time, tro,i represents the start time of the river heat wave in the i-th coupled event, and tre,i represents the end time of the river heat wave in the i-th coupled event;

[0221] Coupled atmospheric heatwave duration is the sum of the duration of atmospheric heatwaves in all river-atmosphere coupled heatwave events within a set period for each basin;

[0222] When a river heat wave event in the search window is coupled with multiple atmospheric heat wave events, the coupled atmospheric heat wave time is the sum of all coupled atmospheric heat wave events in the window, expressed as:

[0223] ;

[0224] Where M represents the number of atmospheric heat wave events coupled to the ith river-atmosphere coupled heat wave event, tao,ij represents the start time of the jth atmospheric heat wave event coupled to the ith river-atmosphere coupled heat wave event, and tae,ij represents the end time of the jth atmospheric heat wave event coupled to the ith river-atmosphere coupled heat wave event. represents the sum of the durations of all atmospheric heat wave events in the i-th river-atmosphere coupled heat wave event;

[0225] Coupled heat wave response lag time The sum of the time distances from the start time of the atmospheric heat wave to the start time of the river heat wave in each river-atmosphere coupled heat wave event within a set time for each basin;

[0226] When a river heat wave event is coupled with multiple atmospheric heat wave events, the start time of the first coupled atmospheric heat wave event is selected for calculation, and the expression is:

[0227] ;

[0228] in, represents the starting time of the first atmospheric heat wave event coupled in the ith river-atmosphere coupled heat wave event;

[0229] Coupled heat wave time difference The sum of the time difference between the duration of coupled river heat waves and the duration of coupled atmospheric heat waves in each river-atmosphere coupled heat wave event in each basin within a set time period is expressed as:

[0230] .

[0231] The coupling rate of the river-atmosphere coupled heat wave event is the ratio of the number of river-atmosphere coupled heat wave events occurring in each basin within a set time to the number of river heat wave events. If no river heat wave event occurs in the basin within the set time, the coupling rate is calculated as NA. The expression is:

[0232] ;

[0233] Among them, CR represents the coupling rate of the coupling event in the basin within the set time, represents the number of river-atmosphere coupled heatwave events that occurred in the basin within a set time period. represents the number of riverine heatwave events that occurred in the basin over a set period of time; When , the coupling ratio CR is NA.

[0234] 4) Evaluate the annual variation trend of the basic characteristic indicators of river heat wave events and atmospheric heat wave events in step 2) and the temporal characteristic indicators and coupling rate of river-atmosphere coupled heat wave events in step 3), and evaluate the correlation between the basic characteristic indicators of river heat wave events and atmospheric heat wave events in step 2) and the temporal characteristic indicators and coupling rate of river-atmosphere coupled heat wave events in step 3);

[0235] Among them, in step 4), the basic characteristic indicators of the river heat wave events and the atmospheric heat wave events in step 2) and the annual change trend of the time indicators and coupling rate of the river-atmosphere coupled heat wave events in step 3) are evaluated based on linear regression and Theil Sens slope method. The annual change trend is expressed in absolute value and percentage value respectively, and the percentage trend change formula is:

[0236] ;

[0237] Among them, C End and C Start Respectively represent the average of the end date and the start date, Nyears represents the number of years of the time series span, Indicates the annual percentage change.

[0238] Wherein, in step 4), the Pearson correlation coefficient is used to evaluate the correlation between the basic characteristic indicators of the river heat wave event and the atmospheric heat wave event in step 2) and the temporal characteristic indicators and coupling rate of the river-atmosphere coupled heat wave event in step 3);

[0239] The Pearson correlation coefficient indicates the strength and direction of the linear relationship between two variables. The Pearson correlation coefficient calculation formula is as follows:

[0240] ;

[0241] Where r is the Pearson correlation coefficient, and are the observed values ​​of two variables, and is the mean of the two variables, and the Pearson correlation coefficient ranges from -1 to 1, where -1 indicates a perfect negative correlation, 1 indicates a perfect positive correlation, and 0 indicates no correlation.

[0242] The present invention also provides a river-atmosphere coupled heat wave assessment system, comprising: a data acquisition module, a first calculation module, an event coupling module, a second calculation module, a first assessment module and a second assessment module;

[0243] A data acquisition module is used to obtain river water temperature data of multiple river basins in a large range within a set time;

[0244] The first calculation module is used to calculate the heat wave temperature threshold of the study area, define the river heat wave events and atmospheric heat wave events in the study area, and calculate the basic characteristic indicators of the river heat wave events and atmospheric heat wave events;

[0245] An event coupling module is used to couple river heat wave events with atmospheric heat wave events to define river-atmosphere coupled heat wave events;

[0246] The second calculation module is used to calculate the relevant time characteristic indicators and coupling rate of river-atmosphere coupled heat wave events;

[0247] The first evaluation module is used to evaluate the basic characteristic indicators of river heat wave events and atmospheric heat wave events, the temporal characteristic indicators of river-atmosphere coupled heat wave events, and the annual variation trend of the coupling rate;

[0248] The second evaluation module is used to evaluate the correlation between the basic characteristic indicators of river heat wave events and atmospheric heat wave events, the temporal characteristic indicators of river-atmosphere coupled heat wave events, and the coupling rate.

[0249] The present invention has many specific application paths. The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements can be made without departing from the principle of the present invention. These improvements should also be regarded as the protection scope of the present invention.

Claims

1. A river-atmosphere coupled heat wave assessment method, characterized in that: Here are the steps: 1) Obtain daily continuous river surface water temperature data for multiple river basins over a large area within a set time; 2) Calculate the heat wave temperature threshold of the study area, define the river heat wave events and atmospheric heat wave events in the study area, and calculate the basic characteristic indicators of river heat wave events and atmospheric heat wave events; 3) Couple river heat wave events with atmospheric heat wave events to define river-atmosphere coupled heat wave events; Calculate relevant temporal characteristic indicators and coupling rates of river-atmosphere coupled heat wave events; 4) Evaluate the annual variation trend of the basic characteristic indicators of river heat wave events and atmospheric heat wave events in step 2) and the temporal characteristic indicators and coupling rate of river-atmosphere coupled heat wave events in step 3), and evaluate the correlation between the basic characteristic indicators of river heat wave events and atmospheric heat wave events in step 2) and the temporal characteristic indicators and coupling rate of river-atmosphere coupled heat wave events in step 3); The step 1) specifically includes: 11) Obtain meteorological and hydrological driving data, river surface water temperature observation data and basin attribute data for multiple basins in a large range within a set time; 12) Preprocessing the data obtained in step 11), including deleting duplicate time series data, deleting erroneous records, deleting outliers and standardizing the data, to obtain preprocessed meteorological and hydrological driving data, river surface water temperature observation data and watershed attribute data; then converting the date of the data into time characteristic data; dividing the processed meteorological and hydrological driving data, river surface water temperature observation data, watershed attribute data and time characteristic data into a training set and a test set according to a certain ratio; 13) Setting hyperparameters for the LSTM model, inputting the training set data processed in step 12) into the LSTM model for training, and updating the model parameters, with the goal of minimizing the loss function L; 14) inputting the test set data processed in step 12) into the LSTM model trained in step 13), and outputting the predicted river surface water temperature data layer by layer through the trained LSTM model; evaluating the model performance after denormalizing the predicted river surface water temperature data and the observed river surface water temperature data, and obtaining the continuous daily river surface water temperature data; The step 2) specifically includes: 21) Calculate the heat wave temperature threshold of the study area based on the river surface water temperature data obtained in step 1) and the daily maximum temperature data used; The daily temperature ensemble T of the five days before and after the d-day of all years in the study area during the reference period is d In ascending order, T d The expression is: Among them, T y,i represents the daily temperature on the i-th day of the y-th year; ∪ represents the union of sets, represents the daily temperature set of 5 days before and after the d-th day in the year 1 to m; The heat wave temperature threshold on day d is the daily temperature set T d The daily temperature corresponding to the 90th percentile after ascending order, the river heat wave temperature threshold and the atmospheric heat wave temperature threshold were calculated using the river surface water temperature and the daily maximum temperature as the daily temperature; 22) Define river heat wave events and atmospheric heat wave events in the study area, and set basic parameters for heat wave identification, including: the maximum interval length allowed for atmospheric heat wave events and river heat wave events is 2 days, the shortest atmospheric heat wave duration is 3 days, and the shortest river heat wave duration is 5 days; 23) Calculate the basic characteristic indicators of river heat wave events and atmospheric heat wave events, including heat wave frequency, heat wave duration and average heat wave intensity; The step 3) specifically includes: 31) Obtain the start date and end date of each river heat wave event and atmospheric heat wave event, create a search window for each river heat wave event based on the start and end dates of the river heat wave event, search for overlapping atmospheric heat wave events and river heat wave events on the search window and couple them, and define the matched coupled events as river-atmosphere coupled heat wave events; 32) Calculating the relevant time characteristic indicators and coupling rate of river-atmosphere coupled heat wave events based on the basic characteristic indicators of river heat wave events and atmospheric heat wave events calculated in step 2); In the step 31), the detect_event function is used to determine the start date tRhw,start and the end date tRhw,end of each river heat wave event; A search window is created for each river heat wave event, and overlapping atmospheric heat wave events and river heat wave events are searched and coupled on the search window; Define the start time tswb and end time tswe of the search window. The start time of the search window is defined as 120 hours before the start date of the river heat wave event, specifically: tswb = tRhw, start - 120 hours; The end time of the search window is defined as the end date of the river heat wave event, specifically: tswe=tRhw,end; Compare whether the start date tAhw,start and the end date tAhw,end of the atmospheric heat wave event are within the search window [tswb,tswe] to determine whether the atmospheric heat wave event is coupled with the river heat wave event. The coupling condition is: tswb≤tAhw,start≤tswe or tswb≤tAhw,end≤tswe; Or, tAhw,start≤tswe and tAhw,end≥tswb; All matching coupled events are recorded and stored in the event list as river-atmosphere coupled heat wave events; if no overlapping atmospheric heat wave events are found, the river heat wave events are defined as independent events; The relevant time characteristic indicators of the river-atmosphere coupled heat wave event in step 32) include: The coupled river heat wave duration Tr is the sum of the river heat wave durations in all river-atmosphere coupled heat wave events in each basin within a set time, and is expressed as: Where N is the total number of river-atmosphere coupled heat wave events occurring within a set time, tro,i represents the start time of the river heat wave in the i-th coupled event, and tre,i represents the end time of the river heat wave in the i-th coupled event; The coupled atmospheric heat wave duration Ta is the sum of the durations of atmospheric heat waves in all river-atmosphere coupled heat wave events within a set period for each basin; When a river heat wave event in the search window is coupled with multiple atmospheric heat wave events, the coupled atmospheric heat wave time is the sum of all coupled atmospheric heat wave events in the window, expressed as: Where M represents the number of atmospheric heat wave events coupled in the ith river-atmosphere coupled heat wave event, tao,ij represents the start time of the jth atmospheric heat wave event coupled in the ith river-atmosphere coupled heat wave event, and tae,ij represents the end time of the jth atmospheric heat wave event coupled in the ith river-atmosphere coupled heat wave event. represents the sum of the durations of all atmospheric heat wave events in the i-th river-atmosphere coupled heat wave event; Coupled heat wave response lag time T LR The sum of the time distances from the start time of the atmospheric heat wave to the start time of the river heat wave in each river-atmosphere coupled heat wave event within a set time for each basin; When a river heat wave event is coupled with multiple atmospheric heat wave events, the start time of the first coupled atmospheric heat wave event is selected for calculation, and the expression is: Wherein, tao,i1 represents the starting time of the first atmospheric heat wave event coupled in the i-th river-atmosphere coupled heat wave event; Coupled heat wave time difference T HD The sum of the time difference between the duration of coupled river heat waves and the duration of coupled atmospheric heat waves in each river-atmosphere coupled heat wave event in each basin within a set time period is expressed as:

2. The method according to claim 1, characterized in that The step 13) specifically includes: Set hyperparameters for the LSTM model, including defining the total number of training cycles, model save frequency, batch size, time step, number of units in the LSTM hidden layer, and dropout rate; The training set data processed in step 12) is input into the LSTM model for training, including the pre-processed meteorological and hydrological driving data, river surface water temperature observation data and watershed attribute data. During the training process, the model generates river surface water temperature simulation data according to the LSTM network forward propagation; The model parameters are updated through the back-propagation algorithm, the goal is to minimize the loss function L, the expression of the loss function L is as follows: Among them, y m represents the simulated data of river surface water temperature, y obs represents the river surface water temperature observation data, std(y obs ) 2 represents the variance of the river surface water temperature observations, ε is 1×10 -8 , n is the number of river surface water temperature observation data samples; Calculate the gradient of the loss function L with respect to the model parameters θ The parameter θ includes all weight matrices and bias terms in the LSTM model; Use the gradient descent method to update the model parameters. The update rule is: Where k represents the current number of iterations, θ(k) is the parameter value at the kth iteration, η is the learning rate, is the gradient of the loss function with respect to the parameter θ, and θ(k+1) is the updated parameter value; The model parameters are gradually optimized through multiple iterations until the loss function converges to the minimum value or the predetermined number of training cycles is reached, completing the model training.

3. The method according to claim 2, characterized in that The step 14) specifically includes: The pre-processed meteorological and hydrological driving data and watershed attribute data in the test set are input into the trained LSTM model; the trained LSTM model is used to calculate and output the river surface water temperature prediction data layer by layer; the specific process is as follows: At each time step t, the trained LSTM model receives input data x t , the forget gate is calculated to obtain the river surface water temperature forecast information of the previous time step that the model decides needs to retain under the current meteorological and hydrological conditions; The calculation formula of the forget gate is as follows: f t =σ(W f ·[h t-1 ,x t ]+b f ); Among them, f t represents the water temperature information retention rate, σ represents the sigmoid function, W f represents the weight matrix of the forget gate, h t-1 represents the predicted river surface temperature at the previous time step, x t represents the meteorological and hydrological driving data and watershed attribute data input at the current time step, b f Represents the bias term of the forget gate; The influence ratio of meteorological and hydrological driving data and watershed attribute data input at the current time step on the update of river surface water temperature forecast state is calculated through the input gate. t And the candidate unit status The value of The input gate calculation formula is as follows: i t =σ(W i ·[h t-1 ,x t ]+b i ); Among them, i t represents the meteorological input impact rate, W i represents the weight matrix of the input gate, b i represents the bias vector of the input gate, represents the candidate unit state, tanh represents the hyperbolic tangent activation function, W C The weight matrix representing the candidate unit state, b C A bias vector representing the candidate cell state; Update the cumulative water temperature prediction information C of the current time step t , C t Including the meteorological and hydrological input of the current time step and the river surface water temperature forecast information of the previous time step, the calculation formula is as follows: Among them, C t and C t-1 Respectively represent the cumulative water temperature prediction information of the current time step and the previous time step, * represents element-by-element multiplication; By calculating the activation value o of the output gate t To control the information output ratio of the accumulated water temperature prediction information of the current time step, calculate the river surface water temperature prediction information h of the current time step t As output information, the calculation formula is as follows: the t =σ(W o ·[h t-1 ,x t ]+b o ); h t =o t *tanh(C t ); Among them, t represents the activation value of the output gate, h t represents the river surface water temperature forecast information at the current time step, W o represents the weight matrix of the output gate, b o represents the bias vector of the output gate; The predicted river surface temperature at the last time step h t As the feature representation of the entire time series, it is passed to the output layer to output the river surface water temperature prediction data. The expression is as follows: Among them, Wout represents the weight matrix of the output layer, bout represents the bias term, Represents the river surface water temperature prediction data output by the model.

4. The method according to claim 3, characterized in that In the step 14), the predicted data of the river surface water temperature and the observed data of the river surface water temperature are denormalized to evaluate the model performance and obtain the daily continuous river surface water temperature data, which specifically includes the following steps: The calculation formula for denormalization is as follows: Y=Y norm ×σ+μ; Among them, Y represents the water temperature data after denormalization, Y norm represents the normalized water temperature data, σ represents the standard deviation of the water temperature data, μ represents the average value of the water temperature data, and the water temperature data includes the predicted data of river surface water temperature and the observed data of river surface water temperature; The model was evaluated using the denormalized river surface temperature prediction data and river surface temperature observation data as follows: In the formula, y mod represents the predicted river surface temperature data, y obs represents the river surface water temperature observation data, n represents the total number of paired river surface water temperature prediction data and observation data during the model testing period, represents the average value of the predicted river surface water temperature data, represents the average value of the river surface water temperature observation data, Pbias represents the percentage deviation, NSE represents the Nash efficiency coefficient, RMSE represents the root mean square error, and Pcorr represents the Pearson correlation coefficient; after evaluating the reliability of the model output results based on the four evaluation parameters of Pbias, NSE, RMSE and Pcorr, the denormalized river surface water temperature prediction data is the daily continuous river surface water temperature data obtained.

5. The method according to claim 1, characterized in that In step 4), the basic characteristic indicators of the river heat wave events and the atmospheric heat wave events in step 2) and the annual variation trend of the time indicators and coupling rate of the river-atmosphere coupled heat wave events in step 3) are evaluated based on linear regression and Theil Sens slope method. The annual variation trend is expressed in absolute value and percentage value respectively, and the percentage trend change formula is: T [%, the -1 ]=100*(C End / C start -1) / N years ; Among them, C End and C Start Represents the average of the end date and the start date, Nyears represents the number of years of the time series span, T[%,yr -1 ] indicates the annual percentage change; In the step 4), the Pearson correlation coefficient is used to evaluate the correlation between the basic characteristic indicators of the river heat wave event and the atmospheric heat wave event in step 2) and the temporal characteristic indicators and coupling rate of the river-atmosphere coupled heat wave event in step 3); The Pearson correlation coefficient indicates the strength and direction of the linear relationship between two variables. The Pearson correlation coefficient calculation formula is as follows: Among them, r is the Pearson correlation coefficient, x i and i are the observed values ​​of two variables, and is the mean of the two variables, and the Pearson correlation coefficient ranges from -1 to 1, where -1 indicates a perfect negative correlation, 1 indicates a perfect positive correlation, and 0 indicates no correlation.

6. A river-atmosphere coupled heat wave assessment system, characterized in that: include: A data acquisition module, a first calculation module, an event coupling module, a second calculation module, a first evaluation module and a second evaluation module; A data acquisition module is used to obtain river water temperature data of multiple river basins in a large range within a set time; The first calculation module is used to calculate the heat wave temperature threshold of the study area, define the river heat wave events and atmospheric heat wave events in the study area, and calculate the basic characteristic indicators of the river heat wave events and atmospheric heat wave events; An event coupling module is used to couple river heat wave events with atmospheric heat wave events to define river-atmosphere coupled heat wave events; The second calculation module is used to calculate the relevant time characteristic indicators and coupling rate of river-atmosphere coupled heat wave events; The first evaluation module is used to evaluate the basic characteristic indicators of river heat wave events and atmospheric heat wave events, the temporal characteristic indicators of river-atmosphere coupled heat wave events, and the annual variation trend of the coupling rate; The second evaluation module is used to evaluate the correlation between the basic characteristic indicators of river heat wave events and atmospheric heat wave events, the temporal characteristic indicators of river-atmosphere coupled heat wave events, and the coupling rate; The data acquisition module specifically includes: 11) Obtain meteorological and hydrological driving data, river surface water temperature observation data and basin attribute data for multiple basins in a large range within a set time; 12) Preprocessing the data obtained in step 11), including deleting duplicate time series data, deleting erroneous records, deleting outliers and standardizing the data, to obtain preprocessed meteorological and hydrological driving data, river surface water temperature observation data and watershed attribute data; then converting the date of the data into time characteristic data; dividing the processed meteorological and hydrological driving data, river surface water temperature observation data, watershed attribute data and time characteristic data into a training set and a test set according to a certain ratio; 13) Setting hyperparameters for the LSTM model, inputting the training set data processed in step 12) into the LSTM model for training, and updating the model parameters, with the goal of minimizing the loss function L; 14) inputting the test set data processed in step 12) into the LSTM model trained in step 13), and outputting the predicted river surface water temperature data layer by layer through the trained LSTM model; evaluating the model performance after denormalizing the predicted river surface water temperature data and the observed river surface water temperature data, and obtaining the continuous daily river surface water temperature data; The first calculation module specifically includes: 21) Calculate the heat wave temperature threshold of the study area based on the river surface water temperature data obtained in step 1) and the daily maximum temperature data used; The daily temperature ensemble T of the five days before and after the d-day of all years in the study area during the reference period is d In ascending order, T d The expression is: Among them, T y,i represents the daily temperature on the i-th day of the y-th year; ∪ represents the union of sets, represents the daily temperature set of 5 days before and after the d-th day in the year 1 to m; The heat wave temperature threshold on day d is the daily temperature set T d The daily temperature corresponding to the 90th percentile after ascending order, the river heat wave temperature threshold and the atmospheric heat wave temperature threshold were calculated using the river surface water temperature and the daily maximum temperature as the daily temperature; 22) Define river heat wave events and atmospheric heat wave events in the study area, and set basic parameters for heat wave identification, including: the maximum interval length allowed for atmospheric heat wave events and river heat wave events is 2 days, the shortest atmospheric heat wave duration is 3 days, and the shortest river heat wave duration is 5 days; 23) Calculate the basic characteristic indicators of river heat wave events and atmospheric heat wave events, including heat wave frequency, heat wave duration and average heat wave intensity; The event coupling module specifically includes: Obtain the start date and end date of each river heat wave event and atmospheric heat wave event, create a search window for each river heat wave event based on the start and end dates of the river heat wave event, search for overlapping atmospheric heat wave events and river heat wave events on the search window and couple them, and define the matched coupled events as river-atmosphere coupled heat wave events; The second calculation module specifically includes: Based on the calculated basic characteristic indicators of river heat wave events and atmospheric heat wave events, the relevant time characteristic indicators and coupling rate of river-atmosphere coupled heat wave events are calculated; Use the detect_event function to determine the start date tRhw,start and end date tRhw,end for each river heat wave event; A search window is created for each river heat wave event, and overlapping atmospheric heat wave events and river heat wave events are searched and coupled on the search window; Define the start time tswb and end time tswe of the search window. The start time of the search window is defined as 120 hours before the start date of the river heat wave event, specifically: tswb = tRhw, start - 120 hours; The end time of the search window is defined as the end date of the river heat wave event, specifically: tswe=tRhw,end; Compare whether the start date tAhw,start and the end date tAhw,end of the atmospheric heat wave event are within the search window [tswb,tswe] to determine whether the atmospheric heat wave event is coupled with the river heat wave event. The coupling condition is: tswb≤tAhw,start≤tswe or tswb≤tAhw,end≤tswe; Or, tAhw,start≤tswe and tAhw,end≥tswb; All matching coupled events are recorded and stored in the event list as river-atmosphere coupled heat wave events; if no overlapping atmospheric heat wave events are found, the river heat wave events are defined as independent events; The relevant temporal characteristic indicators of the river-atmosphere coupled heat wave event include: The coupled river heat wave duration Tr is the sum of the river heat wave durations in all river-atmosphere coupled heat wave events in each basin within a set time, and is expressed as: Where N is the total number of river-atmosphere coupled heat wave events occurring within a set time, tro,i represents the start time of the river heat wave in the i-th coupled event, and tre,i represents the end time of the river heat wave in the i-th coupled event; The coupled atmospheric heat wave duration Ta is the sum of the durations of atmospheric heat waves in all river-atmosphere coupled heat wave events within a set period for each basin; When a river heat wave event in the search window is coupled with multiple atmospheric heat wave events, the coupled atmospheric heat wave time is the sum of all coupled atmospheric heat wave events in the window, expressed as: Where M represents the number of atmospheric heat wave events coupled in the ith river-atmosphere coupled heat wave event, tao,ij represents the start time of the jth atmospheric heat wave event coupled in the ith river-atmosphere coupled heat wave event, and tae,ij represents the end time of the jth atmospheric heat wave event coupled in the ith river-atmosphere coupled heat wave event. represents the sum of the durations of all atmospheric heat wave events in the i-th river-atmosphere coupled heat wave event; Coupled heat wave response lag time T LR The sum of the time distances from the start time of the atmospheric heat wave to the start time of the river heat wave in each river-atmosphere coupled heat wave event within a set time for each basin; When a river heat wave event is coupled with multiple atmospheric heat wave events, the start time of the first coupled atmospheric heat wave event is selected for calculation, and the expression is: Wherein, tao,i1 represents the starting time of the first atmospheric heat wave event coupled in the i-th river-atmosphere coupled heat wave event; Coupled heat wave time difference T HD The sum of the time difference between the duration of coupled river heat waves and the duration of coupled atmospheric heat waves in each river-atmosphere coupled heat wave event in each basin within a set time period is expressed as:

Images