A method and device for forecasting daily composite salt-drought index in estuary areas

By collecting daily runoff and salinity data in the estuary area, calculating standardized indices, and decomposing the composite salinity and drought index using Hilbert transform and random forest models, the problem of low accuracy in predicting composite salinity and drought disasters was solved, achieving high precision and stability in daily forecasts.

CN118760832BActive Publication Date: 2024-11-26SUN YAT SEN UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411218413.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-02
Publication Date
2024-11-26
Estimated Expiration
2044-09-02

AI Technical Summary

Technical Problem

In existing technologies, the accuracy of predicting complex saline-drought disasters in estuary areas is low, and the time resolution determined by month is coarse, making it difficult to accurately identify the start and end times of events.

Method used

By collecting daily runoff and salinity data in the estuary area, calculating the marginal distribution functions of runoff and salinity, generating standardized indices, and decomposing the composite salinity and drought index using Hilbert transform and random forest model, a daily composite salinity and drought index forecasting method and device are constructed to improve temporal resolution and prediction accuracy.

Benefits of technology

It has achieved accurate daily composite salinity and drought index forecasting in the estuary area, avoiding overfitting and improving the accuracy and stability of the forecast results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118760832B_ABST
    Figure CN118760832B_ABST
Patent Text Reader

Abstract

The invention discloses a method and device for predicting daily composite salty drought index in estuary areas, which are used to solve the problem of low prediction accuracy of composite salty drought disasters in estuaries. The method includes: calculating the runoff marginal distribution function of daily runoff data and the salinity marginal distribution function of salinity data; calculating the standardized runoff index according to the runoff marginal distribution function, and calculating the standardized salinity index according to the salinity marginal distribution function; calculating the daily-scale composite salty drought index according to the runoff marginal distribution function and the salinity marginal distribution function to obtain a daily-scale composite salty drought index sequence; using the standardized runoff index and the standardized salinity index to verify the applicability of the daily-scale composite salty drought index; when the verification is passed, decomposing the daily-scale composite salty drought index sequence into multiple sub-signal components; constructing a random forest model of each sub-signal component to obtain sub-composite salty drought index forecast data; adding and reconstructing each sub-composite salty drought index forecast data to generate a composite salty drought index forecast value.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of estuary salt drought intensity prediction, and in particular to a method and device for predicting daily composite salt drought index in estuary areas. Background Art

[0002] Compound events are a combination of factors or hazards that usually lead to more serious ecological and socioeconomic impacts than a single event. Such events have become a hot topic in the field of extreme event science. With the intensification of climate change, extreme events, especially drought-related events, are occurring at an increasingly rapid rate, such as sudden droughts, which are expected to become more severe in the future. Compound droughts and saltwater upwelling events may exacerbate water shortages, crop yields, primary productivity declines, and tree deaths in estuaries, seriously affecting ecosystems, the economy, and society.

[0003] The temporal resolution of compound salty drought events determined on a monthly basis is relatively coarse, which may hinder the accurate identification of the start and end time of the event. In addition, compound salty drought events occurring simultaneously in a short period of time will have a greater impact. Summary of the invention

[0004] The invention provides a method and device for predicting daily composite salt-drought index in estuary areas, which are used to solve the technical problem of low prediction accuracy of composite salt-drought disasters in estuaries.

[0005] The present invention provides a method for forecasting a daily composite salt-drought index in an estuary area, comprising:

[0006] Collect daily runoff and salinity data for pre-defined estuary areas;

[0007] Calculating the runoff marginal distribution function of the daily runoff data and the salinity marginal distribution function of the salinity data;

[0008] Calculating a normalized runoff index of the daily runoff data according to the runoff marginal distribution function, and calculating a normalized salinity index of the salinity data according to the salinity marginal distribution function;

[0009] Calculating a daily-scale composite salt-drought index according to the runoff marginal distribution function and the salinity marginal distribution function to obtain a daily-scale composite salt-drought index sequence;

[0010] The applicability of the daily-scale composite salt-drought index is verified by using the standardized runoff index and the standardized salinity index;

[0011] When the verification is passed, the daily scale composite salt-drought index sequence is decomposed into a plurality of sub-signal components;

[0012] Constructing random forest models for each sub-signal component respectively, and obtaining sub-composite salt-drought index forecast data for each sub-signal component according to the random forest model;

[0013] The forecast data of each sub-composite salt-drought index are added and reconstructed to generate the forecast value of the composite salt-drought index.

[0014] Optionally, the step of collecting daily runoff and salinity data of a preset estuary area comprises:

[0015] Collect initial daily runoff data and initial salinity data for the pre-determined estuary area;

[0016] Interpolation is performed on the initial daily runoff data and the initial salinity data to obtain daily runoff data and salinity data.

[0017] Optionally, the step of calculating the normalized runoff index of the daily runoff data according to the runoff marginal distribution function, and the step of calculating the normalized salinity index of the salinity data according to the salinity marginal distribution function, comprises:

[0018] Calculating the runoff marginal probability of daily runoff data according to the runoff marginal distribution function;

[0019] The runoff marginal probability is normalized to obtain a normalized runoff index;

[0020] Calculating the salinity marginal probability of the salinity data according to the salinity marginal distribution function;

[0021] The salinity marginal probability is normalized to obtain a normalized salinity index.

[0022] Optionally, the step of calculating the daily-scale composite salt-drought index according to the runoff marginal distribution function and the salinity marginal distribution function to obtain a daily-scale composite salt-drought index sequence comprises:

[0023] Calculating the ratio of the runoff marginal distribution function to the salinity marginal distribution function, and obtaining a daily-scale composite salt-drought index based on the ratio;

[0024] The daily-scale composite salt-drought indexes of consecutive days are arranged in chronological order to obtain a daily-scale composite salt-drought index sequence.

[0025] Optionally, when the verification is passed, the step of decomposing the daily-scale composite salt-drought index sequence into a plurality of sub-signal components comprises:

[0026] When the verification is passed, the instantaneous amplitude and instantaneous frequency of the input signal of the daily-scale composite salt-drought index are calculated based on the Hilbert transform;

[0027] determining a first sequence of local maxima and a first sequence of local minima of the instantaneous amplitude;

[0028] Performing interpolation processing on the local maximum value sequence and the local minimum value sequence to obtain a first curve and a second curve;

[0029] Calculating an instantaneous average value and an instantaneous envelope according to the first curve and the second curve;

[0030] Calculating a second local maximum sequence and a second local minimum sequence of the instantaneous frequency;

[0031] Calculating a local cutoff frequency using the first local maximum sequence, the first local minimum sequence, the second local maximum sequence, and the second local minimum sequence;

[0032] generating a local average curve according to the local cutoff frequency;

[0033] Calculate the cutoff condition;

[0034] When the cutoff condition does not meet the preset threshold, the input signal is updated using the local average curve, and the step of calculating the instantaneous amplitude and instantaneous frequency of the input signal of the daily-scale composite salt-drought index based on the Hilbert transform is returned;

[0035] When the cutoff condition meets a preset threshold, the decomposition is completed and the input signal is used as a sub-signal component.

[0036] Optionally, the steps of respectively constructing a random forest model for each sub-signal component and obtaining sub-composite salt-drought index forecast data for each sub-signal component according to the random forest model include:

[0037] Performing normalization processing on the sub-signal components to obtain normalized data;

[0038] Inputting the normalized data into a random forest model to obtain a predicted value of each sub-signal component;

[0039] The predicted values ​​are subjected to inverse normalization processing to obtain sub-composite salt-drought index forecast data.

[0040] Optionally, it also includes:

[0041] The root mean square error, Nash efficiency coefficient, correlation coefficient, mean absolute error and Kling-Gupta efficiency coefficient were used to evaluate the accuracy of the composite salt-drought index forecast data.

[0042] The present invention also provides a daily composite salt-drought index forecasting device for estuary areas, comprising:

[0043] The collection module is used to collect daily runoff data and salinity data of the preset estuary area;

[0044] A marginal distribution function calculation module, used for calculating the runoff marginal distribution function of the daily runoff data and the salinity marginal distribution function of the salinity data;

[0045] A standardized runoff index and a standardized salinity index calculation module, used to calculate the standardized runoff index of the daily runoff data according to the runoff marginal distribution function, and to calculate the standardized salinity index of the salinity data according to the salinity marginal distribution function;

[0046] A daily scale composite salty drought index sequence generation module is used to calculate the daily scale composite salty drought index according to the runoff marginal distribution function and the salinity marginal distribution function to obtain a daily scale composite salty drought index sequence;

[0047] A suitability judgment module, used for verifying the suitability of the daily-scale composite salt-drought index by using the standardized runoff index and the standardized salinity index;

[0048] A decomposition module, used for decomposing the daily scale composite salt-drought index sequence into a plurality of sub-signal components when the verification is passed;

[0049] A sub-composite salty drought index forecast data calculation module is used to construct a random forest model of each sub-signal component, and obtain the sub-composite salty drought index forecast data of each sub-signal component according to the random forest model;

[0050] The composite salty drought index forecast value generation module is used to add and reconstruct the forecast data of each sub-composite salty drought index to generate a composite salty drought index forecast value.

[0051] The present invention also provides an electronic device, the device comprising a processor and a memory:

[0052] The memory is used to store program codes and transmit the program codes to the processor;

[0053] The processor is used to execute the daily composite salt-drought index forecasting method for estuary areas as described in any one of the above items according to the instructions in the program code.

[0054] The present invention also provides a computer-readable storage medium, which is used to store program codes, and the program codes are used to execute the daily composite salt-drought index forecasting method for estuary areas as described in any one of the above items.

[0055] It can be seen from the above technical scheme that the present invention has the following advantages: the present invention calculates the daily-scale composite salt-drought index sequence by collecting daily runoff data and salinity data of a preset estuary area to improve the accuracy of time resolution, and decomposes the daily-scale composite salt-drought index sequence to predict each sub-signal component separately, thereby making the prediction result more accurate and avoiding the overfitting problem generated during the operation process. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0057] Figure 1 A flowchart of the steps of a method for forecasting a daily composite salt-drought index in an estuary area provided by an embodiment of the present invention;

[0058] Figure 2 It is a daily runoff and salinity variation diagram of a daily composite salt drought index forecasting method for an estuary area of ​​the present invention;

[0059] Figure 3 It is a daily scale composite salty drought index sequence line chart of a daily composite salty drought index forecasting method for an estuary area of ​​the present invention;

[0060] Figure 4 It is a TVE-EMD decomposition result diagram of an estuary composite salty drought index sequence of a daily composite salty drought index forecasting method for an estuary area of ​​the present invention;

[0061] Figure 5 It is a test set component forecast result diagram based on TVF-EMD-RF decomposition of a daily composite salt drought index forecast method for estuary areas of the present invention;

[0062] Figure 6 It is a sub-component prediction evaluation index diagram based on TVF-EMD-RF decomposition of a daily composite salt-drought index forecasting method for estuary areas of the present invention;

[0063] Figure 7 It is a test set evaluation index diagram of the TVF-EMD-RF hybrid model of the A estuary of the present invention;

[0064] Figure 8 This is a prediction effect index diagram of the RF model of estuary A of the present invention;

[0065] Fig. 9A structural block diagram of a device for forecasting daily composite salt and drought index in an estuary area provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0066] The embodiment of the present invention provides a method and device for predicting a daily composite salt-drought index in an estuary area, which are used to solve the technical problem of low prediction accuracy of composite salt-drought disasters in estuaries.

[0067] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0068] See also Figure 1 , Figure 1 A flowchart of the steps of a method for predicting a daily composite salt-drought index in an estuary area provided by an embodiment of the present invention.

[0069] The present invention provides a method for forecasting a daily composite salt-drought index in an estuary area, which may specifically include the following steps:

[0070] Step 101, collecting daily runoff data and salinity data of a preset estuary area;

[0071] In the embodiment of the present invention, an estuary area to be studied may be selected, and daily runoff data and salinity data of the estuary area may be collected to prepare for the distributed simulation.

[0072] In one example, step 101 may include the following sub-steps:

[0073] S11, collect initial daily runoff data and initial salinity data of the preset estuary area;

[0074] S12, interpolating the initial daily runoff data and the initial salinity data to obtain daily runoff data and salinity data.

[0075] In a specific implementation, the collected initial daily runoff data and initial salinity data of the estuary area can be preprocessed, such as performing corresponding interpolation processing on missing values ​​to improve the initial daily runoff data and initial salinity data to obtain daily runoff data and salinity data.

[0076] Step 102, calculating the runoff marginal distribution function of the daily runoff data and the salinity marginal distribution function of the salinity data;

[0077] In the embodiment of the present invention, the runoff marginal distribution function may be calculated using daily runoff data, and the salinity marginal distribution function may be calculated using salinity data.

[0078] In the specific implementation, the pre-processed daily runoff data can be fitted by applying Lognorm, Logistic, Normal, Gamma, and Weibull distributions to obtain the fitted distribution function. Then, the KS test is performed on the fitted distribution function, the statistical value of each distribution function is calculated, and the distribution function with the smallest statistical value is selected as the optimal distribution function to calculate the marginal distribution function of runoff and the marginal distribution function of salinity.

[0079] Step 103, calculating a standardized runoff index of daily runoff data according to the runoff marginal distribution function, and calculating a standardized salinity index of salinity data according to the salinity marginal distribution function;

[0080] In the embodiment of the present invention, the normalized runoff index of the daily runoff data may be calculated according to the runoff distribution function, and the normalized salinity index of the salinity data may be calculated according to the salinity distribution function.

[0081] In one example, the steps of calculating the normalized runoff index of daily runoff data according to the runoff marginal distribution function and calculating the normalized salinity index of salinity data according to the salinity marginal distribution function may include the following sub-steps:

[0082] S31, calculating the runoff marginal probability of the daily runoff data according to the runoff marginal distribution function;

[0083] S32, normalizing the marginal probability of runoff to obtain a standardized runoff index;

[0084] S33, calculating the salinity marginal probability of the salinity data according to the salinity marginal distribution function;

[0085] S34, normalizing the salinity marginal probability to obtain a standardized salinity index.

[0086] In a specific implementation, the runoff marginal probability of daily runoff data can be calculated according to the runoff marginal distribution function, and then normalized to obtain the standardized runoff index SRI. And the salinity marginal probability can be calculated according to the salinity marginal distribution function, and then normalized to obtain the standardized salinity index SSI. Then the SSI value is multiplied by -1 to make it consistent with the SRI standard, that is, negative values ​​represent drought and high salinity conditions, and positive values ​​represent wet and freshwater conditions.

[0087] Step 104, calculating the daily scale composite salt-drought index according to the runoff marginal distribution function and the salinity marginal distribution function, and obtaining a daily scale composite salt-drought index sequence;

[0088] In the embodiment of the present invention, the severity of the compound salty drought event can be expressed by the ratio between the runoff marginal distribution function and the salinity marginal distribution function, and the daily scale compound salty drought index SNSI can be calculated, and a daily scale compound salty drought index sequence within a certain time interval can be generated.

[0089] In one example, the step of calculating the daily-scale composite salt-drought index according to the runoff marginal distribution function and the salinity marginal distribution function to obtain the daily-scale composite salt-drought index sequence may include the following sub-steps:

[0090] S41, calculating the ratio of the runoff marginal distribution function to the salinity marginal distribution function, and obtaining the daily scale composite salt-drought index according to the ratio;

[0091] S42, arranging the daily-scale composite salt-drought indexes of multiple consecutive days in chronological order to obtain a daily-scale composite salt-drought index sequence.

[0092] In the specific implementation, to define the relative status of drought and salinity conditions, the marginal probability ratio of runoff and salinity is defined as variable Z:

[0093]

[0094] in, is the marginal distribution function of runoff; is the marginal distribution function of salinity.

[0095] The SDSI based on variable Z distribution can be expressed as:

[0096]

[0097] in, is the standard normal distribution, is the marginal cumulative distribution, which is fitted based on the empirical Gringorton location partition formula.

[0098] Step 105, using a standardized runoff index and a standardized salinity index to verify the applicability of a daily-scale composite salt-drought index;

[0099] In the embodiment of the present invention, the applicability and reliability of the daily-scale SDSI in different estuary areas can be verified by comparative analysis of the changing trends of the standardized runoff index SRI and the standardized salinity index SSI, analysis of typical historical droughts and severe saltwater tide upstream events in the estuary area, and correlation analysis.

[0100] Step 106, when the verification is passed, decomposing the daily scale composite salt-drought index series into multiple sub-signal components;

[0101] In the embodiment of the present invention, a signal decomposition method may be used to decompose the original SDSI sequence into equal number of sub-signal components to improve the nonlinearity and non-stationarity of the data.

[0102] Taking the TVF-EMD signal decomposition method as an example, step 106 may include the following sub-steps:

[0103] S601, when the verification is passed, calculating the instantaneous amplitude and instantaneous frequency of the input signal of the daily-scale composite salt-drought index based on Hilbert transform;

[0104] S602, determining a first local maximum value sequence and a first local minimum value sequence of instantaneous amplitudes;

[0105] S603, performing interpolation processing on the local maximum value sequence and the local minimum value sequence to obtain a first curve and a second curve;

[0106] S604, calculating an instantaneous average value and an instantaneous envelope according to the first curve and the second curve;

[0107] S605, calculating a second local maximum value sequence and a second local minimum value sequence of the instantaneous frequency;

[0108] S606, calculating a local cutoff frequency using the first local maximum sequence, the first local minimum sequence, the second local maximum sequence, and the second local minimum sequence;

[0109] S607, generating a local average curve according to the local cutoff frequency;

[0110] S608, calculating the cut-off condition;

[0111] S609, when the cutoff condition does not meet the preset threshold, the input signal is updated using the local average curve, and the step of calculating the instantaneous amplitude and instantaneous frequency of the input signal of the daily-scale composite salt-drought index based on the Hilbert transform is returned;

[0112] S610: When the cutoff condition meets the preset threshold, the decomposition is completed and the input signal is used as a sub-signal component.

[0113] In the specific implementation, the TVF-EMD signal decomposition method first needs to calculate the instantaneous amplitude of the daily-scale SDSI input signal based on the Hilbert transform. and instantaneous frequency :

[0114]

[0115]

[0116] in, is the input signal The Hilbert transform of .

[0117] Next, determine the first local maximum sequence of instantaneous amplitudes and the first local minimum sequence By interpolation and Get the first curve and the second curve , then calculate the instantaneous average and the instantaneous envelope , the calculation formula is as follows:

[0118]

[0119]

[0120] Then, through interpolation and Get and , then calculate and :

[0121]

[0122]

[0123] in, and are respectively the second local maximum sequence and the second local minimum sequence of the instantaneous frequency.

[0124] Then calculate the local cutoff frequency as follows:

[0125]

[0126] Then, readjust Solve intermittent problems.

[0127] Defining Signals , with signal The extreme points of are nodes, and the approximation signal sequence is obtained through B-spline interpolation. The approximation result is , represents the local average curve.

[0128] Calculation cut-off condition ,if ,determination If it is a local narrowband signal, the decomposition ends; otherwise, , repeat the above steps. The calculation formula to reach the cut-off condition is The details are as follows:

[0129]

[0130]

[0131]

[0132] in, is the instantaneous bandwidth of the two Loughlin component signals; is the weighted average instantaneous frequency of a single component signal. Through the decomposition of the above steps, the value of the original data sequence are all mode components The sum of .

[0133] Step 107, constructing a random forest model for each sub-signal component respectively, and obtaining sub-composite salt-drought index forecast data for each sub-signal component according to the random forest model;

[0134] Step 108, adding and reconstructing the sub-composite salt-drought index forecast data to generate a composite salt-drought index forecast value.

[0135] In the embodiment of the present invention, a hybrid model can be formed by combining a decomposition and reconstruction method with a random forest (RF) model to obtain the final forecast value of the composite salt-drought index.

[0136] The supervised learning algorithm in RF machine learning combines Bootstrap sampling and decision tree algorithms. It has the advantages of small computational complexity and easy implementation, so it is widely used in variable prediction. The algorithm consists of a large number of decision trees for classification and regression analysis, and each decision tree is generated by a random vector. Its main ideas are reflected in two aspects: one is random sampling, and the other is the selection of random features. RF divides the original sequence of explanatory variables into multiple decision trees, and predicts the prediction objects separately, so that the prediction results are more accurate and can avoid the overfitting problem generated during the operation process.

[0137] In one example, step 107 may include the following sub-steps:

[0138] S71, performing normalization processing on the sub-signal components to obtain normalized data;

[0139] S72, inputting the normalized data into the random forest model to obtain the predicted value of each sub-signal component;

[0140] S73, performing a denormalization process on the predicted value to obtain sub-composite salt-drought index forecast data.

[0141] In the specific implementation, the sub-component sequences obtained by SDSI are normalized respectively and used as the input data of the RF model. The constructed sub-sequence RF models are trained to obtain the predicted values ​​of each sub-sequence component, and the obtained predicted values ​​are denormalized. Finally, the result data of each sub-component are added together to obtain the final prediction result.

[0142] Furthermore, in the embodiment of the present invention, five evaluation indicators can be used to comprehensively evaluate the forecast accuracy of the composite salt-drought index, namely, root mean square error (RMSE), Nash efficiency coefficient (NSE), correlation coefficient (R), mean absolute error (MAE) and Kling-Gupta efficiency coefficient (KGE). Their expressions are:

[0143]

[0144]

[0145]

[0146]

[0147]

[0148] in, represents the number of samples, represents the observed data, is the average value of the observed data; Represents the forecast data, is the average value of the predicted data; is the relative rate of change, is the bias ratio, and are the standard deviations of the two series, and are the means of the two series respectively.

[0149] The present invention calculates the daily-scale composite salt-drought index sequence by collecting daily runoff data and salinity data of a preset estuary area to improve the accuracy of time resolution, and decomposes the daily-scale composite salt-drought index sequence to predict each sub-signal component separately, thereby making the prediction result more accurate and avoiding the overfitting problem generated during the operation process.

[0150] For ease of understanding, the embodiments of the present invention are described below by using specific examples:

[0151] 1) By collecting runoff and salinity data from the A, B, and C estuaries, including daily runoff and salinity data from the A estuary during the dry season from 1999 to 2016, daily measured runoff and salinity data from the B estuary from 1994 to 2014, and daily measured runoff and salinity data from the C estuary from 1986 to 2014, the coincidence of low runoff (30th percentile) and high salinity events (70th percentile) were evaluated using coincidental events and the likelihood multiplier method ( Figure 2 ).

[0152] 2) The optimal probability distribution function of daily runoff and salinity data in different estuary areas was determined. Five different types of distribution functions were selected for fitting, and the KS test index was used to evaluate and select the optimal function. The marginal distribution functions of the runoff and salinity series were calculated respectively. The marginal fitting evaluation of the function is shown in Table 1. Based on the principle that the smaller the KS evaluation index value, the better the goodness of fit, the table shows that the Lognorm function achieves the minimum value in the five evaluation indicators of runoff and salinity data, thus ensuring the robustness of the results.

[0153] Table 1 Goodness of fit of marginal distribution of runoff and salinity (black bold indicates the optimal value)

[0154]

[0155] 3) The univariate daily scale standardized runoff index (SRI) and standardized coastal salinity index (SSI) were calculated using the marginal distribution function, and the composite salt-drought index SDSI was calculated based on the ratio of the marginal distribution function, as follows: Figure 3 As shown in (a), the correlation coefficients between SDSI and the univariate index are all greater than 0.70, which further demonstrates that the composite index is very sensitive to capturing composite events. In addition, the daily SDSI is effective in identifying composite salt-drought disasters in the A estuary area, especially in the four most severe hydrological droughts and saltwater upwelling events in 2004-2005, 2009-2010, 2011-2012 and 2021-2022. During the dry season of 2004-2005, the A estuary area experienced severe hydrological drought and saltwater upwelling, which led to ecological damage and directly affected the water use for residents' daily life, crop irrigation and industrial water use, becoming a typical composite salt-drought event. Figure 3 (b) and Figure 3 (c) Further demonstrates the effectiveness of SDSI in identifying compound salty drought events at estuary B in 2012 and estuary C in 2002. These two events were equally severe, verifying the reliability of SDSI in identifying compound salty drought events, thus proving that the index has wide applicability and practical value in estuary areas.

[0156] 4) The TVF-EMD method is used to decompose the composite salt-drought index of estuary A to ensure the accuracy and stability of the results. The specific settings are as follows: the bandwidth threshold is set to 0.3, the B-spline order is 26, and the number of decomposition layers is uniformly set to 10 layers. This number of decomposition layers is based on the optimal number of decomposition layers calculated by ESMD. The decomposition results are shown in Figure 2. Figure 4 As shown in the figure. With the increase of decomposition layers, the frequency of the subcomponents gradually decreases, the wavelength gradually increases, the amplitude gradually decreases, and the period becomes longer. The decomposed subsequences show an obvious linear trend, which can effectively reflect the changes in compound salt-drought disasters. The decomposed subsequences are more stable than the original sequences. These subsequences can provide good data input for machine learning models, thereby improving the accuracy of predictions.

[0157] 5) Build the RF model for each sub-signal component separately, determine the composite salty drought index of the first three days as the input factor through partial correlation, take the data of the fourth day as the output, forecast each sub-signal component, and finally reconstruct these sub-signal components to obtain the final composite salty drought index forecast value. Combine TVF-EMD and RF models to build a hybrid model. It should be pointed out that the decomposition method combined with the parameter setting of the RF model is the same. Prediction is performed using the RF model and the 10 sub-components of TVF-EMD. Figure 5 The comparison between the measured and predicted values ​​of the test set is shown, and the results show that the TVF-EMD-RF hybrid model has a good prediction effect on each sub-component. Figure 6 As shown in the figure, overall, the prediction performance of each subcomponent based on the TVF-EMD-RF model during the test period was good, with R values ​​above 0.9. Although the R values ​​of the first and second subcomponents are low, this is because the first subcomponent has greater randomness and noise, which leads to its rapid and irregular changes, making its prediction relatively difficult. As the number of decomposition layers increases, the frequency of the subcomponent decreases, becomes smoother, and exhibits periodicity, so the R value gradually increases. This is because the low-frequency signal is mainly affected by long-term factors, making its fluctuations relatively stable and therefore easier to predict. Figure 7 This shows that the TVF-EMD-RF hybrid model performs well in prediction. Figure 8 The prediction effect based on RF model alone is significantly improved. Therefore, the TVF-EMD-RF hybrid model has great potential in the effective prediction of compound salt-drought events in estuaries.

[0158] See also Fig. 9 , Fig. 9 A structural block diagram of a device for forecasting daily composite salt and drought index in an estuary area provided by an embodiment of the present invention.

[0159] The embodiment of the present invention provides a device for forecasting daily composite salt and drought index in an estuary area, comprising:

[0160] The collection module 901 is used to collect daily runoff data and salinity data of a preset estuary area;

[0161] A marginal distribution function calculation module 902 is used to calculate the runoff marginal distribution function of the daily runoff data and the salinity marginal distribution function of the salinity data;

[0162] A standardized runoff index and standardized salinity index calculation module 903, used to calculate the standardized runoff index of daily runoff data according to the runoff marginal distribution function, and to calculate the standardized salinity index of salinity data according to the salinity marginal distribution function;

[0163] A daily scale composite salt-drought index sequence generation module 904 is used to calculate the daily scale composite salt-drought index according to the runoff marginal distribution function and the salinity marginal distribution function to obtain a daily scale composite salt-drought index sequence;

[0164] The applicability judgment module 905 is used to verify the applicability of the daily scale composite salt-drought index by using the standardized runoff index and the standardized salinity index;

[0165] A decomposition module 906 is used to decompose the daily scale composite salt-drought index series into multiple sub-signal components when the verification is passed;

[0166] The sub-composite salty drought index forecast data calculation module 907 is used to construct a random forest model of each sub-signal component, and obtain the sub-composite salty drought index forecast data of each sub-signal component according to the random forest model;

[0167] The composite salty-drought index forecast value generating module 908 is used to add and reconstruct the sub-composite salty-drought index forecast data to generate a composite salty-drought index forecast value.

[0168] In the embodiment of the present invention, the collection module 901 includes:

[0169] The initial daily runoff data and initial salinity data collection submodule is used to collect the initial daily runoff data and initial salinity data of the preset estuary area;

[0170] The daily runoff data and salinity data collection submodule is used to interpolate the initial daily runoff data and initial salinity data to obtain daily runoff data and salinity data.

[0171] In the embodiment of the present invention, the normalized runoff index and normalized salinity index calculation module 903 includes:

[0172] The runoff marginal probability calculation submodule is used to calculate the runoff marginal probability of daily runoff data according to the runoff marginal distribution function;

[0173] The standardized runoff index calculation submodule is used to standardize the runoff marginal probability and obtain the standardized runoff index;

[0174] A salinity marginal probability calculation submodule is used to calculate the salinity marginal probability of salinity data according to the salinity marginal distribution function;

[0175] The standardized salinity index calculation submodule is used to standardize the salinity marginal probability to obtain the standardized salinity index.

[0176] In the embodiment of the present invention, the daily scale composite salt-drought index sequence generation module 904 includes:

[0177] The daily scale composite salt-drought index calculation submodule is used to calculate the ratio of the runoff marginal distribution function and the salinity marginal distribution function, and obtain the daily scale composite salt-drought index based on the ratio;

[0178] The daily scale composite salty drought index sequence generation submodule is used to arrange the daily scale composite salty drought indexes of multiple consecutive days in chronological order to obtain the daily scale composite salty drought index sequence.

[0179] In the embodiment of the present invention, the decomposition module 906 includes:

[0180] The instantaneous amplitude and instantaneous frequency calculation submodule is used to calculate the instantaneous amplitude and instantaneous frequency of the input signal of the daily scale composite salt-drought index based on the Hilbert transform when the verification is passed;

[0181] A first local maximum value sequence and a first local minimum value sequence determining submodule, used to determine a first local maximum value sequence and a first local minimum value sequence of instantaneous amplitudes;

[0182] A first curve and a second curve generating submodule, used for performing interpolation processing on a local maximum value sequence and a local minimum value sequence to obtain a first curve and a second curve;

[0183] An instantaneous average value and instantaneous envelope calculation submodule, used to calculate the instantaneous average value and the instantaneous envelope according to the first curve and the second curve;

[0184] A second local maximum value sequence and a second local minimum value sequence calculation submodule, used for calculating a second local maximum value sequence and a second local minimum value sequence of the instantaneous frequency;

[0185] A local cutoff frequency calculation submodule, used to calculate the local cutoff frequency using the first local maximum value sequence, the first local minimum value sequence, the second local maximum value sequence and the second local minimum value sequence;

[0186] A local average curve generating submodule, used for generating a local average curve according to a local cutoff frequency;

[0187] A cut-off condition calculation submodule, used for calculating the cut-off condition;

[0188] A return submodule is used for updating the input signal using the local average curve when the cutoff condition does not meet the preset threshold, and returning the step of calculating the instantaneous amplitude and instantaneous frequency of the input signal of the daily-scale composite salt-drought index based on the Hilbert transform;

[0189] The cutoff submodule is used to terminate the decomposition when the cutoff condition meets a preset threshold value, and use the input signal as a sub-signal component.

[0190] In the embodiment of the present invention, the sub-composite salt-drought index forecast data calculation module 907 includes:

[0191] A normalization submodule, used for normalizing the sub-signal components to obtain normalized data;

[0192] The prediction submodule is used to input the normalized data into the random forest model to obtain the predicted value of each sub-signal component;

[0193] The anti-normalization submodule is used to perform anti-normalization processing on the predicted value to obtain the sub-composite salt-drought index forecast data.

[0194] In an embodiment of the present invention, it also includes:

[0195] The accuracy evaluation submodule is used to evaluate the accuracy of the composite salt-drought index forecast data using root mean square error, Nash efficiency coefficient, correlation coefficient, mean absolute error and Kling-Gupta efficiency coefficient.

[0196] An embodiment of the present invention further provides an electronic device, the device comprising a processor and a memory:

[0197] The memory is used to store the program code and transmit the program code to the processor;

[0198] The processor is used to execute the daily composite salt-drought index forecasting method for estuary areas according to the instructions in the program code.

[0199] The embodiment of the present invention further provides a computer-readable storage medium, which is used to store program codes, and the program codes are used to execute the daily composite salt-drought index forecasting method for estuary areas in the embodiment of the present invention.

[0200] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0201] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0202] It will be appreciated by those skilled in the art that the embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the embodiments of the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the embodiments of the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0203] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0204] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0205] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0206] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention.

[0207] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0208] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for forecasting daily composite salt-drought index in estuary areas, characterized in that: include: Collect daily runoff and salinity data for pre-defined estuary areas; Calculating the runoff marginal distribution function of the daily runoff data and the salinity marginal distribution function of the salinity data; Calculating a normalized runoff index of the daily runoff data according to the runoff marginal distribution function, and calculating a normalized salinity index of the salinity data according to the salinity marginal distribution function; Calculating a daily-scale composite salt-drought index according to the runoff marginal distribution function and the salinity marginal distribution function to obtain a daily-scale composite salt-drought index sequence; The applicability of the daily-scale composite salt-drought index is verified by using the standardized runoff index and the standardized salinity index; When the verification is passed, the daily scale composite salt-drought index sequence is decomposed into a plurality of sub-signal components; Constructing random forest models for each sub-signal component respectively, and obtaining sub-composite salt-drought index forecast data for each sub-signal component according to the random forest model; The forecast data of each sub-composite salt-drought index are added and reconstructed to generate a forecast value of the composite salt-drought index; Wherein, when the verification is passed, the step of decomposing the daily scale composite salt-drought index sequence into a plurality of sub-signal components comprises: When the verification is passed, the instantaneous amplitude and instantaneous frequency of the input signal of the daily-scale composite salt-drought index are calculated based on the Hilbert transform; determining a first sequence of local maxima and a first sequence of local minima of the instantaneous amplitude; Performing interpolation processing on the first local maximum value sequence and the first local minimum value sequence to obtain a first curve and a second curve; Calculating an instantaneous average value and an instantaneous envelope according to the first curve and the second curve; Calculating a second local maximum sequence and a second local minimum sequence of the instantaneous frequency; Calculating a local cutoff frequency using the first local maximum sequence, the first local minimum sequence, the second local maximum sequence, and the second local minimum sequence; generating a local average curve according to the local cutoff frequency; Calculate the cutoff condition; When the cutoff condition does not meet the preset threshold, the input signal is updated using the local average curve, and the step of calculating the instantaneous amplitude and instantaneous frequency of the input signal of the daily-scale composite salt-drought index based on the Hilbert transform is returned; When the cutoff condition meets a preset threshold, the decomposition ends and the input signal is used as a sub-signal component; The steps of respectively constructing random forest models for each sub-signal component and obtaining sub-composite salt-drought index forecast data for each sub-signal component according to the random forest model include: Performing normalization processing on the sub-signal components to obtain normalized data; Inputting the normalized data into a random forest model to obtain a predicted value of each sub-signal component; The predicted values ​​are subjected to inverse normalization processing to obtain sub-composite salt-drought index forecast data.

2. The method according to claim 1, characterized in that The step of collecting daily runoff data and salinity data of a preset estuary area comprises: Collect initial daily runoff data and initial salinity data for the pre-determined estuary area; The initial daily runoff data and the initial salinity data are interpolated to obtain daily runoff data and salinity data.

3. The method according to claim 1, characterized in that: The step of calculating the normalized runoff index of the daily runoff data according to the runoff marginal distribution function, and the step of calculating the normalized salinity index of the salinity data according to the salinity marginal distribution function, comprises: Calculating the runoff marginal probability of daily runoff data according to the runoff marginal distribution function; The runoff marginal probability is normalized to obtain a normalized runoff index; Calculating the salinity marginal probability of the salinity data according to the salinity marginal distribution function; The salinity marginal probability is normalized to obtain a normalized salinity index.

4. The method according to claim 1, characterized in that The step of calculating the daily-scale composite salt-drought index according to the runoff marginal distribution function and the salinity marginal distribution function to obtain a daily-scale composite salt-drought index sequence comprises: Calculating the ratio of the runoff marginal distribution function to the salinity marginal distribution function, and obtaining a daily-scale composite salt-drought index based on the ratio; The daily-scale composite salt-drought indexes of consecutive days are arranged in chronological order to obtain a daily-scale composite salt-drought index sequence.

5. The method according to claim 1, characterized in that Also includes: The root mean square error, Nash efficiency coefficient, correlation coefficient, mean absolute error and Kling-Gupta efficiency coefficient were used to evaluate the accuracy of the forecast value of the composite salt-drought index.

6. A daily composite salt-drought index forecasting device for estuary areas, characterized in that: include: The collection module is used to collect daily runoff data and salinity data of the preset estuary area; A marginal distribution function calculation module, used for calculating the runoff marginal distribution function of the daily runoff data and the salinity marginal distribution function of the salinity data; A standardized runoff index and a standardized salinity index calculation module, used to calculate the standardized runoff index of the daily runoff data according to the runoff marginal distribution function, and to calculate the standardized salinity index of the salinity data according to the salinity marginal distribution function; A daily scale composite salty drought index sequence generation module is used to calculate the daily scale composite salty drought index according to the runoff marginal distribution function and the salinity marginal distribution function to obtain a daily scale composite salty drought index sequence; A suitability judgment module, used for verifying the suitability of the daily-scale composite salt-drought index by using the standardized runoff index and the standardized salinity index; A decomposition module, used for decomposing the daily scale composite salt-drought index sequence into a plurality of sub-signal components when the verification is passed; A sub-composite salty drought index forecast data calculation module is used to construct a random forest model of each sub-signal component, and obtain the sub-composite salty drought index forecast data of each sub-signal component according to the random forest model; A composite salty drought index forecast value generation module is used to add and reconstruct the forecast data of each sub-composite salty drought index to generate a composite salty drought index forecast value; The decomposition modules include: The instantaneous amplitude and instantaneous frequency calculation submodule is used to calculate the instantaneous amplitude and instantaneous frequency of the input signal of the daily scale composite salt-drought index based on the Hilbert transform when the verification is passed; A first local maximum value sequence and a first local minimum value sequence determining submodule, used to determine a first local maximum value sequence and a first local minimum value sequence of instantaneous amplitudes; A first curve and a second curve generating submodule, used for performing interpolation processing on the first local maximum value sequence and the first local minimum value sequence to obtain the first curve and the second curve; An instantaneous average value and instantaneous envelope calculation submodule, used to calculate the instantaneous average value and the instantaneous envelope according to the first curve and the second curve; A second local maximum value sequence and a second local minimum value sequence calculation submodule, used for calculating a second local maximum value sequence and a second local minimum value sequence of the instantaneous frequency; A local cutoff frequency calculation submodule, used to calculate the local cutoff frequency using the first local maximum value sequence, the first local minimum value sequence, the second local maximum value sequence and the second local minimum value sequence; A local average curve generating submodule, used for generating a local average curve according to a local cutoff frequency; A cut-off condition calculation submodule, used for calculating the cut-off condition; A return submodule is used for updating the input signal using the local average curve when the cutoff condition does not meet the preset threshold, and returning the step of calculating the instantaneous amplitude and instantaneous frequency of the input signal of the daily-scale composite salt-drought index based on the Hilbert transform; A cut-off submodule, used to terminate the decomposition when the cut-off condition meets a preset threshold, and use the input signal as a sub-signal component; Among them, the sub-composite salt-drought index forecast data calculation module includes: A normalization submodule, used for normalizing the sub-signal components to obtain normalized data; The prediction submodule is used to input the normalized data into the random forest model to obtain the predicted value of each sub-signal component; The anti-normalization submodule is used to perform anti-normalization processing on the predicted value to obtain the sub-composite salt-drought index forecast data.

7. An electronic device, characterized in that: The device comprises a processor and a memory: The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the daily composite salt-drought index forecasting method for estuary areas according to any one of claims 1-5 according to the instructions in the program code.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store program codes, and the program codes are used to execute the daily composite salt-drought index forecasting method for estuary areas according to any one of claims 1-5.