A spatial composite salt drought event monitoring and early warning method and system

Abstract

The application provides a space composite salt drought event monitoring and early warning method and system, and relates to the technical field of hydrology and water resources.Application of the prior art to upstream drought and estuary salt tide is independently monitored, the spatial coupling correlation is ignored, so that the cross-regional composite disaster risk cannot be accurately identified and evaluated, and the early warning lacks pertinence.The application obtains and pre-processes upstream runoff and estuary salinity data, respectively constructs a standardized hydrological drought index and a standardized coastal salinity index, further establishes a joint distribution model based on a binary Copula function, generates a unified space composite salt drought index, finally sets a threshold value for event determination in combination with a run theory, and divides two-level early warning according to a single index threshold value.The method realizes quantitative identification and fine grading early warning of the space composite event of upstream drought and estuary salt tide, improves the accuracy and objectivity of monitoring, and provides reliable technical support for disaster prevention decision-making.

Description

Technical Field

[0001] This invention relates to the field of hydrological and water resources application technology, and in particular to a method and system for monitoring and early warning of spatial complex salinity and drought events. Background Technology

[0002] With the intensification of global climate change and the increasing impact of human activities, water security issues in river basins are becoming increasingly complex and multifaceted. One typical disaster chain manifests as follows: upstream areas frequently experience hydrological droughts due to reduced rainfall and rising temperatures, leading to a sharp decrease in downstream runoff; simultaneously, estuaries and nearshore areas are affected by multiple factors, including rising sea levels, enhanced tidal dynamics, and weakened backwater effects from upstream freshwater, resulting in intensified saltwater intrusion and abnormally high salinity. Upstream drought and estuarine saltwater intrusion are spatially interconnected and may overlap or occur consecutively in time, forming spatially compounded saltwater-drought events. These compound events not only lead to a shortage of available freshwater resources in the region but also trigger water quality-related water scarcity, posing a serious threat to the water supply security of urban and rural areas along the river, agricultural irrigation, wetland ecological health, and the estuarine water environment.

[0003] However, existing research and practice often treat upstream drought and estuarine saltwater intrusion as independent events, and monitoring and early warning systems are often separate. This neglects the inherent physical connection and statistical dependence between the two through river system connectivity and water balance. When upstream water flow remains consistently low, freshwater dynamics in the estuary weaken, making it easier for saltwater intrusion to advance upstream. This causes the water shortage effect of drought and the water quality deterioration effect of saltwater intrusion to spatially couple and superimpose, resulting in amplified impacts. Relying solely on a single drought index or salinity threshold is insufficient to accurately identify and assess the complete risk profile of this cross-regional, multi-factor composite disaster, let alone scientifically quantify its probability of occurrence, coupling strength, and overall hazard level.

[0004] Therefore, there is an urgent need to develop a monitoring and early warning method that can couple upstream hydrological drought with the upstream intrusion of saltwater in the estuary with higher accuracy. Summary of the Invention

[0005] To address this, the present invention provides a method and system for monitoring and early warning of spatially combined drought and salinity events, which overcomes the problems in the prior art where upstream drought and estuary salinity are monitored independently, ignoring their spatial coupling relationship, resulting in the inability to accurately identify and assess cross-regional combined disaster risks and insufficient targeted early warning.

[0006] To achieve the above objectives, the present invention provides a method for monitoring and early warning of space-related combined salinity and drought events, comprising: S1, acquire runoff data of the upstream area of ​​the target river and salinity data of the estuary area; S2, preprocess the runoff data of the upstream area and the salinity data of the estuary area; S3. Based on the preprocessed runoff data of the upstream area, standardized hydrological drought indicators are determined, and based on the preprocessed salinity data of the estuary area, standardized coastal salinity indicators are determined. S4. Based on the standardized hydrological drought index and the standardized coastal salinity index, a spatial composite salinity and drought index is constructed by combining the joint distribution function. S5, Based on the run theory, the spatial composite salinity and drought index is compared with the first threshold. S6, when the spatial composite salinity and drought index is less than or equal to the first threshold, it is determined that a spatial composite salinity and drought event has occurred. S7. When it is determined that a spatial compound salinity and drought event has occurred, the standardized hydrological drought index and the standardized coastal salinity index are compared with the second threshold respectively. S8, when both the standardized hydrological drought index and the standardized coastal salinity index are less than the second threshold, a first-level warning is generated; otherwise, a second-level warning is generated. Among them, the severity of the spatial complex saline-drought event represented by the first-level warning is higher than that of the second-level warning.

[0007] Furthermore, the preprocessing of the runoff data in the upstream area and the salinity data in the estuary area includes: S21, perform quality checks on the runoff data of the upstream area and the salinity data of the estuary area, and remove data that is obviously abnormal, missing or does not conform to physical meaning; S22, imput the missing data after removal. The imputation method includes one of linear interpolation, interpolation based on the mean of adjacent time periods, or interpolation based on historical data of the same period. S23, unify the interpolated runoff data of the upstream region and the salinity data of the estuary region to the same time scale and timestamp; S24. Standardize or normalize runoff and salinity data that are unified to the same time scale and timestamp to make them comparable calculation benchmarks.

[0008] Furthermore, the process of determining the standardized hydrological drought index and the standardized coastal salinity index includes: S31, the runoff data and salinity data are fitted using several preset probability distribution functions respectively, wherein the preset probability distribution functions include log-normal distribution, logistic distribution, normal distribution, gamma distribution and Weibull distribution; S32, calculate the goodness-of-fit evaluation index corresponding to the fitting result of each probability distribution function. The goodness-of-fit evaluation index includes the Kolmogorov-Smirnov test statistic, the von Mises test statistic, the chi-square test statistic, the Akaike information criterion value, and the Bayesian information standard value. S33. Based on the goodness-of-fit evaluation index, determine the optimal probability distribution function, and based on the optimal probability distribution function, estimate the shape parameters and scale parameters using the maximum likelihood method to obtain the optimal runoff marginal distribution function and the optimal salinity marginal distribution function with determined parameters. S34, calculate the cumulative probability values ​​of runoff data and salinity data for any given time period under the optimal runoff marginal distribution function and the optimal salinity marginal distribution function; S35, input the cumulative probability value into the standard normal distribution function, and the output value is the standardized hydrological drought index and the standardized coastal salinity index for a given period. The negative value of the standardized hydrological drought index represents the runoff drought state, and the positive value represents the runoff wet state. The standardized coastal salinity index is processed by sign reversal, with negative values ​​representing the high salinity state and positive values ​​representing the low salinity state.

[0009] Furthermore, the construction of a spatial composite salinity-drought index based on the standardized hydrological drought index and the standardized coastal salinity index, combined with a joint distribution function, includes: S41, based on the optimal marginal distribution functions of the standardized hydrological drought index and the standardized coastal salinity index, these are used as the basis for constructing the joint distribution function; S42, select one from a number of pre-defined candidate binary Copula functions as the initial joint distribution function model to characterize the dependency structure between the standardized hydrological drought index and the standardized coastal salinity index, wherein the pre-defined candidate binary Copula functions include at least Clayton, Frank, Gaussian, Gumbel and Joe. S43, use the maximum likelihood method to estimate all the undetermined parameters of the selected initial joint distribution function model; S44, based on AIC, BIC and LOGLIK goodness-of-fit indices, evaluates and compares the fitting results of all preset candidate binary Copula functions; S45, Based on the evaluation results of the goodness-of-fit index, determine the optimal joint distribution function model; S46. Based on the optimal joint distribution function model, calculate the probability of simultaneous occurrence of upstream drought events and estuary high salinity events within a given time period. S47, the joint occurrence probability is mapped to a standard normal distribution through normal quantile transformation, and the obtained standard normal distribution value is the quantitative value of the spatial composite salinity and drought index.

[0010] Furthermore, the process of determining the first threshold specifically includes: S51, based on standardized hydrological drought indicators and standardized coastal salinity indicators from historical periods, constructs a historical spatial composite salinity and drought indicator sequence. S52, Apply run theory to analyze the historical spatial composite salinity and drought index sequence, and identify all negative runs in the sequence whose values ​​are continuously lower than multiple candidate thresholds; S53. Based on the physical definition of a complex saltwater drought event and the characteristics of typical historical events, a criterion for screening significant runs is established, wherein the criterion includes at least one of the following: run duration, run average intensity, and run cumulative intensity. S54, the runs that meet the significance run screening criteria are determined to be historical spatial complex saltwater and drought events; S55, by comparing the degree of consistency between the historical spatial composite saline-drought events identified under different candidate thresholds and the historical real records, the optimal candidate threshold is selected as the first threshold for monitoring and early warning.

[0011] Furthermore, the process of determining the degree of matching includes: S551, determine the timing and intensity data of confirmed spatial composite saline-drought events in historical records; S552, count the number of runs identified for each candidate threshold within the historical period where the occurrence period of the composite saltwater drought event overlaps with the occurrence period of at least one historical real event, and use this count as the number of correctly identified candidate events. S553, calculate the ratio of the number of correctly identified candidate events to the total number of historical real events, and use it as the first evaluation indicator; S554, For each correctly identified candidate event, calculate the deviation between its average run strength and the strength recorded in historical real events to generate a second evaluation index. S555, comprehensively evaluate the degree of consistency by combining the first evaluation index and the second evaluation index; wherein, the higher the value of the first evaluation index and the lower the value of the second evaluation index, the higher the degree of consistency.

[0012] Furthermore, runoff data of the upstream area of ​​the target river is obtained through hydrological observation stations and satellite remote sensing technology; salinity data of the estuary area is obtained through water quality monitoring stations and tidal stations.

[0013] Furthermore, the spatial composite salinity and drought index reflects the probability characteristics of the simultaneous occurrence of upstream drought disasters and salinity intrusion disasters, providing a quantitative basis for assessing the risk of composite disasters.

[0014] The present invention also provides a space-based combined salinity and drought event monitoring and early warning system, the system being used to implement any of the space-based combined salinity and drought event monitoring and early warning methods described above, the system comprising: The data acquisition module is used to acquire runoff data from the upstream area of ​​the target river and salinity data from the estuary area. The preprocessing module, connected to the data acquisition module, is used to preprocess the runoff data of the upstream area and the salinity data of the estuary area. The standard index generation module, connected to the preprocessing module, is used to determine standardized hydrological drought indexes based on the preprocessed runoff data of the upstream area, and to determine standardized coastal salinity indexes based on the preprocessed salinity data of the estuary area. A composite index generation module, connected to the standard index generation module, is used to construct a spatial composite salinity and drought index based on the standardized hydrological drought index and the standardized coastal salinity index, combined with a joint distribution function. The threshold comparison module, connected to the composite index generation module, is used to compare the spatial composite salinity and drought index with the first threshold based on the run theory. A composite event identification module, connected to the threshold comparison module, is used to determine that a spatial composite salinity and drought event has occurred when the spatial composite salinity and drought index is less than or equal to the first threshold. The early warning classification module, connected to the composite event identification module, is used to compare the standardized hydrological drought index and the standardized coastal salinity index with a second threshold when it is determined that a spatial composite salinity and drought event has occurred. The early warning output module, connected to the early warning classification module, is used to generate a first-level early warning when both the standardized hydrological drought index and the standardized coastal salinity index are less than the second threshold, and otherwise generate a second-level early warning.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: First, by constructing a unified spatial composite salinity and drought index, the monitoring information of two independent disasters, upstream drought and estuary salinity and intrusion, is effectively integrated, overcoming the limitations of spatial fragmentation and independent element analysis in traditional methods, and realizing the overall quantitative characterization and risk identification of cross-regional, multi-driving factor composite events. Secondly, by introducing a joint distribution model based on the Copula function and a judgment and grading mechanism based on run theory and dual thresholds, the problems of difficulty in quantifying the probability of compound events, discontinuous identification of event duration, and coarse classification of warning levels are solved, significantly improving the objectivity, accuracy, and refinement of monitoring and warning. Third, it provides a complete technical process and system implementation from data preprocessing, indicator standardization, joint modeling to event identification and early warning output, which enhances the operability and portability of the method and provides reliable technical support for unified scheduling of water resources in the basin and emergency decision-making for drought and salinity prevention. Attached Figure Description

[0016] Figure 1 A flowchart illustrating a method for monitoring and early warning of space-related combined salinity and drought events, provided by an embodiment of the present invention; Figure 2 This is a structural block diagram of a space-based composite saline-drought event monitoring and early warning system provided in an embodiment of the present invention. Detailed Implementation

[0017] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0018] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0019] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0020] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0021] Example 1 like Figure 1 As shown, the present invention provides a method for monitoring and early warning of space-related combined salinity and drought events, comprising: S1, acquire runoff data of the upstream area of ​​the target river and salinity data of the estuary area; Runoff data of the upstream area of ​​the target river is obtained through hydrological observation stations and satellite remote sensing technology; salinity data of the estuary area is obtained through water quality monitoring stations and tidal stations.

[0022] S2, preprocessing the runoff data of the upstream area and the salinity data of the estuary area, including: S21, perform quality checks on the runoff data of the upstream area and the salinity data of the estuary area, and remove data that is obviously abnormal, missing or does not conform to physical meaning; S22, imput the missing data after removal. The imputation method includes one of linear interpolation, interpolation based on the mean of adjacent time periods, or interpolation based on historical data of the same period. S23, unify the interpolated runoff data of the upstream region and the salinity data of the estuary region to the same time scale and timestamp; S24. Standardize or normalize runoff and salinity data that are unified to the same time scale and timestamp to make them comparable calculation benchmarks.

[0023] In one possible implementation, this example uses the Pearl River, specifically selecting the Modaomen Channel as the main study area. Daily average runoff data from the upstream region are collected, spanning from January 1, 1999 to December 31, 2023. Simultaneously, daily surface salinity data from fixed monitoring stations in the estuary area by marine environmental monitoring departments are collected, with the time span synchronized with the runoff data.

[0024] Judgments are made based on physical plausibility. For example, negative or extremely high values ​​in runoff data, such as exceeding twice the historical maximum flow rate without supporting typhoon or rainstorm records, are considered outliers. Similarly, high salinity values ​​in freshwater estuaries, approaching those of the open ocean (e.g., consistently exceeding 30 PSU), without corresponding reports of strong saltwater intrusion, are also considered outliers. Upon discovery, such values ​​are removed from the sequence and marked as missing. Obvious unit errors or decimal point misalignments are checked and corrected; for example, a flow rate mistakenly recorded as 100 is corrected to 10.0. Data with gaps in records or during periods of instrument malfunction in the original data are explicitly marked as missing.

[0025] For the marked missing data, the interpolation method is selected based on the duration of the missing data and the characteristics of data variation. For short-term missing data, such as consecutive missing data of 3 days or less, spline interpolation is used. For missing data over a longer period or when spline interpolation is not applicable, interpolation based on the historical average for the same period is used. For example, for the missing runoff data for February 20th of a certain year, the average runoff data for February 20th of all years from 2000 to 2023 is used to fill in the missing data.

[0026] Because the observation timestamps of runoff and salinity data may not be completely consistent—for example, runoff data might be the daily average at 12:00 Beijing time, while salinity data might be the average of multiple observations on the same day—to ensure subsequent correlation analysis, both are unified to a date dimension. Specifically, a complete date series from January 1, 2000 to December 31, 2023 is generated, ensuring that each day of this date series has exactly one corresponding value for both runoff and salinity data. Ultimately, this results in two runoff and salinity series of identical length, strictly arranged in chronological order.

[0027] Hydrological time series often exhibit trends or seasonality. To ensure the stability of subsequent statistical distribution fitting, an enhanced Dickey-Fuller test was used to perform unit root tests on the runoff and salinity series. The test results showed that both series rejected the null hypothesis of unit roots at a significance level of 0.05, indicating that the series are basically stationary. To eliminate the possible influence of long-term slow changes, a linear fitting removal method was adopted. That is, a straight line was fitted to the long-term trend of the runoff and salinity series respectively, and then the trend line was subtracted from the original series to obtain a detrended new series for subsequent analysis. This step can prevent long-term trends caused by climate change from interfering with the judgment of the relative states of drought and salinity. For diurnal scale data, obvious intra-annual seasonal cycles, such as flood season and dry season, are characteristics of the intra-annual seasonal cycle rather than noise that needs to be removed. This method aims to identify compound events during the dry season, thus preserving seasonality. In practice, subsequent distribution fitting and index calculation will be performed separately for the dry season, such as the subsequence from October to March of the following year, thereby avoiding the influence of seasonal differences.

[0028] Through the above steps, a complete, continuous, time-aligned, and statistically stable daily runoff and salinity data sequence during the dry season was finally obtained for the spatial composite salinity and drought risk assessment of Modaomen. This data was obtained after strict quality control and preprocessing.

[0029] This invention addresses the technical problem of noise, gross errors, and unreliable records directly interfering with analysis in raw monitoring data by performing quality preprocessing, including outlier removal, error correction, and missing data labeling. This improves the accuracy and reliability of input data and reduces the probability of misjudgment due to data quality issues. By employing various time series interpolation methods to specifically fill in missing values, it solves the technical problem of discontinuous data sequences and gaps caused by monitoring interruptions or malfunctions, generating complete and physically meaningful data sequences. This improves the feasibility and continuity of results in all subsequent statistical analysis steps. Furthermore, by performing stationarity tests and detrending on the aligned sequences, it solves the technical problem of long-term trends or abnormal fluctuations interfering with the extraction of short-term statistical features in hydrological time series. This makes the data sequences more consistent with the requirements of steady-state stochastic processes, improving the stability of subsequent probability distribution fitting and the accuracy of marginal distribution function estimation.

[0030] S3. Based on the preprocessed runoff data of the upstream area, standardized hydrological drought indicators are determined, and based on the preprocessed salinity data of the estuary area, standardized coastal salinity indicators are determined. The process of determining the standardized hydrological drought index and the standardized coastal salinity index includes: S31, the runoff data and salinity data are fitted using several preset probability distribution functions respectively, wherein the preset probability distribution functions include log-normal distribution, logistic distribution, normal distribution, gamma distribution and Weibull distribution; S32, calculate the goodness-of-fit evaluation index corresponding to the fitting result of each probability distribution function. The goodness-of-fit evaluation index includes the Kolmogorov-Smirnov test statistic, the von Mises test statistic, the chi-square test statistic, the Akaike information criterion value, and the Bayesian information standard value. S33. Based on the goodness-of-fit evaluation index, determine the optimal probability distribution function, and based on the optimal probability distribution function, estimate the shape parameters and scale parameters using the maximum likelihood method to obtain the optimal runoff marginal distribution function and the optimal salinity marginal distribution function with determined parameters. S34, calculate the cumulative probability values ​​of runoff data and salinity data for any given time period under the optimal runoff marginal distribution function and the optimal salinity marginal distribution function; S35, input the cumulative probability value into the standard normal distribution function, and the output value is the standardized hydrological drought index and the standardized coastal salinity index for a given period. The negative value of the standardized hydrological drought index represents the runoff drought state, and the positive value represents the runoff wet state. The standardized coastal salinity index is processed by sign reversal, with negative values ​​representing the high salinity state and positive values ​​representing the low salinity state.

[0031] In one possible implementation, five common continuous probability distribution functions applicable to hydrological and meteorological data are used to attempt to fit the preprocessed runoff and salinity sequences of the upstream area of ​​the target river. Specifically, these are the log-normal distribution (Lognorm), logis distribution, normal distribution (Normal), gamma distribution (Gamma), and Weibull distribution (Weibull). The parameters of each distribution are initially obtained through the maximum likelihood method. The Kolmogorov-Smirnov test statistic (KS) measures the maximum absolute distance between the theoretical cumulative distribution function and the empirical cumulative distribution function. A smaller value indicates a better fit. The von Mises test statistic (CV) measures the square integral of the difference between the theoretical distribution and the empirical distribution. The smaller the value, the better the fit.

[0032] The chi-square test statistic (AD) is used to compare the observed frequencies with the expected frequencies after grouping the data. The smaller the value, the better the fit.

[0033] The Akaike Information Criterion (AIC) is used to measure the goodness of fit. The smaller the value, the better the model.

[0034] The Bayesian Information Criterion (BIC) is similar to the AIC, but it imposes a heavier penalty on model complexity, making it particularly suitable for scenarios with large sample sizes. A smaller BIC value generally indicates a better model.

[0035] The following table compares the fitting results of each distribution function with each goodness-of-fit evaluation index: Table 1. Goodness of fit of marginal distribution of daily runoff in the upstream region of the target river during the dry season and salinity in the estuary region (bold black indicates the optimal value). Based on the calculated indicators, and following the principle of selecting the smallest indicator value, optimal probability distribution types were selected for both the runoff and salinity sequences. In this embodiment, the log-normal distribution had the minimum values ​​for all five indicators, and was therefore determined as the optimal marginal distribution type for the runoff sequence. Subsequently, the maximum likelihood method was used to accurately estimate the parameters of this distribution, yielding a shape parameter of 0.83 and a scale parameter of 5.4, thus obtaining the optimal marginal distribution function for runoff with completely determined parameters. The same process was performed on the salinity sequence. In this embodiment, the log-normal distribution was ultimately determined as the optimal distribution, and its shape parameter was estimated to be 2.2 and its scale parameter to be 0.4, thus obtaining the optimal marginal distribution function for salinity.

[0036] For any day during the dry season, substitute the observed runoff value into the optimal runoff marginal distribution function to obtain the cumulative probability value corresponding to the runoff state on that day, which is between 0 and 1. Similarly, substitute the salinity value on that day into the optimal salinity marginal distribution function to obtain the cumulative probability value.

[0037] The cumulative probability value corresponding to the daily runoff status is input into the inverse function of the standard normal distribution, i.e., the quantile function, to calculate the standardized hydrological drought index. The standardized hydrological drought index follows a standard normal distribution with a mean of 0 and a standard deviation of 1. Negative values ​​of the standardized hydrological drought index represent runoff drought status, while positive values ​​represent runoff wet status. The degree of drought or wetness is positively correlated with the absolute value of the standardized hydrological drought index. Similarly, the standardized coastal salinity index is calculated. To maintain consistency in the sign meaning of the standardized hydrological drought index (i.e., negative values ​​represent unfavorable conditions), the sign is reversed. Therefore, negative values ​​of the standardized coastal salinity index represent high salinity status, i.e., significant saltwater intrusion, while positive values ​​represent low salinity status. This invention solves the technical problem of inaccurate fitting caused by relying on a single distribution assumption or subjective selection of the distribution function in traditional methods by employing multiple preset probability distribution functions for competitive fitting and combining multiple statistical indicators such as the Kolmogorov-Smirnov test and the Akaike information criterion for comprehensive optimization. This improves the objectivity and accuracy of fitting univariate marginal distributions. By accurately estimating the shape and scale parameters based on the optimal distribution function and using the maximum likelihood method, it addresses the problem of inaccurate parameter estimation affecting probability calculation accuracy when only the distribution type is determined. This yields an optimal marginal distribution function with defined parameters, enabling more accurate conversion of original observations into cumulative probabilities with clear statistical significance, thus improving the accuracy of subsequent index calculations and risk probability derivation. Finally, by mapping the cumulative probability calculated by the optimal marginal distribution function to the standard normal distribution, the solution... This system solves the technical problems of the inability to directly compare and synthesize raw runoff and salinity data sequences due to differences in dimensions and distribution patterns, and the potential interference of sequence non-stationarity with analysis. It generates standardized hydrological drought indicators with a mean of 0 and a standard deviation of 1, and standardized coastal salinity indicators. This achieves dimensionless, normalized, and stationarized processing of different hydrological variables, significantly improving the operability and comparability of the data in probability space. By generating standardized univariate indicators with consistent statistical interpretation through the above system process, it solves the technical problems of traditional risk assessments that directly use raw data or simple normalized data, where threshold settings lack theoretical basis and results from different regions or periods are not comparable. It provides a common language for risk characterization with a solid theoretical foundation and unified calculation methods, improving the accuracy, repeatability, and comparability of results across different dimensions of the entire risk assessment method.

[0038] S4, Based on the standardized hydrological drought index and the standardized coastal salinity index, a spatial composite salinity-drought index is constructed by combining the joint distribution function, including: S41, based on the optimal marginal distribution functions of the standardized hydrological drought index and the standardized coastal salinity index, these are used as the basis for constructing the joint distribution function; S42, select one from a number of pre-defined candidate binary Copula functions as the initial joint distribution function model to characterize the dependency structure between the standardized hydrological drought index and the standardized coastal salinity index, wherein the pre-defined candidate binary Copula functions include at least Clayton, Frank, Gaussian, Gumbel and Joe. S43, use the maximum likelihood method to estimate all the undetermined parameters of the selected initial joint distribution function model; S44, based on AIC, BIC and LOGLIK goodness-of-fit indices, evaluates and compares the fitting results of all preset candidate binary Copula functions; S45, Based on the evaluation results of the goodness-of-fit index, determine the optimal joint distribution function model; S46. Based on the optimal joint distribution function model, calculate the probability of simultaneous occurrence of upstream drought events and estuary high salinity events within a given time period. S47, the joint occurrence probability is mapped to a standard normal distribution through normal quantile transformation, and the obtained standard normal distribution value is the quantitative value of the spatial composite salinity and drought index.

[0039] The spatial composite salinity and drought index reflects the probability characteristics of the simultaneous occurrence of upstream drought disasters and salinity intrusion disasters, providing a quantitative basis for assessing the risk of composite disasters.

[0040] In one possible implementation, the log-normal distribution functions corresponding to the calculated standardized hydrological drought index and the standardized coastal salinity index are used as the basis for constructing the joint distribution. To characterize the dependency structure between the standardized hydrological drought index and the standardized coastal salinity index, ClaytonCopula is selected as the initial joint distribution model from a pre-defined family of candidate binary Copula functions. This family of functions also includes Frank, Gaussian, Gumbel, and Joe Copula for subsequent comparisons.

[0041] The maximum likelihood method is used to estimate the undetermined parameters of all candidate Copula functions to obtain their specific parameterized forms. Subsequently, the goodness-of-fit index for each Copula function is calculated, including the Akaike information criterion, the Bayesian information criterion, and the log-likelihood value. The following is a comparison table of the goodness-of-fit indices for each Copula function: Table 2 Goodness-of-fit indices for the Copula function (bold black indicates the optimal value). In this embodiment, Clayton Copula has the smallest AIC and BIC values ​​and the largest LOGLIK value, indicating that it has the best fit among all candidate models. Therefore, Clayton Copula is determined to be the optimal joint distribution function model, which best characterizes the coupling relationship between upstream drought and estuary saltwater intrusion during the dry season in the study area.

[0042] Based on the established optimal Clayton Copula joint distribution model, for any given period within the dry season, such as January of a certain year, the corresponding standardized hydrological drought index value x and standardized coastal salinity index value y are substituted into the model. By calculating the joint distribution function value C(u, v; θ), where u and v are the cumulative probabilities of x and y under their respective optimal marginal distributions, and θ is the undetermined parameter of the Clayton Copula function, the joint probability of the combined event of upstream drought and estuary high salinity during that period can be obtained.

[0043] The calculated joint occurrence probability is transformed using a normal quantile transformation, that is, the probability value is input into the inverse function of the standard normal distribution and mapped to the corresponding quantile under the standard normal distribution. This quantile is the quantitative value of the Spatial Composite Salinity and Drought Index (SDSI) for the current period. A negative SDSI value indicates that a spatial composite salinity and drought event has occurred, and the smaller the value, the more severe the anomaly of the simultaneous occurrence of upstream drought and high salinity in the estuary, and the higher the composite intensity of the event. Thus, a unified quantitative characterization of cross-regional, multi-factor composite disasters is achieved.

[0044] This invention addresses the technical problem of treating upstream drought and estuary salinity intrusion as independent events and failing to quantify their spatial coupling relationship and joint occurrence probability in traditional methods by introducing a binary Copula function to construct a joint distribution model. This achieves an accurate characterization of the inherent statistical dependence structure of cross-regional, multi-factor composite disasters. By systematically evaluating and optimizing various Copula functions based on goodness-of-fit indices such as AIC, BIC, and LOGLIK, it solves the technical problem of model distortion or poor applicability that may result from single or subjective selection of connection functions, improving the accuracy and reliability of the joint distribution model in representing the hydrological-salinity coupling relationship in specific watersheds. By calculating the simultaneous occurrence probability of the two indicators under the joint distribution and mapping it to a standardized spatial composite salinity-drought index through normal quantile transformation, it solves the technical problems of inconsistent threshold determination criteria and lack of comprehensive quantitative support for early warning level classification in composite event monitoring, improving the accuracy and operability of the monitoring and early warning system. Furthermore, by constructing a modeling process from marginal distribution determination and joint function optimization to probability transformation, it solves the technical problems of methodological fragmentation and weak interpretability of results in composite disaster risk assessment, forming a standardized and portable technical framework and improving the method's ability to be promoted and applied in different watersheds or regions.

[0045] S5, Based on the run theory, the spatial composite salinity and drought index is compared with the first threshold. S6, when the spatial composite salinity and drought index is less than or equal to the first threshold, it is determined that a spatial composite salinity and drought event has occurred. S7. When it is determined that a spatial compound salinity and drought event has occurred, the standardized hydrological drought index and the standardized coastal salinity index are compared with the second threshold respectively. S8, when both the standardized hydrological drought index and the standardized coastal salinity index are less than the second threshold, a first-level warning is generated; otherwise, a second-level warning is generated. Among them, the severity of the spatial complex saline-drought event represented by the first-level warning is higher than that of the second-level warning.

[0046] The process of determining the first threshold specifically includes: S51, based on standardized hydrological drought indicators and standardized coastal salinity indicators from historical periods, constructs a historical spatial composite salinity and drought indicator sequence. S52, Apply run theory to analyze the historical spatial composite salinity and drought index sequence, and identify all negative runs in the sequence whose values ​​are continuously lower than multiple candidate thresholds; S53. Based on the physical definition of a complex saltwater drought event and the characteristics of typical historical events, a criterion for screening significant runs is established, wherein the criterion includes at least one of the following: run duration, run average intensity, and run cumulative intensity. S54, the runs that meet the significance run screening criteria are determined to be historical spatial complex saltwater and drought events; S55, by comparing the degree of consistency between the historical spatial composite saline-drought events identified under different candidate thresholds and the historical real records, the optimal candidate threshold is selected as the first threshold for monitoring and early warning.

[0047] The process of determining the degree of matching includes: S551, determine the timing and intensity data of confirmed spatial composite saline-drought events in historical records; S552, count the number of runs identified for each candidate threshold within the historical period where the occurrence period of the composite saltwater drought event overlaps with the occurrence period of at least one historical real event, and use this count as the number of correctly identified candidate events. S553, calculate the ratio of the number of correctly identified candidate events to the total number of historical real events, and use it as the first evaluation indicator; S554, For each correctly identified candidate event, calculate the deviation between its average run strength and the strength recorded in historical real events to generate a second evaluation index. S555, comprehensively evaluate the degree of consistency by combining the first evaluation index and the second evaluation index; wherein, the higher the value of the first evaluation index and the lower the value of the second evaluation index, the higher the degree of consistency.

[0048] In one possible implementation, a historical spatial composite salinity-drought index sequence is constructed based on standardized hydrological drought indices and standardized coastal salinity indices from historical periods, following the method described in S4. Subsequently, a set of candidate first thresholds is set, for example, [-0.3, -0.5, -0.7]. Using run theory, the entire historical spatial composite salinity-drought index sequence is scanned using each candidate threshold as a standard. Periods in which the spatial composite salinity-drought index value is below the threshold for several consecutive months are identified; each such consecutive period constitutes a negative run. Not all negative runs constitute a composite event with actual impact. Therefore, a significance screening criterion is set. For example, a significant composite event run must simultaneously satisfy a duration greater than or equal to 2 months and an average run intensity, i.e., the mean of the spatial composite salinity-drought index within that period is less than or equal to -1.0. Based on historical documents and disaster records, such as water supply shortage records in estuary areas in 2004-2005 and 2011-2012, these are identified as actual historical spatial composite salinity-drought events. In this embodiment, a total of 5 real historical events were confirmed.

[0049] For each candidate threshold, the number of runs that satisfy the significance criterion is counted to determine how many runs overlap with the time periods of the aforementioned 5 real historical events, for example, at least one month. In this embodiment, the candidate threshold of -0.5 identifies 6 significant runs, 5 of which correspond to the 5 real events, thus the number of correctly identified candidate events is 5.

[0050] The correct recognition rate, i.e., the first evaluation metric, is calculated. For a candidate threshold of -0.5, the first evaluation metric is 1. For a candidate threshold of -0.3, 8 runs are identified due to the threshold being too wide, of which 4 correspond to real events, including false alarms; the first evaluation metric is 0.8. For a candidate threshold of -0.7, only 3 runs corresponding to 3 events are identified due to the threshold being too strict; the first evaluation metric is 0.6. For each correctly identified event, the average absolute deviation between its average run intensity and the intensity level described in the corresponding historical event record is calculated as the second evaluation metric to measure the accuracy of intensity identification.

[0051] The overall assessment principle for matching accuracy is to pursue a high correct recognition rate and low intensity bias. In this example, candidate threshold -0.5 achieves the highest correct recognition rate while maintaining an acceptable low level of intensity bias. Candidate threshold -0.3, although having a reasonable correct recognition rate, introduces false alarms, increasing early warning costs; candidate threshold -0.7 suffers from severe false negatives and has a low correct recognition rate. Therefore, considering all factors, candidate threshold -0.5 is determined to be the optimal candidate threshold and is thus selected as the first threshold for monitoring and early warning.

[0052] Similarly, a second threshold for detection and early warning is determined. In this embodiment, -1.0 is taken as the second threshold.

[0053] During the monitoring period, the spatial composite salinity and drought index value is calculated monthly. When the spatial composite salinity and drought index value for the first time in a certain month is less than or equal to -0.5, it is determined to be the starting point of a spatial composite salinity and drought event. If the spatial composite salinity and drought index value is less than or equal to -0.5 for several consecutive months thereafter, these consecutive months together constitute a spatial composite salinity and drought event. For example, if monitoring found that the spatial composite salinity and drought index value was less than or equal to -0.5 for three consecutive months from December 2023 to February 2024, then it is determined that a spatial composite salinity and drought event occurred during this period.

[0054] For events that have been identified, extract the standardized hydrological drought index and standardized coastal salinity index values ​​for each month during the event duration. If the monthly values ​​of the standardized hydrological drought index and standardized coastal salinity index are all less than the second threshold throughout the entire event period, it indicates that the upstream drought and estuary saltwater intrusion have both reached severe or above levels, and the combined effect is serious. The system generates a first-level warning, prompting the need to activate the highest level of emergency response. If, during the event period, the monthly values ​​of the standardized hydrological drought index and the standardized coastal salinity index are not all simultaneously below the second threshold—for example, in some months the standardized hydrological drought index is below the second threshold but the standardized coastal salinity index is above the second threshold, or vice versa, or neither of them is consistently below the second threshold—it indicates that the intensity of at least one element has not consistently reached an extreme level, generating a Level 2 warning, indicating a high risk and requiring strengthened monitoring and preparedness.

[0055] This invention analyzes continuous SDSI sequences using run-length analysis and sets a first threshold as the event triggering criterion. This solves the technical problems of traditional single-point threshold methods, which cannot effectively identify event duration and easily fragment continuous disaster processes into multiple isolated signals. It achieves accurate capture of the complete lifecycle of composite events, improving the continuity and completeness of event identification. By calibrating the first threshold based on historical SDSI sequences, combined with historical real disaster records and physical criteria, and using a quantitative assessment of the degree of fit, this invention addresses the technical problems of high subjectivity and low matching degree with historical disasters caused by relying on expert experience or simple statistical quantiles for early warning thresholds. This improves the objectivity, scientific rigor, and localization accuracy of monitoring threshold setting, reducing the probability of false alarms and missed alarms. Furthermore, after determining the occurrence of a composite event, a second threshold is introduced to logically determine the warning level based on two single indicators: standardized hydrological drought index and standardized coastal salinity index. This solves the problem of... Relying solely on spatial composite salinity and drought indicators for single-level early warning cannot distinguish the differences in the intensity combination of internal driving factors of an event, resulting in coarse early warning information. This approach enables a more refined differentiation of the severity within composite events, improving the relevance and decision-making support value of early warning information. By constructing a two-stage early warning process that first determines whether an event has occurred and then classifies the event's internal intensity, this approach solves the technical problem of traditional methods having a single early warning trigger condition and being unable to simultaneously consider the probability of event occurrence and the degree of comprehensive impact, thus improving the structural rationality and operational reliability of the entire monitoring and early warning system. By directly linking the early warning level with the specific intensity indicators of the two core elements—upstream drought and estuary salinity intrusion—this approach solves the technical problem of the disconnect between early warning conclusions and the state of disaster-causing factors, hindering the tracing of risk roots and the formulation of targeted mitigation measures. This ensures that the early warning results not only indicate the level of risk but also reveal the main sources of risk, providing a direct basis for the basin to adopt specific response strategies such as water diversion and salinity suppression.

[0056] Example 2 like Figure 2 As shown, the present invention also provides a space-based combined salinity and drought event monitoring and early warning system. The system is used to implement the space-based combined salinity and drought event monitoring and early warning method described in any of Embodiment 1. The system includes: The data acquisition module is used to acquire runoff data from the upstream area of ​​the target river and salinity data from the estuary area. The preprocessing module, connected to the data acquisition module, is used to preprocess the runoff data of the upstream area and the salinity data of the estuary area. The standard index generation module, connected to the preprocessing module, is used to determine standardized hydrological drought indexes based on the preprocessed runoff data of the upstream area, and to determine standardized coastal salinity indexes based on the preprocessed salinity data of the estuary area. A composite index generation module, connected to the standard index generation module, is used to construct a spatial composite salinity and drought index based on the standardized hydrological drought index and the standardized coastal salinity index, combined with a joint distribution function. The threshold comparison module, connected to the composite index generation module, is used to compare the spatial composite salinity and drought index with the first threshold based on the run theory. A composite event identification module, connected to the threshold comparison module, is used to determine that a spatial composite salinity and drought event has occurred when the spatial composite salinity and drought index is less than or equal to the first threshold. The early warning classification module, connected to the composite event identification module, is used to compare the standardized hydrological drought index and the standardized coastal salinity index with a second threshold when it is determined that a spatial composite salinity and drought event has occurred. The early warning output module, connected to the early warning classification module, is used to generate a first-level early warning when both the standardized hydrological drought index and the standardized coastal salinity index are less than the second threshold, and otherwise generate a second-level early warning.

[0057] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A method for monitoring and early warning of spatial complex salinity and drought events, characterized in that, include: S1, acquire runoff data of the upstream area of ​​the target river and salinity data of the estuary area; S2, preprocess the runoff data of the upstream area and the salinity data of the estuary area; S3. Based on the preprocessed runoff data of the upstream area, standardized hydrological drought indicators are determined, and based on the preprocessed salinity data of the estuary area, standardized coastal salinity indicators are determined. S4. Based on the standardized hydrological drought index and the standardized coastal salinity index, a spatial composite salinity and drought index is constructed by combining the joint distribution function. S5, based on run theory, compare the spatial composite salinity and drought index with the first threshold. S6, when the spatial composite salinity and drought index is less than or equal to the first threshold, it is determined that a spatial composite salinity and drought event has occurred. S7. When it is determined that a spatial compound salinity and drought event has occurred, the standardized hydrological drought index and the standardized coastal salinity index are compared with the second threshold respectively. S8, when both the standardized hydrological drought index and the standardized coastal salinity index are less than the second threshold, a first-level warning is generated; otherwise, a second-level warning is generated. Among them, the severity of the spatial complex saline-drought event represented by the first-level warning is higher than that of the second-level warning.

2. The method for monitoring and early warning of space-related combined salinity and drought events according to claim 1, characterized in that, The preprocessing of runoff data from the upstream region and salinity data from the estuary region includes: S21, perform quality checks on the runoff data of the upstream area and the salinity data of the estuary area, and remove data that is obviously abnormal, missing or does not conform to physical meaning; S22, imput the missing data after removal, and the imputation method includes one of linear interpolation, interpolation based on the mean of adjacent time periods or interpolation based on historical data of the same period; S23, unify the interpolated runoff data of the upstream region and the salinity data of the estuary region to the same time scale and timestamp; S24. Standardize or normalize runoff and salinity data that are unified to the same time scale and timestamp to make them comparable calculation benchmarks.

3. The method for monitoring and early warning of space-related combined salinity and drought events according to claim 1, characterized in that, The process of determining the standardized hydrological drought index and the standardized coastal salinity index includes: S31, several preset probability distribution functions are used to fit the runoff data and salinity data respectively, wherein the preset probability distribution functions include log-normal distribution, gamma distribution, Weibull distribution, logistic distribution and normal distribution; S32, calculate the goodness-of-fit evaluation index corresponding to the fitting result of each probability distribution function. The goodness-of-fit evaluation index includes the Kolmogorov-Smirnov test statistic, the von Mises test statistic, the chi-square test statistic, the Akaike information criterion value, and the Bayesian information standard value. S33. Based on the goodness-of-fit evaluation index, determine the optimal probability distribution function, and based on the optimal probability distribution function, estimate the shape parameters and scale parameters using the maximum likelihood method to obtain the optimal runoff marginal distribution function and the optimal salinity marginal distribution function with determined parameters. S34, calculate the cumulative probability values ​​of runoff data and salinity data for any given time period under the optimal runoff marginal distribution function and the optimal salinity marginal distribution function; S35, the cumulative probability value is mapped to a standard normal distribution, and the output value is the standardized hydrological drought index and the standardized coastal salinity index for a given period. The negative value of the standardized hydrological drought index represents the runoff drought state, and the positive value represents the runoff wet state. The standardized coastal salinity index is processed by sign reversal, with negative values ​​representing the high salinity state and positive values ​​representing the low salinity state.

4. The method for monitoring and early warning of space-related combined salinity and drought events according to claim 1, characterized in that, The construction of a spatial composite salinity and drought index based on the standardized hydrological drought index and the standardized coastal salinity index, combined with a joint distribution function, includes: S41, based on the optimal marginal distribution functions of the standardized hydrological drought index and the standardized coastal salinity index, these are used as the basis for constructing the joint distribution function; S42, select one from a number of pre-defined candidate binary Copula functions as the initial joint distribution function model to characterize the dependency structure between the standardized hydrological drought index and the standardized coastal salinity index, wherein the pre-defined candidate binary Copula functions include at least Clayton, Frank, Gaussian, Gumbel and Joe. S43, use the maximum likelihood method to estimate all the undetermined parameters of the selected initial joint distribution function model; S44. Based on AIC, BIC and log-likelihood goodness-of-fit indices, the fitting results of all preset candidate binary Copula functions are evaluated and compared. S45, Based on the evaluation results of the goodness-of-fit index, determine the optimal joint distribution function model; S46. Based on the optimal joint distribution function model, calculate the probability of simultaneous occurrence of upstream drought events and estuary high salinity events within a given time period. S47, the joint occurrence probability is mapped to a standard normal distribution through normal quantile transformation, and the obtained standard normal distribution value is the quantitative value of the spatial composite salinity and drought index.

5. The method for monitoring and early warning of space-related combined salinity and drought events according to claim 1, characterized in that, The process of determining the first threshold specifically includes: S51, based on standardized hydrological drought indicators and standardized coastal salinity indicators from historical periods, constructs a historical spatial composite salinity and drought indicator sequence. S52, Apply run theory to analyze the historical spatial composite salinity and drought index sequence, and identify all negative runs in the sequence whose values ​​are continuously lower than multiple candidate thresholds; S53. Based on the physical definition of a complex saltwater drought event and the characteristics of typical historical events, a criterion for screening significant runs is established, wherein the criterion includes at least one of the following: run duration, run average intensity, and run cumulative intensity. S54, the runs that meet the significance run screening criteria are determined to be historical spatial complex saltwater and drought events; S55, by comparing the degree of consistency between the historical spatial composite saline-drought events identified under different candidate thresholds and the historical real records, the optimal candidate threshold is selected as the first threshold for monitoring and early warning.

6. The method for monitoring and early warning of space-related combined salinity and drought events according to claim 5, characterized in that, The process of determining the degree of matching includes: S551, determine the timing and intensity data of confirmed spatial composite saline-drought events in historical records; S552, count the number of runs identified for each candidate threshold within the historical period where the occurrence period of the composite saltwater drought event overlaps with the occurrence period of at least one historical real event, and use this count as the number of correctly identified candidate events. S553, calculate the ratio of the number of correctly identified candidate events to the total number of historical real events, and use it as the first evaluation indicator; S554, For each correctly identified candidate event, calculate the deviation between its average run strength and the strength recorded in historical real events to generate a second evaluation index. S555, comprehensively evaluate the degree of consistency by combining the first evaluation index and the second evaluation index; wherein, the higher the value of the first evaluation index and the lower the value of the second evaluation index, the higher the degree of consistency.

7. The method for monitoring and early warning of space-related combined salinity and drought events according to claim 1, characterized in that, Runoff data of the upstream area of ​​the target river is obtained through hydrological observation stations and satellite remote sensing technology; salinity data of the estuary area is obtained through water quality monitoring stations and tidal stations.

8. The method for monitoring and early warning of space-related combined salinity and drought events according to claim 1, characterized in that, The spatial composite salinity and drought index reflects the probability characteristics of the simultaneous occurrence of upstream drought disasters and salinity intrusion disasters, providing a quantitative basis for assessing the risk of composite disasters.

9. A space-based composite salinity and drought event monitoring and early warning system, characterized in that, The system is used to implement the space-based composite saline-drought event monitoring and early warning method according to any one of claims 1 to 8, the system comprising: The data acquisition module is used to acquire runoff data from the upstream area of ​​the target river and salinity data from the estuary area. The preprocessing module, connected to the data acquisition module, is used to preprocess the runoff data of the upstream area and the salinity data of the estuary area. The standard index generation module, connected to the preprocessing module, is used to determine standardized hydrological drought indexes based on the preprocessed runoff data of the upstream area, and to determine standardized coastal salinity indexes based on the preprocessed salinity data of the estuary area. A composite index generation module, connected to the standard index generation module, is used to construct a spatial composite salinity and drought index based on the standardized hydrological drought index and the standardized coastal salinity index, combined with a joint distribution function. The threshold comparison module, connected to the composite index generation module, is used to compare the spatial composite salinity and drought index with the first threshold based on the run theory. A composite event identification module, connected to the threshold comparison module, is used to determine that a spatial composite salinity and drought event has occurred when the spatial composite salinity and drought index is less than or equal to the first threshold. The early warning classification module, connected to the composite event identification module, is used to compare the standardized hydrological drought index and the standardized coastal salinity index with a second threshold when it is determined that a spatial composite salinity and drought event has occurred. The early warning output module, connected to the early warning classification module, is used to generate a first-level early warning when both the standardized hydrological drought index and the standardized coastal salinity index are less than the second threshold, and otherwise generate a second-level early warning.

Images