Regional drought level identification method, system, electronic device and storage medium
By acquiring and analyzing multiple drought characteristic data in the drought data set, the dynamic threshold of drought events is determined, and the problem of low accuracy in determining drought levels in the existing technology is solved, and timely and accurate identification and early warning of drought events is achieved. A more scientific data foundation.
Patent Information
- Application Number
- CN202410262623.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-03-07
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-03-07
AI Technical Summary
The prior art is difficult to effectively identify and early warning of dynamic changes in drought events, especially when the rank determination accuracy is low during the occurrence and development of drought events, and it is impossible to identify the severity of regional droughts in a timely and accurate manner.
By obtaining the drought data set of the target area, multiple drought characteristic data of drought events at different times are extracted, and dynamic thresholds are determined based on the dependence between drought characteristics, and drought levels are dynamically identified, including light drought, heavy drought and special drought levels. .
It has achieved timely and accurate hierarchical identification of drought events during the occurrence and development, reduced the risk of misjudgment, provided a more scientific data basis, and provided technical support for drought resistance measures.
Smart Images

Figure CN118114147B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of natural science research technology, and more specifically to a method for identifying a regional drought level, a system for identifying a regional drought level, an electronic device, a storage medium, and a computer program product. Background Art
[0002] Affected by global warming, the frequency, intensity, range and other characteristics of drought have changed significantly, posing a huge challenge to the early identification and early warning of extreme drought.
[0003] Drought index is one of the commonly used methods to quantify drought. When the drought index is less than a certain threshold, it is judged as a certain degree of drought level. In the prior art, some schemes directly use the difference between the mean and standard deviation of the drought index to divide the drought level. However, this type of method belongs to the point scale division and does not consider the spatiotemporal scale of the drought event. Other schemes quantitatively represent the intensity level of regional drought by multiplying multiple drought characteristics including drought index and drought duration. However, the regional drought characteristics identified by this method are the characteristics of the complete drought event that has occurred extracted after the drought event ends, and are used to evaluate the intensity level of the drought event, which to a certain extent belongs to post-disaster evaluation.
[0004] The occurrence and development of drought events is a dynamic process. The existing technologies have failed to consider the characteristics of the dynamic changes in the extreme nature of drought events during the formation process, making it difficult to achieve effective early warning of drought. Summary of the invention
[0005] The present application is proposed in view of the above problems. In order to at least partially solve the above technical problems, according to one aspect of the present application, a method for identifying regional drought levels is provided, comprising:
[0006] Obtain a drought data set of a target area, wherein the drought data set includes drought data of multiple drought events occurring in the area during a statistical period in various time periods;
[0007] For each drought data, determining multiple characteristic data of multiple drought characteristics of the drought data at least according to the drought data, the associated data of the drought data in the data set, and the duration from a first time period to a second time period, wherein the drought data and the associated data correspond to the same drought event, the first time period is the time period when the drought event corresponding to the drought data starts, the second time period is the collection time period of the drought data and the collection time period of the associated data is before the second time period;
[0008] Analyzing multiple characteristic data sequences of the target area, and determining the dependency relationship between multiple drought characteristics according to the analysis results, wherein the multiple characteristic data sequences are generated by each characteristic data of each drought data in the data set, and each characteristic data in each characteristic data sequence corresponds to the same drought characteristic; and
[0009] According to the dependency relationship between the plurality of drought characteristics, a plurality of thresholds are determined, wherein the plurality of thresholds are used to determine the drought level of a drought event occurring in a target area during its development process.
[0010] Exemplarily, the threshold includes a joint probability threshold, and the dependency between multiple drought characteristics is determined, including:
[0011] Fitting a joint probability distribution model between multiple drought characteristics, wherein the joint probability distribution model is used to describe the distribution of joint probabilities between multiple characteristic data of each drought data in the drought data set;
[0012] Determine multiple thresholds, including:
[0013] A plurality of joint probability thresholds are determined based on the probability distribution of drought indices corresponding to different drought levels in a joint probability distribution model and a pre-fitted drought index probability distribution model, wherein the drought index probability distribution model is used to represent the probability distribution of each drought indices in a drought index library, and each drought index includes a first drought index corresponding to a drought event and a second drought index corresponding to a non-drought event.
[0014] Exemplarily, determining a plurality of joint probability thresholds includes:
[0015] According to the drought index probability distribution model and the drought index thresholds corresponding to different drought levels, determining the proportion of the first drought index corresponding to each drought level in the first drought index group, wherein the first drought index group is composed of the first drought indices in the drought index library;
[0016] According to the joint probability distribution model, the joint probability value between multiple characteristic data of each drought data in the data set is determined;
[0017] Arranging the determined joint probability values in ascending order to determine the ordinal percentile of each joint probability value;
[0018] Multiple joint probability thresholds are determined based on the proportions and the ordinal percentiles of the respective joint probability values.
[0019] Exemplarily, the types of drought levels include light drought, moderate drought, severe drought and extreme drought, the determined coverage includes a first ratio, a second ratio, a third ratio and a fourth ratio corresponding to the light drought, moderate drought, severe drought and extreme drought, respectively, and the multiple joint probability thresholds include the first probability threshold, the second probability threshold and the third probability threshold, and the multiple joint probability thresholds are determined, including:
[0020] According to the joint probability value of the order percentile being equal to the fourth proportion, a third probability threshold is determined, and the third probability threshold represents the dividing threshold between the extreme drought level and the severe drought level;
[0021] Determine a second probability threshold value according to a joint probability value of the order percentile being equal to the fifth proportion, wherein the fifth proportion is equal to the sum of the third proportion and the fourth proportion, and the second probability threshold value represents a dividing threshold value between the severe drought level and the moderate drought level;
[0022] The first probability threshold is determined according to the joint probability value whose order percentile is equal to the sixth proportion, where the sixth proportion is equal to the sum of the second proportion, the third proportion and the fourth proportion, and the first probability threshold represents the dividing threshold between moderate drought and mild drought.
[0023] Exemplarily, the drought data is a drought index, the plurality of characteristic data at least include the cumulative duration and cumulative severity of the drought event corresponding to the drought data up to the second time period, and the plurality of characteristic data of the plurality of drought characteristics of the drought data are determined, including:
[0024] Calculate the total number of unit periods from the first period to the second period as the cumulative number of duration periods;
[0025] The absolute values of the drought data and all associated data of the drought data are summed to determine the cumulative severity.
[0026] Exemplarily, the method further comprises:
[0027] The drought level of the drought event corresponding to the drought data in the specific time period is determined based on at least multiple thresholds and drought data in a specific time period, wherein the specific time period is any time period among the various time periods or the specific time period is any time period in the development process of the drought event occurring in the target area after the statistical period.
[0028] Exemplarily, determining the drought level of the drought event corresponding to the drought data in a specific period includes:
[0029] Matching each characteristic data of drought data in a specific period with the joint probability model to determine the joint probability value corresponding to the drought data in the period as the specific joint probability;
[0030] If the specific joint probability is greater than the first probability threshold, determining that the drought level of the first drought event in the period is a light drought level, wherein the first drought event is the drought event corresponding to the drought data;
[0031] If the specific joint probability is greater than the second probability threshold and less than or equal to the first probability threshold, determining that the drought level of the first drought event in the period is moderate drought level;
[0032] If the specific joint probability is greater than the third probability threshold and less than or equal to the second probability threshold, determining that the drought level of the first drought event in the period is a severe drought level;
[0033] If the specific joint probability is less than or equal to the third probability threshold, it is determined that the drought level of the first drought event in the period is an extreme drought level.
[0034] Exemplarily, the statistical period is from the statistical start period to the current period, the specific period is the current period, and the method further includes:
[0035] If it is determined that the drought level of the drought event in the current period is greater than the preset drought level, prompt information indicating drought warning is output.
[0036] Exemplarily, after determining the plurality of thresholds, the method further comprises:
[0037] In each time period of at least one monitoring period after the statistical period, the drought level of the drought event corresponding to the drought data in the time period is determined based on at least a plurality of thresholds and the drought data in the time period.
[0038] Exemplarily, the at least one monitoring period is a first monitoring period, and the method further comprises:
[0039] At the end of each monitoring period, the drought data of each period in the monitoring period is added to the existing drought data set to obtain an updated drought data set;
[0040] Based on each drought characteristic data of each drought data in the updated drought data set, the dependency relationship between multiple drought characteristics is updated, and multiple thresholds are updated to obtain multiple updated thresholds; and
[0041] For each time period in the next monitoring period, the drought level of the drought event corresponding to the drought data in the time period is determined at least based on the updated multiple thresholds and the drought data in the time period.
[0042] Exemplarily, determining dependencies between multiple drought characteristics includes:
[0043] For each feature data sequence,
[0044] Fitting multiple initial marginal distribution models about the feature data sequence;
[0045] According to the first screening method, a marginal distribution model as the feature data sequence is screened out from a plurality of initial marginal distribution models;
[0046] According to the marginal distribution model of each feature data sequence, multiple initial joint probability distribution models about multiple feature data sequences are fitted;
[0047] According to the second screening method, a first joint probability distribution model between multiple drought characteristics is screened out from each of the fitted initial joint probability distribution models, wherein the first joint probability distribution model is used to represent the probability distribution of multiple drought characteristic data being less than or equal to a specific value at the same time;
[0048] According to the marginal distribution model of each characteristic data sequence and the screened first joint probability distribution model, a second joint probability distribution model between multiple drought characteristics is generated, wherein the second joint probability distribution model is used to represent the probability distribution situation that multiple drought characteristic data are simultaneously greater than or equal to a specific value.
[0049] According to another aspect of the present application, a regional drought level identification system is also provided, comprising:
[0050] An acquisition module is used to acquire a drought data set of a target area, wherein the drought data set includes drought data of multiple drought events occurring in the area during a statistical period in various time periods;
[0051] A first determination module is used to determine, for each drought data, a plurality of characteristic data of a plurality of drought characteristics of the drought data at least according to the drought data, the associated data of the drought data in the data set, and the duration from a first time period to a second time period, wherein the drought data and the associated data correspond to the same drought event, the first time period is a time period when the drought event corresponding to the drought data starts, the second time period is a time period for collecting the drought data, and the time period for collecting the associated data is before the second time period;
[0052] an analysis module, for analyzing a plurality of characteristic data sequences of a target area, and determining a dependency relationship between a plurality of drought characteristics according to the analysis result, wherein the plurality of characteristic data sequences are generated by respective characteristic data of respective drought data in a data set, and each characteristic data in each characteristic data sequence corresponds to the same drought characteristic; and
[0053] The second determination module is used to determine multiple thresholds according to the dependency relationship between multiple drought characteristics, wherein the multiple thresholds are used to determine the drought level of the drought event occurring in the target area during the development process, and different thresholds correspond to different drought levels.
[0054] According to another aspect of the present application, an electronic device is provided, including: a processor and a memory, wherein the memory stores computer program instructions, and the computer program instructions are used to execute the above-mentioned regional drought level identification method when the processor is executed.
[0055] According to another aspect of the present application, a storage medium is provided, on which program instructions are stored, and the program instructions are used to execute the above-mentioned regional drought level identification method when running.
[0056] According to another aspect of the present application, a computer program product is also provided, including computer program instructions, which are used to execute the above-mentioned regional drought level identification method when running.
[0057] As mentioned above, since the prior art usually analyzes the entirety of the drought event that has ended, the scheme for determining the level of the quantified drought event in the prior art is usually only used to determine the drought level of the drought event that has ended, which belongs to post-disaster evaluation to a certain extent, and cannot effectively evaluate or warn the level of the drought event in the process of occurrence and development. In most cases, the accuracy of determining the degree of drought in the process of drought event occurrence and development is low, especially when the amount of data of the existing drought event is limited, the accuracy of the quantified drought level determination standard or the determined determination threshold is lower. According to the above-mentioned regional drought level identification method of the embodiment of the present application, based on the characteristics of the dynamic change of drought level in different periods during the occurrence and development of drought events, on the basis of collecting drought data of different drought events in different periods of occurrence and development, according to the cumulative duration and accumulated drought data of each drought event from the beginning to each period of development, the multiple characteristic data of multiple drought characteristics of each drought event at the end of each period are accurately extracted. And by analyzing the generated multiple characteristic data sequences, the dependency relationship between multiple drought characteristics of the target area is accurately determined. Finally, according to the determined dependencies, multiple thresholds that can dynamically identify the drought level of drought events during their occurrence and development are quickly and accurately determined. This scheme fully considers the characteristics of the dynamic changes in the extreme nature of drought events during their formation, extracts and analyzes feature data from the perspective of drought data at each period of each drought event during its occurrence and development, and provides a more scientific and complete data basis for dynamically identifying the drought level of drought events. It does not rely on complete drought events, making drought level identification more timely, more accurate, and more widely applicable. In addition, the correlation between the drought feature data extracted for each drought data and the severity of drought in this scheme is high, which can ensure that the determined threshold is more accurate. Moreover, the scheme that determines the threshold for determining the drought level based on the dependency between each drought feature can more comprehensively understand the complexity of drought, effectively reduce the risk of misjudgment, and further improve the accuracy of the determined threshold. This scheme can not only quantitatively analyze the level of regional drought events that have occurred, but also dynamically distinguish the severity level of regional drought events that are occurring, providing technical support for effective drought relief measures, and has certain practical significance and social value.
[0058] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0059] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.
[0060] Figure 1 A schematic flow chart of a method for identifying regional drought levels according to an embodiment of the present application is shown;
[0061] Figure 2 A graph showing the change of drought data over time in a drought event according to one embodiment of the present application;
[0062] Figure 3a A schematic diagram showing the dynamic change process of drought levels of various drought events in different target areas during a statistical period at different time periods according to an embodiment of the present application;
[0063] Figure 3b A schematic diagram showing the proportion of different drought levels in different target areas during the development of drought events during the statistical period according to an embodiment of the present application;
[0064] Figure 4 A flow chart showing a method for identifying regional drought levels according to another embodiment of the present application is shown;
[0065] Figure 5 A schematic block diagram of a regional drought level identification system according to an embodiment of the present application is shown; and
[0066] Figure 6 A schematic block diagram of an electronic device according to an embodiment of the present application is shown. DETAILED DESCRIPTION
[0067] In order to make the purpose, technical solutions and advantages of the present application more obvious, the example embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application described in the present application, all other embodiments obtained by those skilled in the art without creative work should fall within the scope of protection of the present application.
[0068] In order to at least partially solve the above problems, an embodiment of the present application provides a method for identifying regional drought levels. Figure 1 FIG. 2 shows a schematic flow chart of a method for identifying a regional drought level according to an embodiment of the present application. Figure 1As shown, the method includes the following steps S120, S140, S160 and S180.
[0069] Step S120, obtaining a drought data set of the target area, wherein the drought data set includes drought data of multiple drought events occurring in the area during the statistical period in various time periods.
[0070] The target area may be any suitable area of desired concern, and this application is not limited thereto.
[0071] The number of target regions can be one or more. For example, the number of target regions is 3, namely North China, the Yangtze River Basin and Yunnan Province. In this step, drought datasets of these three target regions can be obtained respectively.
[0072] The drought data may be any suitable drought index that can reflect the drought situation of the target area. In the embodiment of the present application, each drought data in the drought data set corresponds to a drought event. That is, the drought data of the target area is the value of the drought index when the target area is in a drought state. Exemplarily, a drought data may include one or more drought index values. Each drought data may be a value of the drought index of a drought event that has occurred in the target area or a drought event that is occurring in a certain unit period, which is screened out by a preset screening method. Exemplarily, the drought index may include one or more of the standardized precipitation index (SPI), the standardized precipitation evapotranspiration index (SPEI), the Palmer drought index (PDSI), and the meteorological drought comprehensive index (MCI). Of course, the drought index may also be represented by data such as precipitation, soil moisture, and soil temperature. Taking the drought index as SPEI as an example, various suitable calculation methods may be used to count the SPEI values of each unit period of the target area in the statistical period, and all SPEI values less than the SPEI drought threshold may be screened out, thereby generating a drought data set with a timestamp. The statistical period may have a longer time span, such as a statistical period greater than or equal to 10 years. The unit period (time scale) can be any time scale such as one week, one month, two months, three months, half a year, etc. In a specific example, the target area includes North China, the Yangtze River Basin and Yunnan Province, the statistical period can be from 1960 to the present, the unit period can be one month, the drought index can be SPEI, and the SPEI drought threshold can be -0.5. Specifically, the monthly precipitation and temperature (maximum temperature, minimum temperature, average temperature) sequences of the three target areas during the above statistical period can be obtained from, for example, the National Meteorological Science Data Center. Based on the precipitation and temperature data, and referring to the method for calculating the standard precipitation evapotranspiration index in the "Meteorological Drought Level", the SPEI values of each of the three target areas in each month can be calculated respectively, and the SPEI values of the months with SPEI less than or equal to -0.5 can be screened out, thereby generating a drought data set with timestamps for each target area.
[0073] In some examples, the target area may include multiple sites, and each drought data may be obtained by screening the drought index values of each site in the target area in the same period of the statistical period and performing statistical analysis using a suitable statistical method.
[0074] For example, the target area includes North China (including Inner Mongolia), which covers an area of 880,000 km 2, including 401 meteorological stations in the region. Therefore, in North China, the values of drought index of each of the 401 meteorological stations can be obtained in each month during the statistical period. Similarly, taking the drought index as SPEI and the SPEI drought threshold as -0.5 as an example, the drought stations in North China for each month (meteorological stations with SPEI values less than or equal to -0.5) can be screened out, and the drought data for each month in North China can be calculated based on the drought index of these drought stations.
[0075] Specifically, we can first count the number of stations (represented as m) whose drought index SPEI is less than or equal to -0.5 in the i-th month (time scale) of the statistical period in North China, and then use the following formula to calculate the ratio of the sum of the drought indexes of the m drought stations to the total number of all stations n (for example, 401) in North China as the drought data R of North China for that month: d,i :
[0076]
[0077] Among them, Drought_index k It represents the value of the drought index SPEI of the kth drought station in the North China region in the i-th month. For example, if the SPEI values of all stations in the target area in a certain month are greater than -0.5, the drought data of that month can be determined as 0, or set as a missing value and not counted.
[0078] According to the embodiment of the present application, the drought data R of the target area in each period during the statistical period is calculated by the above method. d,i After that, an initial drought data set can be generated. Then, drought data that meets the filtering requirements in the initial drought data set can be filtered out according to the drought data threshold R0. d,i Taking the drought index SPEI as an example, the drought data threshold R0 may be the SPEI value -0.5, and the drought data that meets the filtering requirements may be the drought data with a value greater than -0.5. Thus, the filtered drought data set of the target area is obtained. Figure 2 A graph showing the change of drought data over time in a drought event according to an embodiment of the present application is shown. As shown in the figure, each dot in the figure represents a drought data, and according to the drought data threshold, the drought data represented by the white dots can be filtered out, and the remaining drought data (drought data represented by the gray dots and black dots in the figure) are generated into a drought data set according to the collection time.
[0079] Of course, in addition to the above method for obtaining the drought dataset of the target area, other suitable methods may be used to obtain the drought dataset of the target area, which is not limited in the present application.
[0080] Step S140, for each drought data, multiple characteristic data of multiple drought characteristics of the drought data are determined at least according to the drought data, the associated data of the drought data in the data set, and the duration from the first time period to the second time period. Wherein, the drought data and the associated data correspond to the same drought event, the first time period is the time period when the drought event corresponding to the drought data starts, the second time period is the collection time period of the drought data, and the collection time period of the associated data is before the second time period. It can be understood that more drought events may occur during the statistical period. For the same target area, different drought events occur at different times, and the occurrence times of different drought events are not continuous. In other words, for two adjacent drought events, there is at least a preset number of unit time periods between the end time period of the previous drought event and the start time period of the next drought event. Taking the unit time period as one month as an example, the preset number can be 2. Exemplarily and not restrictively, after obtaining the above-mentioned drought data set with timestamps, each drought data can be divided into multiple drought event groups according to the relationship between the collection time periods of each drought data. It can be understood that the collection time period of drought data is the time period for collecting the original data used to calculate the drought data, that is, the corresponding time period of the drought data. Specifically, every two adjacent drought data in the drought data set can be traversed, and if the number of unit time periods between the collection time periods of the two drought data is greater than or equal to the preset number, the two drought data are divided into different drought event groups. For example, the unit time period is one month, the preset number is 2, and the collection time period (timestamp) of each drought data can be represented by a sequence number. For example, the start time period of the statistical period (the first month) can be represented by "1", the second month can be represented by "2", and the nth month can be represented by "n". Exemplarily, if the time period sequence numbers corresponding to the two adjacent drought data A and B in the drought data set are "10" and "15" respectively, the number of unit time periods between the collection time periods of the two drought data is 5, which is greater than the preset number 2, so the two drought data can be divided into different drought event groups. For example, drought data A can be divided into drought event group A, and drought data B can be divided into drought event group B. Exemplarily, the collection time period of the previous drought data can also be used as the end time period of the corresponding drought event, and the collection time period of the subsequent drought data can be used as the start time period of the corresponding drought event. For example, the period corresponding to drought data A is taken as the end period of drought event A, and the period corresponding to drought data B is taken as the start period of drought event B. It can be understood that for a complete drought event, the drought event usually corresponds to a start period and an end period, while for an incomplete drought event, it usually only corresponds to a start period. Therefore, if the last drought event that occurred during the statistical period is still developing, the last drought event only corresponds to the start period. Figure 2, three drought events occurred in the target area at different times, of which Drought Event 1 and Drought Event 2 were complete drought events, both of which had a start period (the period of drought data collection represented by the light gray dots in the figure) and an end period (the period of drought data collection represented by the dark gray dots in the figure), while Drought Event 3 was a developing drought event, which only had a start period but no end period.
[0081] Exemplarily, for each drought data, other drought data that fall in the same drought event group as the drought data and whose collection period is before the collection period of the drought data can be used as the associated data of the drought data. For example, drought data A is obtained by statistical analysis of the original drought index collected in the 6th month after the start of the statistical period, and the collection period of the drought data can be represented by the serial number "6", then the drought data with the collection period serial number less than 6 can be screened out from the drought event group where the drought data A is located as the associated data of the drought data A. It can be understood that, under normal circumstances, the collection period of each drought data in each drought event group is usually continuous. For example, the unit period can be one month. If a drought event lasts for a total of 10 months from the beginning to the end (or until the end of the statistical period), the drought data of the 10 months can usually be included in the drought event group. In some special examples, affected by some interference factors (such as data missing, data statistical errors, climate abrupt changes, etc.), there may be a lack of drought index in individual periods during the duration of the drought event or the statistical drought data is filtered out because it meets the filtering requirements. Therefore, the collection period of some drought data in the same drought event group may also be discontinuous. For example, a drought event group for a drought event that lasted 10 months may only include drought data for 9 months.
[0082] According to the embodiment of the present application, multiple drought characteristics can be extracted for each drought data. Considering that drought has a cumulative effect in the development process, such as the duration is increasing, the cumulative water shortage degree is also increasing, so the drought characteristics can at least include a feature that can represent the cumulative duration of the drought event and a feature that can represent the cumulative drought intensity of the drought event. Of course, other arbitrary suitable drought characteristics such as the drought peak value representing the drought event, the coverage area of the drought event, etc. can also be extracted for each drought data. It can be understood that for each drought event, the drought characteristics of the drought data of different time periods can accurately characterize the drought characteristics of the drought event at different time periods.
[0083] It should be noted that the schemes for counting the duration or cumulative drought intensity of drought events in the prior art are usually based on complete drought events, and only a fixed value is used to represent the duration or cumulative drought intensity of a drought event. In other words, the prior art only extracts the overall characteristics of a complete drought event, so it is only suitable for post-disaster evaluation. In the drought level identification method of the embodiment of the present application, based on the characteristic that the drought level changes dynamically at different time periods during the occurrence and development of a drought event, the drought characteristics of each drought event at each time period during its occurrence and development are extracted and analyzed. Refer again to Figure 2 , in the prior art, only the feature data of the overall drought characteristics of the first two complete drought events (drought event 1 and drought event 2) can be extracted. In the embodiment of the present application, the drought characteristics of the drought data of 6 time periods during the duration of drought event 1 can be extracted and analyzed, the drought characteristics of the drought data of 9 time periods in drought event 2 can be extracted and analyzed, and the drought characteristics of the drought data of 10 time periods in the unfinished drought event 3 can also be extracted and analyzed, that is, the feature data of the drought characteristics of 25 drought data can be extracted. As a result, not only the number level of the extracted feature data of the drought event is greatly increased, even for the situation where there are fewer drought events during the statistical period, the relationship between different drought characteristics can be scientifically and effectively analyzed, thereby improving the applicability of the scheme. In addition, by analyzing the drought characteristics of each drought event in each time period during the occurrence and development process, the dynamic analysis of any drought event can be accurately realized, thereby providing a data basis for dynamically identifying the drought level of the drought event.
[0084] Specifically, for each drought data, any appropriate feature extraction method can be used to extract each drought feature to obtain feature data corresponding to each drought feature. For example, if the number of drought features is 3, 3 feature data can be extracted for each drought data.
[0085] Exemplarily, the drought data is a drought index. The plurality of characteristic data at least includes the cumulative duration time period and the cumulative severity of the drought event corresponding to the drought data up to the second time period. That is, the cumulative duration time period is characteristic data representing the cumulative duration time of each time period in the process of occurrence and development of the drought event, and the cumulative severity is characteristic data representing the cumulative drought intensity of each time period in the process of occurrence and development of the drought event.
[0086] Step S140 determines a plurality of characteristic data of a plurality of drought characteristics of the drought data, including step S141 and step S142.
[0087] Step S141, calculating the total number of unit time periods from the first time period to the second time period as the cumulative number of continuous time periods.
[0088] It can be understood that, usually, for a target area, if a drought event has not ended, then the drought event has a persistence rule: ① At the i-th time scale, if R d,i If the drought threshold is less than the target area, it is considered that regional drought has occurred in the target area, and the duration of regional drought begins to accumulate; ② When regional drought has not occurred in the i+1 time scale, but regional drought has occurred in the i+2 time scale, considering that drought recovery takes time, the i-th time scale, the i+1 time scale, and the i+2 time scale are considered to be continuous drought events (that is, regional droughts within one time scale are considered to be continuous drought events). Figure 2 Taking the drought data of period T8 in drought event 3 as an example, the first period is period T0, and the second period is period T8. Therefore, the total number of unit periods from period T0 to period T8 is 9, that is, the drought event lasted for a total of 9 periods from the beginning to period T8, so the cumulative duration of the drought data can be recorded as 9.
[0089] Step S142, summing the absolute values of the drought data and all associated data of the drought data to determine the cumulative severity. In this example, there is no drought index with opposite signs in the drought data set. For example, the drought data is the value of the statistical drought index SPEI, with Figure 2 Taking the drought data of drought event 3 as an example, the drought data of each period of drought event 3 can be represented as R d,j (R d,j ≤-0.5), j represents the cumulative duration of drought event 3 in the corresponding period (for example, the drought data in period T8 can be expressed as R d,9 ), the cumulative severity S of each drought data can be calculated using the following formula:
[0090]
[0091] Taking the calculation of the cumulative severity S of drought data in period T8 as an example, the cumulative duration number D=9 can be substituted into the above formula, and the absolute values of the 9 drought data between periods T0 to T8 can be summed to obtain a value representing the cumulative severity of the drought event in this period.
[0092] The above scheme for calculating the characteristic data of the cumulative duration and cumulative drought intensity of drought events in each period during their occurrence and development has simpler calculation logic and less calculation amount, thus having higher recognition efficiency and facilitating real-time diagnosis of drought events in the target area.
[0093] Of course, other suitable calculation methods can also be used to obtain the characteristic data representing the cumulative duration of each period of the drought event during the occurrence and development process and the characteristic data representing the cumulative drought intensity of each period of the drought event during the occurrence and development process. In one example, for the case where the signs of the drought data in the drought data set are different, various suitable data processing methods can also be used to process the various drought data in the drought data set first, so that the values of the processed drought data are all non-negative numbers, and then the sum of the numerical values of these drought data is calculated, and then the calculated sum can be used as the characteristic data characterizing the cumulative drought intensity. For example, the value of each drought data can be subtracted from the preset maximum value so that the values of each drought data after processing are all positive numbers. In another example, for the case where the aforementioned drought data threshold is a negative value, the various drought data in the drought data set are also negative values, and the drought data threshold can also be used to subtract the drought data (for example, using the SPEI value -0.5 to subtract the actual SPEI value of each month in the drought event), so that the drought data at the beginning of each drought event can be converted to 0 as the drought intensity value of the period, and the drought intensity value of each period during the continuous process of the drought event can be guaranteed to be not less than 0.
[0094] Step S160, analyzing multiple characteristic data sequences of the target area, and determining the dependency relationship between multiple drought characteristics according to the analysis results. The multiple characteristic data sequences are generated by each characteristic data of each drought data in the data set, and each characteristic data in each characteristic data sequence corresponds to the same drought characteristic.
[0095] As described above, the multiple characteristic data include the cumulative duration number (for the sake of convenience, it can be called the first characteristic data) and the cumulative severity (for the sake of convenience, it can be called the second characteristic data) as an example. Figure 2 As mentioned above, for the three drought events shown in the figure, a drought data set consisting of 25 drought data can be obtained. Thus, 2 feature data can be obtained for each drought data, that is, 25 first feature data and 25 second feature data can be obtained for the data in the entire drought data set. Thus, the 25 first feature data can be used to generate a first feature data sequence in the order of the drought data collection period, and the 25 second feature data can be used to generate a second feature data sequence in the order of the drought data collection period. By analogy, for other feature data, corresponding feature data sequences can also be generated.
[0096] After generating the characteristic data sequences corresponding to each characteristic data, a variety of suitable methods can be used to analyze and process these characteristic data sequences, and the dependency relationship between multiple drought characteristics can be determined based on the processing results. Specifically, these characteristic data sequences can be analyzed and studied using methods such as correlation analysis, causal analysis, and Granger causality test to determine the dependency relationship between multiple drought characteristics. The dependency relationship between multiple drought characteristics can be represented by a variety of suitable mathematical models, including but not limited to structural equation models, joint probability distribution models, and the like.
[0097] For example, the dependency between multiple drought characteristics can be represented by a joint probability distribution model. Taking into account the above multidimensional drought characteristics, a joint probability model can be constructed using the Copula function to reflect the extreme nature of regional drought events through probability. The assumption is that the smaller the joint probability, the more extreme the corresponding drought event, the higher the level, and the higher the risk of damage that may be caused.
[0098] It can be understood that the Copula function can flexibly simulate the dependency structure between random variables according to the marginal distribution of random variables, has the advantage of connecting the distribution of multi-dimensional variables, and has fewer restrictions on the marginal distribution functions obeyed by multi-variables. Therefore, the Copula function can be used to calculate the joint probability between multi-dimensional drought characteristics. According to Sklar's theorem, assuming X1, X2, ..., X m are multiple continuous random variables, whose marginal distribution functions are F1, F2, ..., F m , if F(X1, X2, ..., X m ) is the joint distribution function of d-dimensional random variables, then there exists a Copula function C such that:
[0099] F(x1, x2, ..., x m )=C(F1(x1), F2(x2),..., Fm(x m ))=C(u1,u2,...,u m )
[0100] Among them, θ is a parameter to be determined.
[0101] The example of multiple characteristic data including the cumulative duration and cumulative severity of drought events corresponding to the drought data up to the second period is used for explanation. Among them, the characteristic data sequence of the cumulative duration can be called Dc sequence, c represents the collection period of drought data, and the characteristic data sequence of the cumulative severity can be called Sc sequence. The following drought characteristic joint probability formula can be used to represent the two-dimensional joint probability P that the cumulative duration and cumulative severity of drought events in the target area are both greater than or equal to a certain value:
[0102] P(Dc≥d∩Sc≥s)=1-F D (d)-F S (s)+C DS (F D (d), F S (s))
[0103] In the formula, (F D (d) means D c The sequence obeys the one-dimensional distribution function, F S (s) means S c The sequence obeys the one-dimensional distribution function, C DS Indicates D c Sequence and S c The optimal Copula function that the sequence obeys. The smaller P is, the smaller the probability of the corresponding drought event occurring in the past is, the more extreme it is, and the higher the level is.
[0104] By way of example and not limitation, step S160 determines the dependency relationship between multiple drought characteristics, including:
[0105] Step S161, for each feature data sequence, fit multiple initial marginal distribution models about the feature data sequence. Step S162, for each feature data sequence, select one marginal distribution model of the feature data sequence from multiple initial marginal distribution models according to the first screening method.
[0106] Exemplarily, in step S161, each feature data sequence may be first statistically analyzed, and the initial marginal distribution model for each feature data sequence may be fitted using a plurality of candidate functions of one-dimensional distribution functions. Table 1 is a candidate function information table of marginal distribution functions of one-dimensional variables according to an embodiment of the present application. Specifically, in step S161, six initial marginal distribution models for each feature data sequence may be fitted according to the six candidate functions listed in Table 1. For example, six one-dimensional distribution functions for the Dc sequence and six one-dimensional distribution functions for the Sc sequence may be fitted. In step S162, a screening method such as the Kolmogorov Smimov (KS) method may be used to select one of the six initial marginal models determined for the Dc sequence and the Sc sequence as the marginal model of the sequence. Specifically, taking the Dc sequence as an example, the KS statistic of the six one-dimensional distribution functions fitted for the Dc sequence may be tested, and a one-dimensional distribution function corresponding to the minimum value of the KS statistic after passing the confidence level of 0.05 may be used as the optimal one-dimensional distribution function of the sequence.
[0107] Table 1
[0108]
[0109]
[0110] In the case where the target area includes multiple ones, the optimal one-dimensional distribution function of each characteristic data sequence of the target area can be determined for each target area. Taking the case where the target area includes North China, the Yangtze River Basin and Yunnan Province, and the characteristic data includes the number of duration periods and the cumulative severity as an example, as shown in Table 2 below, the optimal one-dimensional distribution function of each Dc sequence and Sc sequence determined for each target area can be determined. As shown in Table 2, the optimal one-dimensional distribution function of the same characteristic data sequence determined for different target areas may be the same function or different functions.
[0111] Table 2
[0112]
[0113] Step S163, according to the marginal distribution model of each characteristic data sequence, multiple initial joint probability distribution models about multiple characteristic data sequences are fitted. Step S164, according to the second screening method, a joint probability distribution model between multiple drought characteristics is screened out from each of the fitted initial joint probability distribution models. Among them, the first joint probability distribution model represents the probability distribution of multiple drought characteristic data being less than or equal to a specific value at the same time. The specific value can be the value of the drought characteristic data of any drought data in the drought data set.
[0114] It can be understood that the execution logic of step S163 and step S164 is similar to that of step S161 and step S162 described above. The multidimensional variable Copula function performs well in processing nonlinearity and correlation of extreme events, so the initial joint probability distribution model can be represented by the Copula function. Table 3 is a candidate function information table of the multidimensional variable Copula function according to an embodiment of the present application. Take the characteristic data including the cumulative number of duration periods and the cumulative severity as an example. Table 3 is a candidate function information table of the multidimensional variable Copula function. As shown in Table 3, for the Dc sequence and Sc sequence determined for each target area, the initial joint probability distribution model between the two sequences can be fitted according to the three candidate functions listed in Table 3 (Clayton Copula function, Gumbel-Hougaard Copula function, Frank Copula function) (d=2 in the table). That is, three initial joint probability distribution models can be obtained (copula functions after matching the three candidate functions with the two sequences). The parameter estimation of the Copula function adopts the maximum likelihood method. Then, the optimal Copula function can be screened based on the statistical values of the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC) of the empirical probability and theoretical probability of the random variable. The Copula function with the smallest statistical values of AIC and BIC is taken as the optimal Copula function.
[0115] Table 3
[0116]
[0117] In the case where there are multiple target areas, the optimal first joint probability distribution model between multiple drought characteristics of the drought data set of each target area can be obtained according to the above method. For example, the target areas include North China, the Yangtze River Basin and Yunnan Province. Table 4 below is a list of the optimal copula functions between the Dc sequence and the Sc sequence of each target area screened by the above screening method, as the first joint probability distribution model between the respective drought characteristics.
[0118] Table 4
[0119]
[0120] Step S165, generating a second joint probability distribution model between multiple drought characteristics based on the marginal distribution model of each characteristic data sequence and the selected joint probability distribution model. The second joint probability distribution model is used to represent the probability distribution of multiple drought characteristic data being greater than or equal to a specific value at the same time. The specific value can be the value of the drought characteristic data of any drought data in the drought data set.
[0121] Exemplarily, taking the characteristic data including the cumulative number of duration periods and the cumulative severity as an example, for each target area, the optimal marginal distribution function of each characteristic data sequence selected in Table 2 and the optimal two-dimensional copula function between the two characteristic data sequences listed in Table 4 can be substituted into the above-mentioned drought characteristic joint probability formula to obtain the second joint probability distribution function between the two drought characteristics, which serves as the second joint probability distribution model between the two drought characteristics.
[0122] It can be understood that the joint probability distribution model can comprehensively consider the information between multiple drought characteristics, rather than processing each characteristic separately, which helps to integrate various information sources, make full use of multidimensional data, reduce information loss, and improve the comprehensiveness and accuracy of analysis, so that the dependencies between multiple drought characteristics can be more comprehensively and accurately characterized, so that drought monitoring, prediction and response can be more intuitive and accurate. In addition, the above-mentioned scheme of screening and determining the marginal distribution model of each characteristic data sequence and the joint probability distribution model between multiple drought characteristics can significantly improve the accuracy of the fitted joint probability distribution model, thereby further improving the accuracy of drought monitoring and drought event level identification.
[0123] Step S180, determining multiple thresholds according to the dependency relationship between the multiple drought characteristics, wherein the multiple thresholds are used to determine the drought level of the drought event occurring in the target area during the development process.
[0124] According to the embodiment of the present application, a variety of suitable methods can be used to further determine the drought level during the development process for dynamically identifying the drought event occurring in the target area according to the dependency determined in step S160. The drought level that a drought event may be in during the development process can be multiple, and the type of drought level can be reasonably set according to actual needs. The following takes the types of drought levels including light drought level, moderate drought level, severe drought level and extreme drought level as an example to expand and explain various implementation methods of this step.
[0125] In one example, as mentioned above, for each target area, a joint probability distribution model between multiple drought characteristics can be fitted for the target area first. The joint probability distribution model represents the joint probability distribution of multiple drought characteristic data that are simultaneously greater than or equal to a certain value. According to the "law of large numbers" in steady-state statistical properties, the probability of extreme events occurring in nature is usually small, while the probability of general events occurring is large. In this step, the range of values of the joint probability values of the drought data corresponding to each period represented by the joint probability distribution model can be classified according to this law and the joint probability distribution represented by the joint probability distribution model, so as to determine the joint probability threshold for dividing the drought level of the drought event in the development process. Exemplarily and non-restrictively, the proportion of drought indices corresponding to different drought levels in the meteorological index library can be determined according to the existing meteorological drought level classification standard, and then the joint probability threshold for dividing the drought level belonging to the drought event in the development process can be determined according to the proportion and the joint probability distribution model. After determining the joint probability threshold, the thresholds of each drought characteristic data used to divide the drought level can also be further determined according to the joint probability threshold.
[0126] In another example, the functional relationship formula between each characteristic data and the variable (for example, called the extreme degree variable) that characterizes the extreme degree of the drought event can also be fitted according to the determined dependency relationship. Further, the extreme degree threshold of the drought event for dynamically diagnosing the drought level of the drought event during the occurrence and development process is set according to the functional relationship formula. In this way, at any time during the occurrence and development process of the drought event in the target area, the corresponding actual drought characteristic data can be extracted according to the acquired drought data, and the actual drought characteristic data can be substituted into the above-mentioned functional relationship formula to determine the value of the actual extreme degree variable. Finally, the drought level of the drought event in the current period can be quickly determined according to the determined extreme degree threshold and the value of the actual extreme degree variable. Exemplarily, the drought characteristic data threshold corresponding to the extreme degree threshold can also be determined. Thus, after extracting the drought characteristic data of the drought data in a certain period, the drought characteristic data can be compared with the corresponding drought characteristic data threshold to quickly and accurately determine the drought level of the drought event in the current period.
[0127] It can be understood that determining the drought level of a drought event is actually to quantitatively determine the extreme degree of the drought event, so determining multiple thresholds is the process of quantifying the extreme determination standard of drought. As mentioned above, since the prior art usually analyzes the drought event as a whole that has ended, the scheme of quantifying the drought event level determination standard in the prior art is usually only used to realize the determination of the drought level of the drought event that has ended, which belongs to post-disaster evaluation to a certain extent, and cannot effectively evaluate or warn the level of the drought event in the process of occurrence and development. In most cases, the determination accuracy of the drought level in the process of drought event occurrence and development is low, especially when the data volume of the existing drought event is limited, the quantified drought level determination standard or the determined determination threshold is less accurate. According to the above-mentioned regional drought level identification method of the embodiment of the present application, based on the characteristics of the dynamic change of drought level in different periods during the occurrence and development of drought events, on the basis of collecting drought data of different drought events in different periods during the occurrence and development, according to the cumulative duration and accumulated drought data of each drought event from the beginning to each period during the development, the multiple characteristic data of multiple drought characteristics of each drought event at the end of each period are accurately extracted. And by analyzing the generated multiple feature data sequences, the dependencies between multiple drought features in the target area are accurately determined. Finally, according to the determined dependencies, multiple thresholds that can dynamically identify the drought level of drought events during their occurrence and development are quickly and accurately determined. This scheme fully considers the characteristics of the dynamic changes in the extreme nature of drought events during the formation process, extracts and analyzes feature data from the perspective of drought data in each period of each drought event during its occurrence and development, and provides a more scientific and complete data basis for the dynamic identification of the drought level of drought events. It does not rely on complete drought events, making drought level identification more timely, more accurate, and more widely applicable. In addition, the correlation between the drought feature data extracted for each drought data and the severity of drought in this scheme is high, which can ensure that the determined threshold is more accurate. Moreover, the scheme of determining the threshold for determining the drought level based on the dependency between each drought feature can more comprehensively understand the complexity of drought, effectively reduce the risk of misjudgment, and further improve the accuracy of the determined threshold. This program can not only conduct quantitative analysis on the levels of regional drought events that have already occurred, but also dynamically determine the severity of regional drought events that are currently occurring, providing technical support for the effective implementation of drought relief measures, and has certain practical significance and social value.
[0128] Exemplarily, the threshold comprises a joint probability threshold. Step S160 determines the dependency relationship between a plurality of drought characteristics, including step S161.
[0129] Step S161, fitting a joint probability distribution model between multiple drought characteristics. In this step, a variety of suitable methods can be used to fit the joint probability distribution model between multiple drought characteristics. Exemplarily and not restrictively, the method of steps S161 to S164 above can be used to fit the joint probability distribution model.
[0130] Step S180 determines multiple thresholds, including step S181. Step S181, determines multiple joint probability thresholds according to the probability distribution of drought indices corresponding to different drought levels in the joint probability distribution model and the pre-fitted drought index probability distribution model. The drought index probability distribution model is used to represent the probability distribution of each drought index in the drought index library, and each drought index includes a first drought index corresponding to a drought event and a second drought index corresponding to a non-drought event.
[0131] According to an embodiment of the present application, the probability distribution model of the drought index can be a probability distribution model of any form that is fitted according to the distribution of the drought index in the existing drought index library. The drought index in the drought index library can be the drought index of any referenced area in different continuous periods during a longer observation period. Optionally, the drought index library is a monthly precipitation and temperature (maximum temperature, minimum temperature, average temperature) sequence of each station in the target area from January 1960 to January 2024, which is counted by the National Meteorological Science Data Center. For example, the data sequence generated by these drought indices can be statistically analyzed according to the method mentioned in the "Meteorological Drought Level", and the probability distribution model of the drought index can be fitted. For example, a probability distribution model representing the difference sequence of precipitation and potential evapotranspiration each month can be fitted. And according to the appropriate drought level judgment standard, the proportion of drought index corresponding to each drought level in the drought index library can be determined. Alternatively, according to the drought level classification method of various drought indices in the "Meteorological Drought Level", it can be approximately deduced that the probability distribution of drought index in different periods in the same region can be converted into a standard normal distribution, and the drought index interval is usually divided into five categories: no drought, light drought, moderate drought, severe drought and extreme drought according to the commonly used 0.5 standard deviation, 1 standard deviation, and 1.5 standard deviation. Thus, the cumulative probability of drought index corresponding to different drought levels within the entire drought index value interval can be determined, which is approximately equal to the coverage rate of drought index of each drought level in the entire drought index interval. And the calculated coverage rate can be converted to obtain the proportion of drought index of each drought level in the first drought index interval corresponding to all drought levels. Then, according to the determined proportion and the joint probability distribution model, multiple joint probability thresholds for dynamically evaluating the drought level of drought events can be determined. As mentioned above, the joint probability distribution model can represent the joint probability distribution of multiple drought characteristic data that are simultaneously greater than or equal to a certain value. Therefore, in this step, a variety of suitable methods can be used to match the proportion (coverage) of the determined drought index corresponding to different drought levels with the value range of the joint probability value represented in the joint probability distribution model, so as to determine multiple joint probability thresholds.
[0132] In one example, the drought index library includes a first drought index corresponding to drought events and a second drought index corresponding to non-drought events. Step S181 determines a plurality of joint probability thresholds, including the following steps S181a, S181b, S181c and S181d.
[0133] Step S181a, according to the drought index probability distribution model and the drought index thresholds corresponding to different drought levels, determine the proportion of the first drought index corresponding to each drought level in the first drought index group, wherein the first drought index group is composed of each first drought index in the drought index library. For example, according to the thresholds for different drought levels divided by the commonly used standard precipitation index (SPI), standard precipitation evapotranspiration index (SPEI), meteorological drought comprehensive index (MCI), etc. and the corresponding probability density values of these thresholds in the standard normal distribution, it can be approximately obtained that the total proportion of drought, flood and normal events is 100% (corresponding to the cumulative probability of the entire standard normal distribution), wherein the proportion of light drought is considered to be about 15%, the proportion of moderate drought is about 9.2%, the proportion of severe drought is about 4.4%, and the proportion of extreme drought is about 2.3%. Then, taking the drought events as a whole, the proportion of light drought is about 15 / (15+9.2+4.4+2.3)=48.5%, the proportion of moderate drought is about 29.8%, the proportion of severe drought is about 14.2%, and the proportion of extreme drought is about 7.5%. For example, the above ratio can be calculated based on the classification rules in the Standardized Precipitation Evaporation Index Drought Classification Table in the Meteorological Drought Classification.
[0134] Step S181b, according to the joint probability distribution model, determine the joint probability value between multiple feature data of each drought data in the data set. Specifically, the feature data of each drought data can be substituted into the determined joint probability distribution model to obtain the joint probability value corresponding to each drought data. For example, Figure 2 The 25 drought data shown in obtain corresponding 25 joint probability values.
[0135] Step S181c, arranging the determined joint probability values in ascending order to determine the order percentiles of the joint probability values. Figure 2 The 25 joint probability values of the 25 drought data in are sorted in order to obtain the sequential percentile of each joint probability value.
[0136] Step S181d, according to the proportions and the order percentiles of each joint probability value, multiple joint probability thresholds are determined. As mentioned above, the types of drought levels can be set according to actual needs. The more types of drought levels are set, the more joint probability thresholds are determined. Exemplarily and non-restrictively, in this step, according to the proportions corresponding to different drought levels and the order percentiles of each joint probability value, the joint probability values corresponding to the order percentiles of these proportions can be selected from each joint probability value as the joint probability threshold for determining the drought level during the development of the drought event. Of course, in some examples, these proportions can also be analyzed and processed according to a suitable analysis method, and then the joint probability threshold is accurately determined according to the proportions corresponding to different drought levels and the order percentiles of each joint probability value.
[0137] The above scheme can quickly and accurately determine the joint probability threshold that can dynamically diagnose the drought level, facilitate dynamic and real-time grasp of the regional drought evolution process, and provide support for drought early warning.
[0138] Exemplarily, the types of drought levels include mild drought, moderate drought, severe drought and extreme drought. The determined ratios include a first ratio, a second ratio, a third ratio and a fourth ratio corresponding to the mild drought, moderate drought, severe drought and extreme drought, respectively. The multiple joint probability thresholds include a first probability threshold, a second probability threshold and a third probability threshold. It can be understood that the sum of the first ratio, the second ratio, the third ratio and the fourth ratio is equal to 1.
[0139] Step S181d determines multiple joint probability thresholds, including the following steps:
[0140] Step S181d.1, determining a third probability threshold according to the joint probability value of the order percentile being equal to the fourth ratio, the third probability threshold representing the dividing threshold between the extreme drought level and the severe drought level;
[0141] Step S181d.2, determining a second probability threshold value according to the joint probability value of the order percentile being equal to the fifth ratio, wherein the fifth ratio is equal to the sum of the third ratio and the fourth ratio, and the second probability threshold value represents the dividing threshold value between the severe drought level and the moderate drought level;
[0142] Step S181d.3, determine the first probability threshold according to the joint probability value of the order percentile equal to the sixth ratio, where the sixth ratio is equal to the sum of the second ratio, the third ratio and the fourth ratio, and the first probability threshold represents the dividing threshold between moderate drought and light drought.
[0143] For example, as shown in Table 5 below, taking drought events as a whole, the proportion of light drought determined can be 48.5% (first proportion), the proportion of moderate drought is about 29.8% (second proportion), the proportion of severe drought is about 14.2% (third proportion), and the proportion of extreme drought is about 7.5% (fourth proportion). The joint probability values corresponding to the various drought data can be arranged in positive order from small to large, and the joint probability values P with order percentiles of 7.5%, 21.7% (fifth proportion), and 51.5% (sixth proportion) are respectively searched as the joint probability thresholds for determining light drought, moderate drought, severe drought, and extreme drought in the process of regional drought development.
[0144] Table 5
[0145]
[0146] In the above example where the target area includes North China, the Yangtze River Basin and Yunnan Province, the joint probability threshold determination method in the above steps can be used to obtain multiple joint probability thresholds corresponding to different drought levels for each target area. For example, as shown in Table 6 below (P j The first probability threshold, the second probability threshold and the third probability threshold determined for North China are 0.4199, 0.1228 and 0.0344 respectively, and the joint probability intervals corresponding to each drought level can be obtained. It can be seen from the table that the difference in the same drought level threshold in different target areas is relatively small. The joint probabilities corresponding to the light drought level in these three target areas are concentrated between 0.4-0.5, the joint probabilities corresponding to the moderate drought level and below are concentrated between 0.12-0.15, and the joint probabilities corresponding to the boundary between the severe drought level and the extreme drought level are concentrated between 0.02-0.035.
[0147] Table 6
[0148]
[0149] It can be understood that regional drought events have the characteristics of spreading and developing from point to surface and from light to heavy in space. In the above scheme, on the one hand, the joint probability value corresponding to each drought data is obtained according to the probability distribution model between the fitted multi-drought characteristics, and the order percentile of each joint probability value is determined by sorting these joint probability values. On the other hand, the proportion of drought index of different drought levels in the existing drought index library is transformed. Then, according to the converted proportion, the joint probability value with the same order percentile as these proportions is screened out from each joint probability value as the joint probability threshold for determining the extreme degree of drought events in the dynamic development process. The joint probability threshold determined by this scheme can not only accurately distinguish different drought levels and significantly improve the recognition accuracy of regional drought, but also the scheme has a small amount of calculation, so the recognition efficiency is also higher, which is helpful to realize the real-time and accurate drought monitoring and early warning management of drought events.
[0150] Exemplarily, the threshold value also includes multiple drought characteristic threshold values, and the multiple drought characteristic threshold values correspond to multiple drought characteristics one by one. Step S180 of determining multiple threshold values also includes: determining multiple drought characteristic data threshold values corresponding to each joint probability threshold value according to multiple joint probability threshold values and a joint probability model.
[0151] The following is an example of drought characteristic data including the cumulative duration and cumulative severity of drought events corresponding to the drought data up to the second period. As shown in Table 6, according to the joint probability threshold obtained by the above method, the difference in the joint probability threshold corresponding to the same drought level in different target areas is small. Considering the high correlation between the cumulative duration and the cumulative severity, the same frequency method can be further used to reversely solve the joint probability threshold corresponding to the drought level of each region, and then obtain the drought characteristic threshold corresponding to the drought level of different regions in each study area. The results are shown in Table 7.
[0152] Table 7
[0153]
[0154]
[0155] As shown in Table 7, the drought characteristic thresholds corresponding to the light drought level in the three target areas are not much different, and the drought characteristic threshold in the Yangtze River Basin is slightly lower. For the drought characteristic thresholds corresponding to the moderate drought level and above, the drought characteristic thresholds in Yunnan Province are generally higher than those in the other two target areas, especially for the drought characteristic thresholds corresponding to the extreme drought level. In Yunnan Province, drought events are identified as extreme drought only when the cumulative duration of drought events is more than 16 months and the cumulative severity is above 17; while for North China or the Yangtze River Basin, drought events are identified as extreme drought only when the cumulative duration of drought events reaches 8 months and the cumulative severity is higher than 6.1 or 8.2.
[0156] Exemplarily, the method 100 further includes step S191.
[0157] Step S191, determining the drought level of the drought event corresponding to the drought data in the specific period of time according to at least a plurality of thresholds and drought data in the specific period of time. The specific period of time is any period of time in each period of time or any period of time in the development process of the drought event occurring in the target area after the statistical period.
[0158] Reference again Figure 2 , Figure 2 For example, it is a part of the drought events counted in the Yangtze River Basin during the statistical period. In this step, a variety of suitable methods can be used to determine the drought level of the corresponding drought event in the period according to the multiple thresholds determined for determining the drought level of the drought event in the development process and the drought data of any period shown in the figure. It is also possible to determine the drought level of drought event 3 at time T10 after time T9, for example, at time T10, according to the drought data of the target area at time T10 and the multiple thresholds determined.
[0159] Exemplarily, step S191 determines the drought level of the drought event corresponding to the drought data in a specific period of time, including the following steps:
[0160] Step S191a, matching each characteristic data of the drought data in a specific period with the joint probability model to determine the joint probability value corresponding to the drought data in the period as the specific joint probability.
[0161] Step S191b: if the specific joint probability is greater than the first probability threshold, it is determined that the drought level of the first drought event in the period is a light drought level, wherein the first drought event is a drought event in the drought data.
[0162] Step S191c: if the specific joint probability is greater than the second probability threshold and less than or equal to the first probability threshold, it is determined that the drought level of the first drought event in this period is moderate drought level.
[0163] Step S191d: if the specific joint probability is greater than the third probability threshold and less than or equal to the second probability threshold, it is determined that the drought level of the first drought event in this period is a severe drought level.
[0164] Step S191e: if the specific joint probability is less than or equal to the third probability threshold, it is determined that the drought level of the first drought event in this period is an extreme drought level.
[0165] Reference again Figure 2 , taking the target area as North China and the specific period as any period Ta from T0 to T9 as an example, the cumulative duration of drought data at time Ta is D a and cumulative severity S a Substituting into the fitted joint probability distribution function, we get the cumulative duration period greater than or equal to D a And the cumulative severity is greater than or equal to S a The two-dimensional joint probability P j , as a specific joint probability. As shown in Table 6, the first probability threshold, the second probability threshold, and the third probability threshold determined for North China are 0.4199, 0.1228, and 0.0344, respectively. j Compared with these three probability thresholds, if the specific joint probability P j >0.4199, then the drought level of drought event 3 in this period is determined to be light drought level; if 0.1228 <P j ≤0.4199, the drought level of drought event 3 in this period is determined to be moderate drought level; if 0.0344 <P j ≤0.1228, then the drought level of drought event 3 during this period is determined to be severe drought level; if P j≤0.0344, it is determined that the drought level of drought event 3 during this period is extremely drought level.
[0166] According to this method, the drought levels of each drought event in the three target areas during the statistical period can be further determined. Figure 2 The drought level of moderate drought event 1 at the end period was light drought level, the drought level of drought event 2 at the end period was determined to be moderate drought level, and the drought level of drought event 3 in period T7 was moderate drought level.
[0167] Figure 3a A schematic diagram showing the dynamic change process of drought levels of various drought events in different target areas during a statistical period at different time periods according to an embodiment of the present application is shown. Figure 3a In the figure, data points of the same color represent the joint probability value of the cumulative number of drought periods and the cumulative drought severity of a drought event in different time periods. It should be noted that due to the limitations of the color display mode, there are cases where data points representing different drought events are displayed in the same color. Figure 3a It can be seen that with the continuous accumulation of drought severity and cumulative duration, the severity of regional drought is increasing.
[0168] Figure 3b A schematic diagram showing the proportion of different drought levels in different target areas during the development of drought events during the statistical period according to an embodiment of the present application is shown. Figure 3b As shown, the target areas basically present a distribution state of gradually decreasing from regional mild drought to regional severe drought, which is consistent with the perception that extreme events occur less frequently and general events occur more frequently in nature.
[0169] In the above scheme, according to the determined joint probability threshold, a dynamic diagnosis of the drought level of a drought event occurring in the target area at any time period within or outside the statistical period during the development process can be achieved quickly and accurately.
[0170] Exemplarily, the statistical period is from the statistical start period to the current period. The specific period is the current period. The method 100 further includes: if it is determined that the drought level of the drought event in the current period is greater than the preset drought level, outputting prompt information indicating drought warning.
[0171] The current period is Figure 2Taking the T9 period shown in as an example, the method of steps S191a to S191e above can be used to quickly diagnose the drought level of drought event 3 in the T9 period. The preset drought level can be arbitrarily set according to actual needs. Taking the preset drought level as a medium drought level as an example, in the above example of determining that the drought level of drought event 3 in the T7 period is a medium drought level, a prompt message such as "The drought event in this area has reached the medium drought level, please execute the emergency response plan!" can be output in the T7 period. Thus, the corresponding organizations can be quickly mobilized to implement disaster relief, land reclamation and other measures, which can have social significance such as reducing drought losses, reducing natural disaster risks, and enhancing social stability to a certain extent.
[0172] Exemplarily, after determining a plurality of thresholds in step S180 , the method 100 further includes step S192 .
[0173] Step S192, in each time period of at least one monitoring period after the statistical period, at least according to multiple thresholds and the drought data of the period, determine the drought level corresponding to the drought event corresponding to the drought data in the period. The monitoring period can also be set according to actual needs, such as half a year, 1 year or 3 years. Taking the statistical period from January 1960 to December 2020 and the monitoring period of 1 year as an example, in this step, the drought characteristic threshold value shown in Table 7 can be determined according to the drought data of the target area during the statistical period. In the first monitoring period after the statistical period, that is, each month from January 2021 to December 2021, drought data such as the SPEI value can be determined according to the drought index of the month to determine whether there is a drought event in the current area. If so, the corresponding drought characteristic data is calculated, and the drought characteristic data is compared with the corresponding drought characteristic threshold value in Table 7 to determine the drought level of the drought event of the month. And when the drought level is greater than the preset drought level, the warning prompt information can be output.
[0174] In this solution, during at least a part of the monitoring period after the statistical period, the drought level of the drought event can be accurately and real-time determined directly based on the determined threshold value, thereby improving the efficiency of drought monitoring.
[0175] Exemplarily, the at least one monitoring period is the first monitoring period. The method 100 further includes step S193, step S194 and step S195.
[0176] Step S193, at the end of each monitoring period, adding the drought data of each time period in the monitoring period to the existing drought data set to obtain an updated drought data set.
[0177] Step S194, based on each drought characteristic data of each drought data in the updated drought data set, the dependency relationship between multiple drought characteristics is updated, and multiple thresholds are updated to obtain multiple updated thresholds.
[0178] Step S195, for each time period in the next monitoring period, determine the drought level corresponding to the drought event corresponding to the drought data in the time period based on at least the updated multiple thresholds and the drought data in the time period.
[0179] In the example above, where the statistical period is from January 1960 to December 2020 and the monitoring period is 1 year, the 12 monthly drought data of the previous year can be added to the drought dataset at the beginning of each year after 2021, and the joint probability distribution model between multiple drought characteristics of the target area can be refitted according to the updated drought dataset using the method in the above steps, and multiple thresholds can be updated according to the updated joint probability distribution model. Then, in the next monitoring period, the drought level of drought events in each period of the monitoring period can be identified in real time and accurately according to the updated threshold.
[0180] In this scheme, the drought threshold can also be updated at a certain frequency to further improve the accuracy of drought level identification, achieving a balance between identification efficiency and identification accuracy.
[0181] Figure 4 FIG. 4 is a flow chart showing a drought level identification method according to another embodiment of the present application. Figure 4As shown, firstly, the drought data of drought events occurring in the target area during the statistical period in each period are obtained to generate a drought data set. Then, drought characteristic data corresponding to multiple drought characteristics are extracted for each drought data as characteristic data of each period in the development process of the drought event, and a characteristic data sequence is generated. For example, the cumulative duration period number and the cumulative drought severity of each drought event up to each period in the development are extracted to generate a cumulative duration period number sequence and a cumulative drought severity sequence. Then, the two characteristic data sequences can be statistically analyzed, and a joint probability distribution model (second joint probability distribution model) between the cumulative duration period number sequence and the cumulative drought severity sequence can be fitted. Furthermore, multiple joint probability thresholds can be determined according to the probability distribution of drought indices corresponding to different drought levels in the joint probability distribution model and the pre-fitted drought index probability distribution model. Specifically, the proportion of the first drought index corresponding to each drought level in the first drought index group can be determined according to the drought index probability distribution model and the drought index threshold corresponding to different drought levels. And the joint probability value between multiple characteristic data of each drought data in the data set can be determined according to the joint probability distribution model. Then, the determined joint probability values can be arranged in ascending order to determine the order percentile of each joint probability value. After that, multiple joint probability thresholds can be determined according to the ratio of drought indexes of different drought levels and the order percentile of each joint probability value. Then, the same frequency method can be used to reversely solve the cumulative duration threshold and the cumulative drought severity threshold corresponding to each drought level according to the determined joint probability threshold. Furthermore, in the monitoring period after the statistical period, the drought level of the drought event in each period of development during the current monitoring period can be dynamically identified in real time according to the determined cumulative duration threshold and the cumulative drought severity threshold and the drought data of the drought event in each period of development during the monitoring period. If it is determined that the drought level is greater than the preset drought level, an early warning prompt message is output. If the current monitoring period ends, the drought data in the monitoring period is updated to the drought data set, and the joint probability distribution model and threshold are updated according to the updated drought data set. In the next monitoring period, the drought level of the drought event in the development process is dynamically identified according to the updated threshold.
[0182] According to another aspect of the present application, a regional drought level identification system is also provided. Figure 5 A schematic block diagram of a regional drought level identification system 500 according to an embodiment of the present application is shown. As shown in the figure, the system 500 includes:
[0183] An acquisition module 510 is used to acquire a drought data set of a target area, wherein the drought data set includes drought data of multiple drought events occurring in the area during a statistical period in various time periods;
[0184] A first determination module 520 is used to determine, for each drought data, a plurality of characteristic data of a plurality of drought characteristics of the drought data according to at least the drought data, the associated data of the drought data in the data set, and a duration from a first time period to a second time period, wherein the drought data and the associated data correspond to the same drought event, the first time period is a time period when the drought event corresponding to the drought data starts, the second time period is a time period for collecting the drought data and the time period for collecting the associated data is before the second time period, wherein the plurality of characteristic data correspond to the plurality of drought characteristics one by one;
[0185] An analysis module 530 is used to analyze multiple feature data sequences of the target area and determine the dependency relationship between multiple drought features according to the analysis results, wherein the multiple feature data sequences are generated by each feature data of each drought data in the data set, and each feature data in each feature data sequence corresponds to the same drought feature; and
[0186] The second determination module 540 is used to determine multiple thresholds according to the dependency relationship between multiple drought characteristics, wherein the multiple thresholds are used to determine the drought level of the drought event occurring in the target area during the development process, and different thresholds correspond to different drought levels.
[0187] According to another aspect of the present application, an electronic device is also provided. Figure 6 A schematic block diagram of an electronic device 600 according to an embodiment of the present application is shown. As shown in the figure, the electronic device 600 includes a processor 610 and a memory 620, in which computer program instructions are stored, and the computer program instructions are used to execute the above-mentioned regional drought level identification method 100 when the processor 610 is running.
[0188] In addition, according to another aspect of the present application, a storage medium is also provided. Program instructions are stored on the storage medium. When the program instructions are executed by a computer or a processor, the computer or the processor executes the corresponding steps of the above-mentioned regional drought level identification method 100 of the embodiment of the present application, and is used to implement the corresponding modules in the above-mentioned regional drought level identification system 500 or the corresponding modules in the above-mentioned electronic device 600 according to the embodiment of the present application. The storage medium may include, for example, a memory card of a smart phone, a storage component of a tablet computer, a hard disk of a personal computer, a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a portable read-only memory (CD-ROM), a USB memory, or any combination of the above-mentioned storage media. The computer-readable storage medium may be any combination of one or more computer-readable storage media.
[0189] According to another aspect of the present application, a computer program product is also provided, including computer program instructions, and the computer program instructions are used to execute the above-mentioned regional drought level identification method 100 when running.
[0190] A person skilled in the art can understand the specific implementation and beneficial effects of the above-mentioned regional drought level identification system, electronic device, storage medium and computer program product by reading the above-mentioned detailed description of the regional drought level identification method. For the sake of brevity, they will not be repeated here.
[0191] Although example embodiments have been described herein with reference to the accompanying drawings, it should be understood that the above example embodiments are merely exemplary and are not intended to limit the scope of the present application to this. Those of ordinary skill in the art may make various changes and modifications therein without departing from the scope and spirit of the present application. All these changes and modifications are intended to be included within the scope of the present application as required by the appended claims.
[0192] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.
[0193] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed.
[0194] In the description provided herein, a large number of specific details are described. However, it is understood that the embodiments of the present application can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.
[0195] Similarly, it should be understood that in order to streamline the present application and help understand one or more of the various inventive aspects, in the description of the exemplary embodiments of the present application, the various features of the present application are sometimes grouped together into a single embodiment, figure, or description thereof. However, the method of the present application should not be interpreted as reflecting the following intention: the claimed application requires more features than the features clearly stated in each claim. More specifically, as reflected in the corresponding claims, the inventive point is that the corresponding technical problem can be solved with features less than all the features of a single disclosed embodiment. Therefore, the claims following the specific embodiment are hereby explicitly incorporated into the specific embodiment, wherein each claim itself serves as a separate embodiment of the present application.
[0196] It will be understood by those skilled in the art that, except for mutually exclusive features, all features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed in this specification may be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature that provides the same, equivalent or similar purpose.
[0197] In addition, those skilled in the art will appreciate that, although some embodiments herein include certain features included in other embodiments but not other features, the combination of features of different embodiments is meant to be within the scope of the present application and form different embodiments. For example, in the claims, any one of the claimed embodiments may be used in any combination.
[0198] The various component embodiments of the present application can be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or a digital signal processor (DSP) can be used in practice to implement some or all functions of some modules in the regional drought level identification system according to the embodiment of the present application. The present application can also be implemented as a device program (e.g., a computer program and a computer program product) for executing a part or all of the methods described herein. Such a program implementing the present application can be stored on a computer-readable medium, or can have the form of one or more signals. Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.
[0199] It should be noted that the above embodiments illustrate the present application rather than limit the present application, and that those skilled in the art may design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets should not be constructed as a limitation to the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of multiple such elements. The present application may be implemented by means of hardware including several different elements and by means of a suitably programmed computer. In a unit claim that lists several devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc. does not indicate any order. These words may be interpreted as names.
[0200] The above is only a specific implementation or description of a specific implementation of the present application, and the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. The protection scope of the present application shall be based on the protection scope of the claims.
Claims
1. A method for identifying regional drought levels, characterized in that: include: Acquire a drought data set of a target area, wherein the drought data set includes drought data of multiple drought events occurring in the area during a statistical period in various time periods; For each drought data, determining multiple characteristic data of multiple drought characteristics of the drought data at least according to the drought data, the associated data of the drought data in the data set, and the duration from a first time period to a second time period, wherein the drought data and the associated data correspond to the same drought event, the first time period is the time period when the drought event corresponding to the drought data starts, the second time period is the collection time period of the drought data, and the collection time period of the associated data is before the second time period; Analyzing a plurality of characteristic data sequences of the target area, and determining the dependency relationship between the plurality of drought characteristics according to the analysis result, wherein the plurality of characteristic data sequences are generated by respective characteristic data of respective drought data in the data set, and each characteristic data in each characteristic data sequence corresponds to the same drought characteristic; and According to the dependency relationship between the plurality of drought characteristics, a plurality of thresholds are determined, wherein the plurality of thresholds are used to determine the drought level of the drought event occurring in the target area during the development process, and the thresholds include a joint probability threshold, Wherein, determining the dependency relationship between the plurality of drought characteristics comprises: Fitting a joint probability distribution model between the multiple drought characteristics, wherein the joint probability distribution model is used to describe the distribution of joint probabilities between multiple characteristic data of each drought data in the drought data set; The determining of multiple thresholds includes: A plurality of joint probability thresholds are determined based on the probability distribution of drought indices corresponding to different drought levels in the joint probability distribution model and a pre-fitted drought index probability distribution model, wherein the drought index probability distribution model is used to represent the probability distribution of each drought indices in a drought index library, and each drought indices includes a first drought index corresponding to a drought event and a second drought index corresponding to a non-drought event.
2. The method for identifying regional drought levels according to claim 1, characterized in that: The determining of multiple joint probability thresholds comprises: According to the drought index probability distribution model and drought index thresholds corresponding to different drought levels, determining the proportion of the first drought index corresponding to each drought level in the first drought index group, wherein the first drought index group is composed of the first drought indices in the drought index library; Determining a joint probability value between a plurality of characteristic data of each drought data in the data set according to the joint probability distribution model; Arranging the determined joint probability values in ascending order to determine the ordinal percentile of each joint probability value; The plurality of joint probability thresholds are determined according to the proportion and the ordinal percentile of each joint probability value.
3. The method for identifying regional drought levels according to claim 2, characterized in that: The types of drought levels include light drought, moderate drought, severe drought and extreme drought, the determined coverage includes a first ratio, a second ratio, a third ratio and a fourth ratio corresponding to the light drought, moderate drought, severe drought and extreme drought in sequence, the multiple joint probability thresholds include a first probability threshold, a second probability threshold and a third probability threshold, and the determining of the multiple joint probability thresholds includes: Determine the third probability threshold value according to the joint probability value whose order percentile is equal to the fourth ratio, wherein the third probability threshold value represents the dividing threshold value between the extreme drought level and the severe drought level; determining the second probability threshold according to a joint probability value whose order percentile is equal to a fifth ratio, wherein the fifth ratio is equal to the sum of the third ratio and the fourth ratio, and the second probability threshold represents a dividing threshold between the severe drought level and the moderate drought level; The first probability threshold is determined according to the joint probability value whose order percentile is equal to the sixth ratio, wherein the sixth ratio is equal to the sum of the second ratio, the third ratio and the fourth ratio, and the first probability threshold represents the dividing threshold between the moderate drought level and the mild drought level.
4. The method for identifying regional drought levels according to claim 1, characterized in that: The drought data is a drought index, the multiple characteristic data at least include the cumulative duration and cumulative severity of the drought event corresponding to the drought data up to the second time period, and the multiple characteristic data for determining the multiple drought characteristics of the drought data include: Calculate the total number of unit time periods from the first time period to the second time period as the cumulative number of continuous time periods; The absolute values of the drought data and all associated data of the drought data are summed to determine the cumulative severity.
5. The method for identifying regional drought levels according to any one of claims 1 to 4, characterized in that: The method further comprises: The drought level of the drought event corresponding to the drought data in the specific time period is determined based at least on the multiple thresholds and the drought data in the specific time period, wherein the specific time period is any time period among the various time periods or the specific time period is any time period in the development process of the drought event occurring in the target area after the statistical period.
6. The method for identifying regional drought levels according to claim 5 as cited in claim 3, characterized in that: Determining the drought level of the drought event corresponding to the drought data in the specific period includes: Matching each characteristic data of the drought data in the specific period with the joint probability model to determine a joint probability value corresponding to the drought data in the specific period as a specific joint probability; If the specific joint probability is greater than the first probability threshold, determining that the drought level of the first drought event in the period is a light drought level, wherein the first drought event is the drought event corresponding to the drought data; If the specific joint probability is greater than the second probability threshold and less than or equal to the first probability threshold, determining that the drought level of the first drought event in the period is moderate drought level; If the specific joint probability is greater than the third probability threshold and less than or equal to the second probability threshold, determining that the drought level of the first drought event in the period is a severe drought level; If the specific joint probability is less than or equal to the third probability threshold, it is determined that the drought level of the first drought event in this period is an extreme drought level.
7. The method for identifying regional drought levels according to claim 5, characterized in that: The statistical period is from the statistical start time period to the current time period, the specific time period is the current time period, and the method further includes: If it is determined that the drought level of the drought event in the current period is greater than the preset drought level, prompt information indicating drought warning is output.
8. The method for identifying regional drought levels according to any one of claims 1 to 4, characterized in that: After determining the plurality of thresholds, the method further comprises: In each time period within at least one monitoring period after the statistical period, the drought level of the drought event corresponding to the drought data in the time period is determined at least based on the multiple thresholds and the drought data in the time period.
9. The method for identifying regional drought levels according to claim 8, characterized in that: The at least one monitoring period is a first monitoring period, and the method further comprises: At the end of each monitoring period, the drought data of each period in the monitoring period is added to the existing drought data set to obtain an updated drought data set; Based on each drought characteristic data of each drought data in the updated drought data set, updating the dependency relationship between the multiple drought characteristics, and updating the multiple thresholds to obtain multiple updated thresholds; and For each time period in the next monitoring period, the drought level of the drought event corresponding to the drought data in the time period is determined at least based on the updated multiple thresholds and the drought data in the time period.
10. The method for identifying regional drought levels according to any one of claims 1 to 4, characterized in that: Determining the dependency relationship between the plurality of drought characteristics comprises: For each feature data sequence, Fitting multiple initial marginal distribution models about the feature data sequence; According to the first screening method, a marginal distribution model as the feature data sequence is screened out from the multiple initial marginal distribution models; According to the marginal distribution model of each feature data sequence, a plurality of initial joint probability distribution models about the plurality of feature data sequences are fitted; According to the second screening method, one is screened out from each of the fitted initial joint probability distribution models as a first joint probability distribution model between the multiple drought characteristics, wherein the first joint probability distribution model is used to represent the probability distribution situation that the multiple drought characteristic data are simultaneously less than or equal to a specific value; A second joint probability distribution model between the multiple drought characteristics is generated based on the marginal distribution models of each characteristic data sequence and the screened first joint probability distribution model, wherein the second joint probability distribution model is used to represent the probability distribution situation that the multiple drought characteristic data are simultaneously greater than or equal to a specific value.
11. A regional drought level identification system, characterized in that: include: An acquisition module is used to acquire a drought data set of a target area, wherein the drought data set includes drought data of multiple drought events occurring in the area during a statistical period in various time periods; A first determination module is used to determine, for each drought data, a plurality of characteristic data of a plurality of drought characteristics of the drought data at least according to the drought data, the associated data of the drought data in the data set, and a duration from a first time period to a second time period, wherein the drought data and the associated data correspond to the same drought event, the first time period is a time period when the drought event corresponding to the drought data starts, the second time period is a time period for collecting the drought data, and the time period for collecting the associated data is before the second time period; an analysis module, configured to analyze a plurality of characteristic data sequences of the target area, and determine the dependency relationship between the plurality of drought characteristics according to the analysis result, wherein the plurality of characteristic data sequences are generated by respective characteristic data of respective drought data in the data set, and each characteristic data in each characteristic data sequence corresponds to the same drought characteristic; and The second determination module is used to determine a plurality of thresholds according to the dependency relationship between the plurality of drought characteristics, wherein the plurality of thresholds are used to determine the drought level of the drought event occurring in the target area during the development process, and different thresholds correspond to different drought levels, wherein the thresholds include a joint probability threshold, The analysis module determines the dependency relationship between the plurality of drought characteristics by performing the following operations: fitting a joint probability distribution model between the plurality of drought characteristics, wherein the joint probability distribution model is used to describe the distribution of the joint probability between the plurality of characteristic data of each drought data in the drought data set; The second determination module determines multiple thresholds by performing the following operations: determining multiple joint probability thresholds according to the probability distribution of drought indices corresponding to different drought levels in the joint probability distribution model and a pre-fitted drought index probability distribution model, wherein the drought index probability distribution model is used to represent the probability distribution of each drought indices in a drought index library, and the each drought indices include a first drought index corresponding to a drought event and a second drought index corresponding to a non-drought event.
12. An electronic device comprising a processor and a memory, characterized in that: The memory stores computer program instructions, which are used by the processor to execute the method for identifying regional drought levels according to any one of claims 1 to 10 when the processor executes the instructions.
13. A storage medium having program instructions stored thereon, characterized in that: The program instructions are used to execute the regional drought level identification method according to any one of claims 1 to 10 when running.
14. A computer program product comprising computer program instructions, characterized in that The computer program instructions are used to execute the method for identifying regional drought levels according to any one of claims 1 to 10 when executed.
