Regional plum entering intelligent forecasting method based on circulation adjustment and element judgment

Abstract

The invention relates to the technical field of intelligent forecasting of plum entering, and discloses an intelligent forecasting method for regional plum entering based on circulation adjustment and element judgment, and the method is characterized in that a plurality of key characteristic quantities influencing plum rain entering are determined, and the membership degree of each characteristic quantity is calculated based on historical data; constructing an in-plum prediction credibility index; on the basis of the long sequence historical data set, a support vector machine, a decision tree or a Logistic regression algorithm is adopted to train an in-plum classification prediction model; acquiring circulation factor data of a global forecasting system in real time, extracting key characteristic quantities, calculating membership degrees of the key characteristic quantities, and inputting the key characteristic quantities into a forecasting model to carry out plum entering condition classification judgment; and judging whether a plum rain monitoring standard is met or not by combining ground element actual condition monitoring and forecast data, and comprehensively determining the plum entering date.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent plum rain forecasting, and more specifically, it relates to an intelligent plum rain forecasting method for a region based on circulation adjustment and element judgment. Background Art

[0002] Traditional plum rain forecasting methods are highly subjective, require a large amount of data, have a cumbersome process for forecasters to analyze, consume a lot of energy in forecasting, and the judgment results are somewhat subjective and not accurate enough. In addition, traditional plum rain forecasting judgments lack structural logic, are not distinct enough in levels, do not well integrate real-time monitoring and the forecasting field for judgment, and also lack the perfect combination of circulation adjustment and elements.

[0003] Therefore, the present invention provides an intelligent plum rain forecasting method for a region based on circulation adjustment and element judgment, which improves the above technical problems. Summary of the Invention

[0004] The embodiments of the present disclosure aim to address the deficiencies of the prior art and provide an intelligent plum rain forecasting method for a region based on circulation adjustment and element judgment. The intelligent plum rain forecasting method for a region based on circulation adjustment and element judgment of the present invention combines characteristic quantities with hierarchical analysis and intelligent recognition to complete an intelligent, accurate, and practical plum rain forecasting method for forecasting the specific start time of the plum rain during the flood season in the Yangtze-Huaihe River Basin every year.

[0005] The above technical objectives of the present invention are achieved through the following technical solutions: An intelligent plum rain forecasting method for a region based on circulation adjustment and element judgment includes the following steps: S1. Determine multiple key characteristic quantities affecting the onset of the plum rain, and calculate the membership degree of each characteristic quantity based on historical data; S2. Construct a plum rain prediction credibility index, which is obtained by weighted summation of the membership degrees of each characteristic quantity, and the weight distribution is determined by the analytic hierarchy process; S3. Based on a long-sequence historical data set, train a plum rain classification prediction model using a support vector machine, decision tree, or Logistic regression algorithm. The inputs of the model include the credibility index and the membership degree of the key characteristic quantity; S4. Real-time obtain the circulation element data of the global forecasting system, extract the key characteristic quantity and calculate its membership degree, and input it into the prediction model for plum rain onset condition classification judgment; S5. Combine the ground element real-time monitoring and forecasting data to determine whether the plum rain monitoring standards are met, including regional precipitation threshold, rain day ratio, and average temperature conditions, and comprehensively determine the plum rain onset date.

[0006] As a preferred technical solution of the present invention, multiple key characteristic quantities affecting the onset of Meiyu include: the wind direction at Kolkata at 500 hPa, the subtropical high ridge line, the South Asian high ridge line, the westerly jet, the northeast cold vortex, the pseudo-equivalent potential temperature at 850 hPa, the low-level jet at 850 hPa, the humidity at 850 hPa, and the -8 °C line at 500 hPa.

[0007] As a preferred technical solution of the present invention, the method for determining the position of the subtropical high ridge line includes: within the range of 10 - 60°N and on the 120°E meridian, identifying the subtropical high center within the range of the 588 geopotential decameter line, and determining the latitude value when its zonal wind u = 0 and du / dy > 0; The method for determining the position of the South Asian high ridge line includes: within the range of 10 - 60°N and 100 - 120°E, calculating the average value of the latitude values at different longitudes within the range of the 1256 geopotential decameter line that satisfy the zonal wind u = 0 and du / dy > 0.

[0008] As a preferred technical solution of the present invention, the assignment rule for membership degree is: the membership degree of the characteristic quantity that meets the Meiyu onset condition is 1, the membership degree of the characteristic quantity that is close to the Meiyu onset condition is 0, and the membership degree of the characteristic quantity that does not meet the Meiyu onset condition is -1.

[0009] As a preferred technical solution of the present invention, the process of weight allocation using the analytic hierarchy process is: constructing a judgment matrix through expert experience, calculating the weight values of each characteristic quantity, and performing a consistency test. When the consistency ratio is less than 0.1, it is determined that the weight allocation is effective.

[0010] As a preferred technical solution of the present invention, the support vector machine adopts a radial basis function kernel or a polynomial kernel, and the formula for the polynomial kernel is: ; where and represent input vectors, represents the dot product of these two vectors, and d is the degree of the polynomial kernel.

[0011] As a preferred technical solution of the present invention, the decision tree adopts the CART algorithm, and uses the Gini index as the splitting criterion. The calculation formula for the Gini index is: ; where represents the proportion of class k in the data set D, D represents the current data set, A represents a certain feature, and is used to split the data set D into two subsets and , represents the total number of samples in the data set D, and respectively represent the subsets and The number of samples in the medium, the Gini coefficient ( ), and the Gini coefficient ( ) respectively represent the Gini coefficients of the subsets and . The value range of the Gini index is between [0, 1], and the smaller the value, the higher the purity of the data set.

[0012] As a preferred technical solution of the present invention, the circulation element data obtained in real time includes: ERA5 reanalysis data, and the forecast data is matched to the monitoring stations based on the grid interpolation method.

[0013] As a preferred technical solution of the present invention, the plum rain monitoring standard further includes: a. The daily precipitation of more than one-third of the stations in the monitoring area is ≥0.1 mm, and the regional daily average precipitation is ≥2.0 mm; b. The proportion of rainy days within 10 days starting from the first rainy day is ≥50%; c. The daily average temperature is ≥22°C.

[0014] As a preferred technical solution of the present invention, the rule for comprehensively determining the beginning date of the plum rain is: if the circulation adjustment date is earlier than the date when the surface element conditions are met, then the date when the surface element conditions are met is taken as the beginning date of the plum rain; if the two dates are the same, then the circulation adjustment date is taken as the beginning date of the plum rain.

[0015] In summary, the present invention has the following beneficial effects: First: By comprehensively screening 9 key circulation characteristic quantities such as the subtropical high ridge line, the South Asian high ridge line, and the 500 hPa - 8°C line, and combining surface precipitation, the proportion of rainy days, and temperature elements, the plum rain onset conditions are quantified in multiple dimensions, effectively avoiding the limitations of a single index and significantly improving the accuracy of the forecast. The embodiment shows that the predicted beginning date of the plum rain in the Yangtze-Huaihe River Basin in 2024 is completely consistent with the actual monitoring results, verifying the reliability of the method.

[0016] Second: Machine learning algorithms such as support vector machines and decision trees are used to construct a classification prediction model, which can automatically process a large amount of historical data and real-time meteorological data, reducing the subjective errors of manual experience judgment. Through model training and integrated optimization, the intelligent identification of plum rain onset conditions is realized, greatly improving the forecast efficiency. At the same time, the concept of fuzzy mathematical membership degree is introduced, converting qualitative evaluations (such as "close to the plum rain onset conditions") into quantitative scores (membership degrees 1, 0, -1), and combining the analytic hierarchy process to assign dynamic weights to the characteristic quantities, making the calculation of the credibility index (T) more scientific and reasonable, and avoiding the defect of equal-weight addition of characteristic quantities in traditional methods.

[0017] Thirdly: Integrate the real-time circulation data of the global forecasting system (such as ERA5 reanalysis data), match it to the monitoring stations through grid interpolation technology, and achieve seamless integration of forecast data and real-time monitoring. Combining with the precipitation forecast of the model for the next week, early warnings can be provided 7-10 days in advance before the onset of the Meiyu season, providing key decision-making support for flood control and dispatching in the Yangtze-Huaihe River Basin and agricultural planning.

[0018] Fourthly: Through the analytic hierarchy process, hierarchically judge the circulation adjustment indicators (such as the jump of the subtropical high ridge line and the northward shift of the westerly jet stream) and surface element conditions (rainy day ratio, temperature threshold), clarify the priority rules (circulation adjustment is a necessary condition, and surface elements are sufficient conditions), avoid indicator conflicts, and enhance the logical rigor of the forecasting process. Brief Description of the Drawings

[0019] Figure 1 It is a flowchart of the intelligent forecasting method for the onset of the Meiyu season in a region based on circulation adjustment and element judgment provided by an embodiment of the present invention; Figure 2 It is a data architecture diagram provided by an embodiment of the present invention; Figure 3 It is a training framework diagram of the classification prediction model for the onset of the Meiyu season provided by an embodiment of the present invention; Figure 4 It is a schematic diagram of the construction of a long-sequence historical data set provided by an embodiment of the present invention; Figure 5 It is a schematic diagram of the principle of the support vector machine algorithm provided by an embodiment of the present invention; Figure 6 It is a flowchart for judging the circulation adjustment indicators provided by an embodiment of the present invention. Detailed Embodiments

[0020] The following further elaborates on the present application with reference to specific embodiments. The following embodiments will help those skilled in the art to further understand the present application, but do not limit the present application in any form. It should be noted that those of ordinary skill in the art can make several modifications and improvements without departing from the concept of the present application. These all fall within the protection scope of the present application.

[0021] In order to make the purpose, technical solutions and advantages of the present application clearer, the following further elaborates on the present application with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0022] It should be noted that if there is no conflict, the various features in the embodiments of the present application can be combined with each other, and all are within the protection scope of the present application. In addition, although functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order from the module division in the device or the flowchart. In addition, the terms "first", "second", "third", etc. used in this article do not limit the data and execution order, but only distinguish the same items or similar items with basically the same functions and effects.

[0023] Unless otherwise defined, all technical and scientific terms used in this specification have the same meaning as those commonly understood by those skilled in the technical field to which this application belongs. The terms used in this specification in the specification of this application are only for the purpose of describing specific embodiments and are not used to limit this application. The term "and / or" used in this specification includes any and all combinations of one or more of the related listed items.

[0024] In addition, the technical features involved in the various embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0025] The embodiments of the present disclosure aim to solve the problem of comprehensive analysis and judgment of the onset of plum rains based on a large amount of data. In view of this, the embodiments of the present disclosure propose a method for intelligent prediction of the onset of regional plum rains based on circulation adjustment and element judgment, which combines characteristic quantities with hierarchical analysis and intelligent recognition to complete an intelligent, accurate and practical method for predicting the onset of plum rains for the specific start time of the annual flood season plum rains in the Yangtze-Huaihe River Basin. This method combines the monitoring of the early circulation and surface element conditions and the prediction of future weather situation changes and elements, comprehensively considers the adjustment of the circulation situation, the specific weather conditions monitoring in different regions and the plum rain monitoring standards, uses the hierarchical analysis method and intelligent recognition judgment, and predicts the onset date in advance according to the indicators for the onset of plum rains in the Yangtze-Huaihe region.

[0026] Please refer to Figure 1 , Figure 1 which shows the flowchart of the method for intelligent prediction of the onset of regional plum rains based on circulation adjustment and element judgment according to the embodiments of the present disclosure. The overall process mainly includes the following 5 steps: Step 1: Determine multiple key characteristic quantities that affect the onset of plum rains, and calculate the membership degree of each characteristic quantity based on historical data.

[0027] S1.1. Multiple key characteristic quantities that affect the onset of plum rains include: the onset of the Indian monsoon (wind direction at Kolkata at 500 hPa), the subtropical high ridge line, the South Asian high ridge line, the westerly jet stream, the northeast cold vortex, the pseudo-equivalent potential temperature at 850 hPa, the low-level jet stream at 850 hPa, the humidity at 850 hPa, and the -8°C line at 500 hPa.

[0028] S1.2. Collect historical data of Meiyu-related factors. Through comparison and composite analysis before and after the onset of Meiyu, determine the change characteristic values of each characteristic quantity before and after the onset of Meiyu and the threshold values of ground truth monitoring elements, and determine the spatial range of the calculation value distribution of each characteristic quantity. As shown in Table 1: Table 1 Historical distribution of key circulation characteristic quantities: ; Continued Table 1: ; S1.3. Convert the qualitative evaluation of meteorological characteristic quantities for the onset of Meiyu into a quantitative evaluation.

[0029] Based on the historical data of meteorological circulation and elements for the onset date of Meiyu and the 10 days before and after, extract their key characteristic quantities respectively and form their membership degrees. Considering comprehensively the meteorological indicators that have obvious effects on Meiyu, based on the principle of membership degree in fuzzy mathematics, convert the qualitative evaluation of meteorological characteristic quantities for the onset of Meiyu into a quantitative evaluation; The specific process is as follows: Let the key characteristic quantity for the onset of Meiyu be T(i), determine the membership degree of each characteristic quantity, and establish: the membership degrees for meeting, approaching, and not meeting the Meiyu onset conditions are 1, 0, and -1 respectively. The key characteristic quantity N of the present invention is 9, and the extraction method and membership degree of each factor are as shown in Table 2: Table 2 Extraction method of characteristic quantities and membership degree assignment rules: ; Continued Table 2: ; In the table, represents the zonal wind at 200 hPa; represents the geopotential height at 500 hPa; represents the latent heat of water vapor; represents the specific heat capacity of air; represents the absolute temperature; represents the vertical velocity; represents the relative humidity; T represents the temperature.

[0030] Step 2: Construct a credibility index for Meiyu onset prediction. The credibility index is obtained by weighted summation of the membership degrees of each characteristic quantity, and the weight distribution is determined by the analytic hierarchy process.

[0031] Construct a credibility index for Meiyu onset prediction based on the circulation situation and elements: , where: T represents the credibility of meeting the plum rain onset conditions (the closer T is to 7, the more conducive to the onset of plum rain); T(G) represents the membership degree of the subtropical high ridge line; T(K) represents the membership degree of the South Asian high ridge line; T(t) represents the membership degree of -8°C at the point of 35N, 120E at 500hPa; T(r) represents the membership degree of 80% relative humidity in the area of (31 - 34, 110 - 120); T(h) represents the membership degree of the 200hPa westerly jet center; T(l) represents the membership degree of the 850hPa low-level jet center within the range of 110 - 120E; T(w) represents the membership degree of the wind direction at Kolkata at 500hPa; and 、 、 、 、 、 respectively represent the weight coefficients of each item.

[0032] Construct a judgment matrix (Table 3) based on the expert experience of each characteristic quantity, and calculate the consistency ratio CR to conduct a consistency test.

[0033] Table 3 Construct the judgment matrix of each characteristic quantity according to expert scoring ; Calculate the consistency index CI = (λmax - n) / (n - 1) using Table 3; where n is the order of the judgment matrix, and λmax is the maximum eigenvalue of the matrix.

[0034] Calculate the consistency ratio CR = CI / RI. Among them, the RI value is usually obtained by looking up the table, or relevant literature or materials can also be referred to.

[0035] If CR < 0.1, it is considered that the consistency of the judgment matrix is acceptable and the weight allocation is reasonable.

[0036] If CR ≥ 0.1, it is considered that the consistency of the judgment matrix is poor and the judgment matrix needs to be corrected.

[0037] On the basis of meeting the consistency test, use the eigenvector method to calculate the weights of each characteristic quantity and obtain 、 、 、 、 、 、 weight coefficients.

[0038] The above-mentioned method calculates the daily plum rain onset credibility index according to the membership degrees of each characteristic quantity, determines the daily plum rain onset category, and forms a historical dataset of the membership degrees of each factor characteristic quantity, the credibility index, and the plum rain onset category. The training dataset and the test dataset are constructed respectively at the ratios of 80% and 20%. The data architecture is as Figure 2 shown, presenting the structure of the historical dataset of the membership degrees of key characteristic quantities, the credibility index, and the classification labels.

[0039] Step 3: Based on the long-sequence historical dataset, use the support vector machine, decision tree or Logistic regression algorithm to train the plum rain onset classification prediction model. The inputs of this model include the credibility index and the membership degrees of key characteristic quantities. As Figure 3 shown, it includes an input layer, a machine learning algorithm layer (support vector machine, decision tree, etc.), a model evaluation layer, and a prediction network layer.

[0040] S3.1 Use the plum rain onset actual data in the Yangtze-Huaihe region from 1991 to 2020 (the plum rain actual record data of the Jiangsu Meteorological Observatory), the ERA5 analysis data (the global meteorological reanalysis dataset released by the European Centre for Medium-Range Weather Forecasts (ECMWF)) and the corresponding characteristic quantity algorithms to calculate and extract different characteristic quantities, and establish the key characteristic quantity data of the long-sequence plum rain meteorological factors according to the membership degree thresholds of different characteristic quantities. Select 10 days before and after the plum rain onset every year for calculation. Calculate the daily plum rain onset prediction credibility index. According to the historical plum rain onset date, it is recorded as -1 before the plum rain onset and 1 after the plum rain onset, and is divided into two categories in total, forming a judgment case library with 20 * 71 cases, and forming a long-sequence dataset including the membership degrees of 7 key characteristic quantities, the credibility index, and whether it is the plum rain onset classification. As Figure 4 shown, presenting the characteristic quantity extraction process based on the ERA5 reanalysis data and the actual data.

[0041] S3.2 Based on the support vector machine and decision tree, construct the plum rain onset date prediction model. This method includes a total of four layers: The first layer: the input layer, input the preprocessed training dataset X and the relevant classification Y.

[0042] The second layer: the data processing network layer of different machine learning methods, train and learn the dataset X and the relevant classification Y through different methods. The steps of different methods are as follows: Support vector machine (SVM): Select the radial basis function (RBF) kernel or polynomial kernel to increase the data dimension and explore complex data relationships. As Figure 5 shown, presenting the hyperplane division and kernel function mapping process.

[0043] Its equation can be expressed as: , where is the hyperplane normal vector, b is the intercept. For any point x, if: , then it is classified as the positive class; It is then classified as a negative class. The SVM model is trained using the polynomial kernel function and the optimized parameters to determine the support vectors.

[0044] The formula for the polynomial kernel function is as follows: ; where and represent the input vectors, represents the dot product of these two vectors, d is the degree of the polynomial kernel, which needs to be selected according to the specific problem.

[0045] Decision tree: Select the improved CART and use the Gini index as the splitting criterion. For a dataset D containing K classes, the formula for its Gini index is: ; where represents the proportion of class k in the dataset D, D represents the current dataset, A represents a certain feature, which is used to split the dataset D into two subsets and , represents the total number of samples in the dataset D, and respectively represent the number of samples in the subsets and , and the Gini coefficient ( ) and the Gini coefficient ( ) respectively represent the Gini coefficients of the subsets and . The value range of the Gini index is between [0, 1], and the smaller the value, the higher the purity of the dataset.

[0046] The third layer: the model evaluation layer, which uses the test set data to evaluate the trained prediction model. Common metrics: Accuracy, Recall, F1 score, etc.

[0047] , , , ; TP represents the number of samples that the model correctly predicts the positive class as the positive class; TN represents the number of samples that the model correctly predicts the negative class as the negative class; FP represents the number of samples that the model incorrectly predicts the negative class as the positive class; FN represents the number of samples that the model incorrectly predicts the positive class as the negative class; Precision represents how many of the samples predicted as positive samples by the model are truly positive samples.

[0048] Fourth layer: Prediction network layer, which determines the final prediction models of the two types of methods according to different methods and tests.

[0049] Step 4: Obtain the circulation element data of the global forecast system in real time, extract key feature quantities and calculate their membership degrees, and input them into the prediction model for classification judgment of the plum rain onset conditions.

[0050] S4.1 Calculate the spatial range to be determined and read according to different feature quantities respectively. For example, when reading the relative humidity, low-level jet, and westerly jet, the different latitude values in the range of 110-120E are read. For the Northeast Cold Vortex, the 500hPa average height in the area of 40-50°N, 125-135°E is applied. The wind in Kolkata is the wind direction at this station at 500hPa. The temperature change at 500hPa - 8℃ is at 35N, 120E.

[0051] S4.2 Use algorithms of different feature quantities (subtropical high ridge line, South Asian high ridge line, westerly jet, pseudo-equivalent potential temperature) to intelligently identify and calculate and extract information of different feature quantities.

[0052] Method for determining the position of the 120E subtropical high ridge line: The latitude where the position of the subtropical high center (within the 588 geopotential decameter range, the zonal wind u = 0, and du / dy > 0) is located on the 120E meridian within the range of 10-60N.

[0053] Method for determining the position of the South Asian high ridge line: The average value of the latitudes where the positions are located at different longitudes that satisfy the zonal wind u = 0 and du / dy > 0 within the range of 10-60N and 100-120E and within the range of 1256 geopotential decameters.

[0054] Method for calculating the position of the westerly jet: The latitude corresponding to the maximum value of the average value of the U-wind values of different grid points at the same latitude within the range of 20-60N in the 200hPa height field at 110-120E.

[0055] Value of pseudo-equivalent potential temperature: The average value of the pseudo-equivalent potential temperature at 850hPa within the range of 20-60N at 110-120°E. When the 340K isotherm > 30N, the plum rain conditions in the Yangtze-Huaihe region are satisfied.

[0056] Wind in Kolkata: Read the actual wind direction and speed at 500hPa at the Kolkata station and the wind direction and speed of the nearest grid point of this station in the forecast field.

[0057] S4.3 Use the analytic hierarchy process to comprehensively judge whether the circulation adjustment index and element index for reaching the plum rain onset conditions can be achieved, and objectively determine the circulation adjustment date.

[0058] The criteria for hierarchical analysis include: (1) The wind direction at 500 hPa in Kolkata changes from westerly to easterly or alternates between east and west. (2) The ridge line of the South Asian High at 110 - 120°E is greater than or equal to 28°N, and the ridge line of the subtropical high at 120°E is greater than or equal to 20°N. (3) The center of the westerly jet at 200 hPa jumps from 32 - 35°N to the range of 36 - 42°N. (4) The centers of the low-level jets at 700 and 850 hPa within the range of 110 - 120°E reach 28 - 30°N, and the 80% humidity area is located at 31 - 34°N. (5) The pseudo-equivalent potential temperature line at 850 hPa within the range of 110 - 120°E, 31 - 34°N is greater than 340 K. (6) The temperature at the point of 35°N, 120°E at 500 hPa is greater than -8°C. When all the above conditions are met, the adjustment of the circulation situation reaches the plum rain onset condition. The specific analysis and discrimination sequence and the threshold values of each index are as Figure 6 shown, demonstrating the hierarchical judgment logic of key indicators such as the jump of the subtropical high ridge line and the northward movement of the westerly jet.

[0059] S5. Combining the ground element real-time monitoring and forecast data, judge whether the plum rain monitoring criteria are met, including the regional precipitation threshold, the proportion of rainy days, and the average temperature condition, and comprehensively determine the plum rain onset date.

[0060] On the basis of meeting the plum rain onset conditions, taking the circulation adjustment day as a necessary condition and the ground elements as a sufficient condition, conduct a comprehensive judgment of the ground element conditions according to the plum rain monitoring criteria in combination with the current precipitation situation and the precipitation forecast of the model for the next week; specific criteria: (1) More than one-third of the monitoring stations in the monitoring area have precipitation greater than or equal to 0.1 mm, and the average daily precipitation in the area is greater than or equal to 2.0 mm, which is defined as a rainy day. (2) Starting from the first rainy day, the proportion of the number of rainy days in the next 10 days to the total number of days in the corresponding period is greater than or equal to 50%. (3) The average daily temperature is greater than or equal to 22°C.

[0061] The calculation formula for the monitoring criteria of the station area (south of the Huaihe River, between the Yangtze and Huaihe Rivers, southern Jiangsu) is: Temperature condition: ; Regional precipitation condition: ; Rainy day persistence condition: , meeting the rainy day condition; In the formula, represents the average temperature of the monitoring stations in the area; represents the average precipitation of the stations in the area; represents the total number of monitoring stations in the monitoring area; N represents the number of stations with a single-station precipitation greater than or equal to 0.1 mm in the area; represents the total precipitation in the next 10 days.

[0062] Example: An intelligent recognition method for determining the date of the onset of the Meiyu season in different regions of the Jianghuai region is formulated according to the monitoring standards for the onset of the Meiyu season. By using real-time monitoring and model precipitation forecasts and adopting a step-by-step judgment method, the intelligent judgment of the date of the onset of the Meiyu season is realized.

[0063] The specific steps are as follows: a: According to the Meiyu monitoring stations and monitoring standards in Jiangsu, standardize and objectify the monitoring standards for the onset of the Meiyu season step by step to lay a foundation for intelligent recognition. Determine the representative stations in the areas south of the Huaihe River, between the Yangtze River and the Huaihe River, and in the middle and lower reaches of the Yangtze River, and set the values of one-third of their stations. The element statistics should meet the following conditions: (1) In the monitoring area, more than 1 / 3 of the monitoring stations have precipitation greater than or equal to 0.1 mm, and the average daily precipitation in the area is greater than or equal to 2 mm; (2) Starting from the first rainy day, the proportion of rainy days in the next 10 days is greater than or equal to 50%; (3) The average temperature is greater than or equal to 22.

[0064] b: Interpolate the grid precipitation forecast and temperature forecast to the Meiyu monitoring stations to achieve a perfect integration of real-time monitoring and forecast fields, and form time-series data of precipitation and temperature.

[0065] c: Construct an automatic algorithm for each factor to determine whether the conditions for the onset of the Meiyu season are met every day. When the three conditions in a are met, the first rainy day is determined as the onset date of the Meiyu season under the monitoring standard.

[0066] When the separately obtained circulation adjustment date occurs before the onset date of the Meiyu season under the monitoring standard, the onset date of the Meiyu season under the monitoring standard is the official onset date of the region. When the two occur on the same date, the circulation adjustment date is the onset date of the region.

[0067] The above is only the preferred embodiment of the present invention. The protection scope of the present invention is not limited to the above embodiments. All technical solutions within the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for those of ordinary skill in the art in this technical field, several improvements and refinements made without departing from the principle of the present invention should also be regarded as within the protection scope of the present invention.

Claims

1. A method for intelligent prediction of the onset of the Meiyu season in a region based on circulation adjustment and factor judgment, characterized in that, The method includes the following steps: S1. Determine multiple key characteristic quantities affecting the onset of Meiyu, and calculate the membership degree of each characteristic quantity based on historical data; S2. Construct a credibility index for Meiyu onset prediction, which is obtained by weighted summation of the membership degrees of each characteristic quantity, and the weight assignment is determined by the analytic hierarchy process; S3. Based on a long-sequence historical data set, train a classification prediction model for Meiyu onset using a support vector machine, decision tree or Logistic regression algorithm. The inputs of the model include the credibility index and the membership degrees of key characteristic quantities; S4. Obtain the circulation element data of the Global Forecast System in real time, extract the key characteristic quantities and calculate their membership degrees, and input them into the prediction model to judge the classification of Meiyu onset conditions; S5. Combine the ground element real-time monitoring and forecast data to judge whether the Meiyu monitoring standards are met, including the regional precipitation threshold, the proportion of rainy days and the average temperature condition, and comprehensively determine the Meiyu onset date.

2. The intelligent prediction method for the onset of plum rains in a region based on circulation adjustment and factor judgment according to claim 1, wherein The multiple key characteristic quantities affecting the onset of Meiyu include: the wind direction at 500 hPa in Kolkata, the subtropical high ridge line, the South Asian high ridge line, the westerly jet, the northeast cold vortex, the pseudo-equivalent potential temperature at 850 hPa, the low-level jet at 850 hPa, the humidity at 850 hPa, and the 500 hPa -8 °C line.

3. The intelligent prediction method for the onset of plum rains in a region based on circulation adjustment and factor judgment according to claim 2, characterized in that, The method for determining the position of the subtropical high ridge line includes: within the range of 10 - 60°N and on the 120°E meridian, identify the subtropical high center within the range of the 588 geopotential decameter line, and judge the latitude value when the zonal wind u = 0 and du / dy > 0; The method for determining the position of the South Asian high ridge line includes: within the range of 10 - 60°N and 100 - 120°E, calculate the average value of the latitude values at different longitudes that satisfy u = 0 and du / dy > 0 within the range of the 1256 geopotential decameter line; 4. The intelligent prediction method for the onset of plum rains in a region based on circulation adjustment and element judgment according to claim 1, characterized in that, The assignment rule of the membership degree is: the membership degree of the characteristic quantity that meets the Meiyu onset condition is 1, the membership degree of the characteristic quantity that is close to the Meiyu onset condition is 0, and the membership degree of the characteristic quantity that does not meet the Meiyu onset condition is -1.

5. The intelligent prediction method for the onset of plum rains in a region based on circulation adjustment and factor judgment according to claim 1, characterized in that, The process of weight assignment using the analytic hierarchy process is: construct a judgment matrix through expert experience, calculate the weight values of each characteristic quantity, and conduct a consistency test. When the consistency ratio is less than 0.1, the weight assignment is determined to be effective.

6. The intelligent prediction method for the onset of plum rains in a region based on circulation adjustment and factor judgment according to claim 1, wherein The support vector machine adopts a radial basis function kernel or a polynomial kernel, and the formula of the polynomial kernel is: ; where and represent input vectors, represents the dot product of these two vectors, and d is the degree of the polynomial kernel.

7. The intelligent prediction method for the onset of plum rains in a region based on circulation adjustment and factor judgment according to claim 1, wherein The decision tree uses the CART algorithm, with the Gini index as the splitting criterion. The calculation formula of the Gini index is: ； Among them, represents the proportion of class k in the dataset D, where D represents the current dataset, and A represents a certain feature used to divide the dataset D into two subsets and , represents the total number of samples in the dataset D, and respectively represent the number of samples in the subsets and . The Gini coefficient ( ) and the Gini coefficient ( ) respectively represent the Gini coefficients of the subsets and . The value range of the Gini index is between [0, 1], and the smaller the value, the higher the purity of the dataset.

8. The intelligent prediction method for the onset of plum rains in a region based on circulation adjustment and factor judgment according to claim 1, characterized in that, The circulation element data obtained in real time includes: ERA5 reanalysis data, and the forecast data is matched to the monitoring stations based on the grid interpolation method.

9. The intelligent prediction method for the onset of plum rains in a region based on circulation adjustment and factor judgment according to claim 1, wherein The Meiyu monitoring standards further include: a. For more than one-third of the stations in the monitoring area, the daily precipitation ≥ 0.1 mm, and the regional daily average precipitation ≥ 2.0 mm; b. The proportion of rainy days ≥ 50% within 10 days starting from the first rainy day; c. The daily average temperature ≥ 22 °C.

10. The intelligent prediction method for the onset of plum rains in a region based on circulation adjustment and factor judgment according to claim 1, wherein, The rule for comprehensively determining the Meiyu onset date is: if the circulation adjustment date is earlier than the date when the ground element conditions are met, then the date when the ground element conditions are met is taken as the Meiyu onset date; if the two dates are the same, then the circulation adjustment date is taken as the Meiyu onset date.

Images