A real-time rainfall and rainfall duration prediction method based on Beidou ZWD

Abstract

The present invention proposes a real-time rainfall and rainfall duration prediction method based on Beidou ZWD, comprising the following steps: step 1, extracting ZWD values, corresponding rainfall amounts and corresponding time series according to historical observation values ​​of Beidou navigation satellites; step 2, calculating the hourly change rate of ZWD values ​​close to 12 hours according to different seasons; step 3, using ZWD values ​​and corresponding hourly change rates, and corresponding rainfall data to establish rainfall and rainfall duration prediction thresholds; step 4, obtaining the observation values ​​of Beidou navigation satellites at the time to be predicted, and extracting the ZWD values ​​at the time to be predicted, and then outputting the rainfall and rainfall duration prediction results according to the established rainfall and rainfall duration prediction thresholds. This method proposes a method for predicting rainfall duration based on Beidou for the first time, which can be effectively applied to real-time detection of rainfall and rainfall duration.

Description

Technical Field

[0001] The present invention relates to the technical field of global satellite navigation meteorology, and in particular to a method for predicting real-time rainfall and rainfall duration based on Beidou ZWD. Background Art

[0002] Rainfall and rainfall duration are closely related to people's daily lives. Small and medium rainfall will affect people's travel preparations in daily life, while heavy rain and rainstorms are likely to cause urban waterlogging or floods. Therefore, accurate rainfall and rainfall duration forecasts can effectively provide references for human travel, and can also provide references for flood control and disaster prevention for relevant departments.

[0003] At present, there are four main methods for real-time rainfall prediction. The first method is the traditional numerical assimilation forecasting method, which is based on the reanalysis of a large number of multi-source meteorological data and the use of mathematical numerical calculation models to predict future sequences, such as radiosonde data, European reanalysis data, and data provided by the National Center for Environmental Prediction and NASA. This method relies heavily on historical data and data assimilation algorithms, and the differences in spatiotemporal resolution of different data sources will lead to reduced prediction accuracy. The second method is based on meteorological remote sensing satellites, such as China's Fengyun series satellites and the United States' Ocean series satellites. This method can analyze the atmosphere from a high level and improve the spatial coverage and spatial resolution of weather forecasts. However, since this type of satellite is a low-orbit satellite and the satellite repetition period interval is long, the temporal resolution is low and it is easily disturbed by the environment, resulting in reduced prediction accuracy. The third type is based on the meteorology of the global navigation satellite system, such as using GPS satellites, Beidou, etc. This type of method uses the signal delay caused by global navigation satellites passing through the troposphere to invert the tropospheric delay, further derive the atmospheric precipitable amount related to rainfall, and then perform weather forecasts. Due to the high coverage of global satellite navigation satellites and high temporal resolution, it has been widely studied. However, the existing methods based on atmospheric precipitable amount require the use of meteorological parameters, which leads to the application of the method first, and the methods directly based on tropospheric delay introduce the error of zenith mechanical delay, resulting in reduced prediction accuracy. In addition, there has been no effective method for predicting rainfall duration. Therefore, studying accurate and efficient real-time rainfall and rainfall duration prediction methods to provide strong support for meteorological departments is a hot and difficult issue that needs to be solved urgently. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention proposes a real-time rainfall and rainfall duration prediction method based on Beidou ZWD, which not only makes up for the deficiency that the meteorological parameters need to be coordinated when predicting rainfall based on atmospheric precipitation, but also solves the problem of reduced prediction accuracy due to the introduction of zenith statics delay in the Beidou zenith total delay method. It is also the first time that a method for predicting rainfall duration based on Beidou has been proposed, which can be effectively applied to real-time detection of rainfall and rainfall duration.

[0005] In order to solve the above technical problems, the technical solution of the present invention is:

[0006] A real-time rainfall and rainfall duration prediction method based on Beidou ZWD, the specific steps are as follows:

[0007] Step 1: Extract valid ZWD and related predictions, mainly including ZWD predictions, corresponding rainfall, and corresponding time series. First, the original observations of Beidou navigation satellites are received through geodetic receivers, mainly including observation files and navigation files. At the same time, the corresponding precise ephemeris, precise clock error and other precise point positioning related error correction products are downloaded in real time from the official website of IGS (International GNSS Service). The above files can be obtained from official channels and are common knowledge in the field. Secondly, the received and downloaded related files are solved. The GEO satellite weighting strategy is used in the solution process to extract accurate tropospheric delay parameters, which are recorded as ZTD here. Finally, the tropospheric static delay formula is used to calculate the tropospheric static delay at the corresponding epoch, and the relationship between ZWD, ZTD and tropospheric static delay is used to calculate the accurate ZWD prediction. The calculation method of tropospheric static delay can be obtained by using the Sa algorithm or the latest GPT3 algorithm. It should be noted that the Sa algorithm and the GPT3 algorithm are both classic algorithms in the field of tropospheric static delay. Both algorithms have specific formulas for reference, which are common knowledge in the field and will not be described in detail here. To extract rainfall, it is necessary to download it from the official website of the Meteorological Bureau or the websites of some meteorological product companies. In addition, the corresponding time series can be obtained directly from the observation file, which records the time information of each set of data, thereby obtaining the time series information corresponding to ZWD.

[0008] Step 2: Calculate the rainfall and rainfall duration prediction thresholds based on ZWD.

[0009] For the rainfall prediction threshold: First, set the initial rainfall prediction threshold; then, count the rainfall events in the historical data, and use the ZWD obtained in the first step and the corresponding time series information to calculate other prediction quantities, including the 1-hour change rate of ZWD, the 2-hour change rate,..., the 11-hour change rate, and the change rate up to the 12-hour change rate, a total of 12 values. Finally, use the correspondence between the 12 change rates and the corresponding rainfall. When at least one of the 12 values ​​is greater than the initial prediction threshold, it is predicted to be a rain event; if there is no value greater than the initial prediction threshold, it is predicted to be a no-rain event; and record all the prediction results, where the prediction results include the prediction of rainfall and the actual rainfall, recorded as P ture ; Rainfall is predicted, but there is no rainfall in reality, recorded as P false ; No rainfall is predicted, but there is rainfall in reality, recorded as P miss ; Calculate the prediction success rate, prediction error rate and prediction loss rate based on the prediction results; Adjust the initial rainfall prediction threshold and repeat the above steps; Select the optimal rainfall prediction threshold based on the prediction success rate, prediction error rate and prediction loss rate. The prediction threshold is stored in the form of a database on the local computer.

[0010] For the rainfall duration prediction threshold:

[0011] First, classify the ZWD value according to the ZWD obtained in the first step and the corresponding rainfall. ZWD is divided into six categories: <230, 230-280, 280-330, 330-380, 380-430, >430. The above classification is mainly for illustration, and the actual process will be adjusted according to rainfall and season. The average rainfall duration of all rainfall durations in each category is averaged to obtain the average rainfall duration; the rainfall duration prediction reference value is determined based on the average rainfall duration; finally, the rainfall duration prediction reference value corresponding to each category is used as a data pair, so that the data pair of each category is used as the rainfall duration prediction threshold. The prediction threshold is stored in the form of a database on the local computer.

[0012] Step 3: Predict rainfall based on the rainfall prediction threshold of ZWD.

[0013] First, collect the original data of Beidou satellite PPP positioning, calculate ZWD and the corresponding hourly change rate. The ZWD calculation process is the same as the first and second steps above, which will not be repeated here.

[0014] Secondly, the ZWD calculated above and the corresponding hourly change rate are used in combination with the rainfall prediction threshold to predict rainfall. The prediction process is to compare the hourly change rate with the established threshold. If 40% of the change rate exceeds the threshold, light rain is predicted. If 60% of the change rate exceeds the threshold, moderate rain is predicted. If 90% of the change rate exceeds the threshold, heavy rain is predicted.

[0015] Step 4: Further predict the duration of rainfall based on the results of ZWD rainfall prediction. First, statistically analyze the rainfall duration of ZWD under different segments based on the previously calculated ZWD and the collected rainfall data to obtain a rainfall duration prediction table based on statistical data. Secondly, combine the results of ZWD rainfall prediction and ZWD value, as well as the rainfall duration prediction table, to predict the duration of the rain. Finally, save and output the predicted rainfall duration results.

[0016] Step 5: Save and output the above-mentioned rainfall and rainfall duration prediction results based on Beidou ZWD, and use the new data as modeling data to update the prediction model to enhance the accuracy and stability of the prediction algorithm.

[0017] The above-mentioned Beidou ZWD-based rainfall and rainfall duration prediction algorithm can effectively solve the shortcomings of the existing PWV-based algorithm that is overly dependent on meteorological data, while avoiding the problem of reduced algorithm accuracy caused by the introduction of zenith statics delay in the algorithm directly using ZTD. In addition, the problem of rainfall duration prediction based on navigation satellites was also experimented for the first time, effectively improving the practicality, accuracy and stability of the rainfall prediction algorithm, and providing guarantees for meteorological and water conservancy departments to make timely and effective decisions.

[0018] The present invention has the following characteristics and beneficial effects:

[0019] The above technical solution can effectively solve the problem of high spatiotemporal resolution prediction of rainfall and rainfall duration, and provide guarantee for meteorological and water conservancy departments to make timely and effective decisions. Beidou satellite observation data is used to extract ZWD parameters, and the relationship between ZWD data and rainfall data is used to establish rainfall prediction thresholds. The relevant thresholds are stored in the form of text on the local computer to prepare for subsequent rainfall prediction. The Beidou GEO satellite weighting strategy can effectively improve the extraction accuracy of ZWD and improve the accuracy of rainfall and rainfall duration prediction. The use of multiple time change rates of ZWD for simultaneous prediction can further improve the accuracy and precision of rainfall prediction. Using the rainfall prediction results and the statistical relationship between rainfall and ZWD, the rainfall duration of each rainfall can be accurately predicted, and the application scenarios and actual use scope of the algorithm can be improved. Through the above algorithm, it can effectively solve the problem of limited use of the PWV algorithm due to excessive reliance on meteorological parameters, and at the same time solve the problem of reduced algorithm accuracy caused by the introduction of zenith static delay in the algorithm directly based on ZTD. In addition, the proposed method also uses navigation satellites for the first time to predict rainfall duration, which further improves the practicality of the algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0021] Figure 1 This is a flow chart of a real-time rainfall and rainfall duration prediction method based on Beidou ZWD.

[0022] Figure 2 This is a flow chart of the real-time rainfall prediction method based on Beidou ZWD.

[0023] Figure 3 This is a flow chart of the real-time rainfall duration prediction method based on Beidou ZWD. DETAILED DESCRIPTION

[0024] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0025] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and the like are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first", "second", and the like may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0026] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood by specific circumstances.

[0027] The present invention mainly provides a real-time rainfall and rainfall duration prediction method based on Beidou navigation satellite ZWD. The method mainly adopts ZWD observation values ​​extracted based on Beidou satellites and useful information such as instantaneous change rate related to ZWD to predict rainfall and rainfall duration in real time, and adopts an adaptive threshold strategy that takes into account time and space differences to set the prediction threshold, so as to effectively overcome the problem that the prediction method fails due to the influence of seasonal fluctuations on rainfall. At the same time, the weight of Beidou GEO (Geostationary orbit, geostationary orbit satellite) satellite in solving ZWD is improved, which can effectively improve the prediction accuracy of rainfall and rainfall duration. The overall method can be divided into two parts. The first part is to establish a rainfall and rainfall duration prediction threshold database by collecting Beidou satellite observation data and rainfall data over the years. First, Beidou epoch observation data is collected. Beidou satellites have provided services in the Asia-Pacific region since December 20, 2012. Historical data can be continuously collected, and tropospheric delay is extracted using Beidou precise single-point positioning technology, and then the corresponding ZWD observation value is solved using the relationship between tropospheric delay and ZWD. Secondly, collect meteorological data of the corresponding time, mainly including rainfall and rainfall duration, which can be obtained from the official website of the meteorological department. Finally, using the correlation between ZWD and rainfall and rainfall duration, establish a corresponding prediction threshold database, which is stored in the form of text on the local computer. The second step is to predict rainfall and rainfall duration based on the database established in the first step. First, collect Beidou satellite observation data for the day of prediction in real time and calculate the corresponding ZWD prediction value. Secondly, use the calculated ZWD observation value and the database established in the first step to predict rainfall. Finally, use the rainfall prediction results and the database established in the first step to predict rainfall duration.

[0028] Specifically, Figure 1 As shown, the following steps are included:

[0029] Step 1, satellite observation value collection. In this embodiment, a large number of original observation values ​​of Beidou satellite PPP (Precise Point Positioning) positioning and the precision products required for solution are collected, mainly including pseudorange, carrier phase observation values, precise ephemeris, precise clock error, and corresponding earth nutation, in preparation for extracting precise ZTD. Among them, the original observation data of Beidou satellite PPP positioning can be directly downloaded from the official website of the local surveying and mapping bureau or seismic bureau, mainly including observation files. The corresponding navigation files and the precision products required for solution can be obtained through the IGS official website.

[0030] It should be noted that IGS is an organization that provides various types of navigation satellite precision products worldwide, and can provide corresponding data products to the world free of charge. The above content is common knowledge in the field of navigation and positioning, and will not be explained in detail here.

[0031] Step 2: Use the weighted least squares algorithm and GEO satellite weighting strategy to solve the high-precision ZTD. The calculation process is as follows:

[0032] First, the data collected in the first step are positioned and solved according to the PPP positioning mode. The parameters can be solved by using the classic least squares algorithm or the extended Kalman filter algorithm. The weighted least squares algorithm makes the parameter results more consistent.

[0033] Therefore, in this embodiment, a weighted least squares algorithm is used to solve the parameters. The weighted least squares algorithm belongs to the commonly used algorithm in the field of parameter estimation, and will not be described in detail here. It should be noted that the parameters solved here include the position parameters of the receiver, the receiver clock difference parameters, and the corresponding ZTD parameters. It can be understood that only the ZTD parameters are used in this embodiment.

[0034] Secondly, in the process of solving ZTD using the weighted least squares algorithm, the GEO satellite weighting strategy is adopted to improve the accuracy of ZTD solution. It should be noted that since there are three types of Beidou satellite orbits, and GEO satellites belong to high-orbit satellites, the signal propagation distance is long and the signal energy loss is large, so in this embodiment, GEO is weighted to reduce the proportion of GEO satellites in the solution and improve the accuracy of parameter solution. The specific process is as follows:

[0035] Assume that the least squares solution expression is expressed as:

[0036] (1)

[0037] Where G represents the Jacobian matrix composed of the observation equation. Δx represents the residual of the parameter to be solved. For the convenience of understanding, it can be simply understood as containing the parameters that need to be solved, among which the ZTD parameters are included. b represents the difference between the input observation value and the output observation value. Formula 1 is the most basic solution formula of least squares, which is common knowledge in the field and will not be elaborated here. In order to further improve the accuracy of parameter solution, it is necessary to weight formula 1 here. The weighted formula is expressed as:

[0038] (2)

[0039] Wherein, W is the weight matrix, which is an N×N diagonal matrix. Since the weights are applied to GEO satellites in this embodiment, the weights are expressed according to the satellite type, and the W weight matrix in this embodiment can be expressed as:

[0040] (3)

[0041] In the formula, the subscript of each coefficient represents the satellite orbit type, which are GEO, IGSO and MEO.

[0042] It should be noted that due to the poor quality of GEO satellite signals, its weight coefficient should be smaller, and the weight coefficients of the three satellite types can be set to 1:3:3. It can be understood that this coefficient can be set manually. If the number of satellites is large enough, the weight of GEO satellites can be further reduced. On the contrary, if the number of satellites is small, the weights of the three satellites are equal.

[0043] Finally, the ZTD result calculated above is saved and output. Through the above process, the ZTD parameters can be accurately calculated using Beidou satellite.

[0044] Step 3: Use the Sa empirical model to calculate ZHD, and use the relationship among ZTD, ZWD, and ZHD to calculate ZWD. The specific process is as follows:

[0045] First, use the Sa empirical model to calculate ZHD. The Sa empirical model is an empirical model algorithm for calculating ZHD proposed by Saastamoinen in 1972. This algorithm can simply and accurately calculate the ZHD of any location in the world and is still used today. The calculation formula is as follows:

[0046] (4)

[0047] In the formula, φ represents the latitude of the Beidou receiver (the device that collects Beidou observation data), and H represents the height of the Beidou receiver, which refers to the altitude, not the height relative to the ground. g Indicates the pressure at the sea level where the receiver is located. The ZHD parameter can be accurately calculated using the above formula.

[0048] Secondly, use the relationship among ZTD, ZWD and ZHD to calculate ZWD. Since ZTD is mainly composed of ZHD and ZWD, the relationship among the three can be used to calculate ZWD. The specific calculation formula is as follows:

[0049] (5)

[0050] Where ZTD is the total zenith delay, which is calculated by step 2. ZHD is the zenith static delay, which is calculated by formula 4. In this way, the accurate ZWD can be obtained.

[0051] Finally, the ZWD result calculated above is saved and output. Through the above process, the ZWD parameters can be accurately calculated using Beidou satellite.

[0052] Step 4: Calculate the hourly change rate of ZWD near 12 hours according to different seasons. The specific process is:

[0053] Calculate the hourly change rate of ZWD, and calculate the change rate within the last 12 hours. The specific calculation formula is as follows:

[0054] (6)

[0055] Where, ΔZWD r Indicates the hourly rate of change of ZWD. ZWD p Indicates the ZWD value at the current time, which is also the ZWD value at the predicted time. i Indicates the ZWD value of the hour to be calculated. For example, if the current time is 12 noon, which is the time to be predicted, the ZWD value of this time is expressed as ZWD p The corresponding change rates of the 12 hours approaching are pushed forward 12 hours, for example, the first one is the ZWD value at 11 am, which corresponds to ZWD i At this time, the first rate of change calculated is expressed as ΔZWD r .

[0056] And so on, until the rate of change of nearly 12 hours is calculated. There are 12 values ​​in total, which can be expressed as: ΔZWD1, ..., ΔZWD 12 .

[0057] Finally, the hourly change rate of the ZWD calculated above within the last 12 hours is saved and output. Through the above process, the ZWD and the corresponding hourly change rate value can be obtained.

[0058] Step 5: Use the ZWD value and the corresponding hourly change rate and the corresponding rainfall data to establish the rainfall and rainfall duration prediction threshold. The ZWD value is calculated by step 3, and the corresponding hourly change rate is obtained by step 4. The corresponding rainfall data can be downloaded from the official website of the local meteorological department or from the products provided by commercial companies. Using the above data, the rainfall and rainfall duration prediction threshold can be established.

[0059] Specifically, the method for establishing rainfall prediction is:

[0060] First, set the initial rainfall prediction threshold;

[0061] Then, the rainfall events in the historical data are counted, and the change rate of the ZWD value 12 hours before each rainfall event is calculated to obtain 12 values. When at least one of the 12 values ​​is greater than the initial prediction threshold, it is predicted to be a rain event; if no value is greater than the initial prediction threshold, it is predicted to be a no-rain event; and all the prediction results are recorded, including the prediction of rainfall and the actual rainfall, which are recorded as P ture ; Rainfall is predicted, but there is no rainfall in reality, recorded as P false ; No rainfall is predicted, but there is rainfall in reality, recorded as P miss ;

[0062] The prediction success rate, prediction error rate and prediction loss rate are calculated based on the prediction results. The expressions are as follows:

[0063] (7)

[0064] Adjust the initial rainfall prediction threshold and repeat the above steps;

[0065] The optimal rainfall prediction threshold is selected based on the prediction success rate, prediction error rate and prediction loss rate.

[0066] In the formula, TPR, WPR and MFR represent the prediction accuracy, prediction error rate and prediction miss rate respectively. Among them, the prediction accuracy rate represents the probability that the predicted rainfall actually rains. The prediction error rate represents the probability that the predicted rainfall actually does not rain. The prediction miss rate represents the probability that the predicted rainfall does not rain, but actually rains. In the formula, P ture ,P false , P miss The definition is shown in the following table:

[0067] Table 1. Rainfall prediction P ture ,P false , P miss Parameter Definition

[0068]

[0069] In addition, R total and P total The definitions are expressed as:

[0070] (8)

[0071] Adjust the initial rainfall prediction threshold and repeat the above steps;

[0072] The optimal rainfall prediction threshold is selected based on the prediction success rate, prediction error rate and prediction loss rate.

[0073] In this embodiment, the initial rainfall prediction threshold is set in the range of 1-10.

[0074] Through the above steps, the prediction success rate, prediction error rate and prediction loss rate under each initial rainfall prediction threshold are calculated.

[0075] The prediction error tolerance rate is determined based on historical experience. In this embodiment, the prediction error tolerance rate is set to 20%, that is, the prediction error rate and the prediction loss rate do not exceed 20%. After determining the prediction error tolerance rate, the value with the highest prediction success rate is selected, and then the corresponding rainfall prediction threshold is selected as the optimal rainfall prediction threshold.

[0076] Furthermore, the method for establishing the rainfall duration prediction threshold is:

[0077] First, obtain all rainfall events in the historical data and calculate the ZWD values ​​of all rainfall events.

[0078] Then all ZWD values ​​are classified into intervals, and after the classification is completed, the duration of rain in each category is counted;

[0079] In this embodiment, ZWD is divided into six categories, namely: <230, 230-280, 280-330, 330-380, 380-430, > 430. The above classification is mainly for illustration, and the actual process will be adjusted according to rainfall and season.

[0080] Finally, all rainfall durations under each category are averaged to obtain the average rainfall duration forecast, and then the rainfall duration forecast reference value is determined based on the average rainfall duration forecast, and then the rainfall duration forecast reference value and the corresponding category are combined to form the rainfall duration forecast threshold.

[0081] Specifically, in this embodiment, the rainfall duration prediction threshold can be obtained by statistical analysis of the data from 15 observation stations in Hong Kong for ten years from 2010 to 2020, which is given in Table 2 below:

[0082] Table 2. Prediction thresholds of rainfall duration obtained based on ZWD statistics

[0083]

[0084] It is understandable that the relevant data can be directly obtained. Thus, a large amount of data can be used for statistical analysis to obtain the corresponding rainfall duration prediction threshold, and the relevant threshold is stored in the form of a database on a local computer.

[0085] Finally, the above rainfall and rainfall duration prediction thresholds are saved in the form of a database on a local computer for use in subsequent rainfall and rainfall duration predictions.

[0086] Step 6: Use the ZWD and ZWD hourly change rate, as well as the corresponding rainfall prediction threshold to make rainfall prediction for the forecast day. Figure 2 As shown,

[0087] First, collect the original data of Beidou satellite PPP positioning, calculate ZWD and the corresponding hourly change rate. The ZWD calculation process is the same as step 2 and step 3 above, and will not be repeated here. The corresponding hourly change rate calculation process is the same as step 4, and will not be repeated here.

[0088] Secondly, the ZWD calculated above and the corresponding hourly change rate are combined with the rainfall prediction threshold to predict rainfall. The prediction threshold has been established in step 5. The prediction process is to compare the hourly change rate with the established threshold. If 40% of the change rate exceeds the threshold, light rain is predicted. If 60% of the change rate exceeds the threshold, moderate rain is predicted. If 90% of the change rate exceeds the threshold, heavy rain is predicted.

[0089] Finally, the above rainfall prediction results are saved and output.

[0090] Step 7: Use the ZWD and rainfall forecast results, as well as the corresponding rainfall duration forecast threshold to predict the rainfall duration for the forecast day. Figure 3 As shown:

[0091] First, collect the original data of Beidou satellite PPP positioning, calculate ZWD and the corresponding hourly change rate. The ZWD calculation process is the same as step 2 and step 3 above, and will not be repeated here. The corresponding hourly change rate calculation process is the same as step 4, and will not be repeated here.

[0092] Secondly, the ZWD calculated above and the corresponding hourly change rate are used in combination with the rainfall prediction threshold to predict rainfall. The prediction process is as described in step 6 and will not be repeated here.

[0093] Finally, the ZWD value and the predicted rainfall result are used to predict the rainfall duration in combination with the rainfall duration prediction threshold. The prediction of rainfall duration must be made on the premise that there is rainfall. If there is rainfall, determine which segment the current ZWD belongs to, and then predict the rainfall duration in combination with the rainfall duration prediction threshold in Table 2, and save and output the results.

[0094] Step 8: Save and output the above ZWD-based rainfall and rainfall duration prediction results.

[0095] According to the records of the above embodiments, it can be known that this embodiment solves the problem that the existing atmospheric precipitation method needs to be overly dependent on meteorological parameters, resulting in limited application, and the method based on the total zenith delay has zenith statics delay and introduces errors, resulting in a decrease in prediction accuracy. The problem that the current method cannot predict the duration of rainfall is also solved. The rainfall prediction method based on Beidou ZWD can effectively solve the problem that the method based on PWV depends on meteorological parameters. Considering that the existing satellite navigation and positioning satellite can only predict rainfall but cannot predict the duration of rainfall, the rainfall duration is predicted based on the results of rainfall prediction, taking into account the short-term change rate, long-term rate, maximum value and seasonal fluctuation factors of ZWD, which can solve the problem that the existing method of this type cannot predict the duration of rainfall. In addition, in order to further eliminate the problems such as low ZWD accuracy caused by the long revisit period of Beidou medium orbit satellites and inclined geosynchronous orbit satellites, the weight strategy of increasing the proportion of Beidou GEO satellite solution is adopted, which can further improve the solution accuracy of ZWD and thus improve the prediction accuracy.

[0096] Finally, in order to further improve the temporal and spatial resolution of rainfall and rainfall duration prediction, an adaptive threshold strategy is used to set the prediction threshold, which can effectively overcome the high error rate of the prediction method caused by the influence of seasonal fluctuations and different station locations on rainfall.

[0097] In summary, the purpose of the present invention is to provide a method suitable for Beidou satellite to predict rainfall and rainfall duration, so as to solve the problems of existing methods over-reliance on meteorological parameters, resulting in limited application and inability to predict rainfall duration, and at the same time solve the problems of low prediction accuracy caused by the existing methods ignoring the influence of factors such as long satellite revisit period, seasonal fluctuations in rainfall, and low temporal and spatial resolution, thereby providing strong guarantee for accurate predictions by meteorological departments.

[0098] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the described embodiments. For those skilled in the art, various changes, modifications, substitutions and variations of these embodiments including components are made without departing from the principles and spirit of the present invention, and still fall within the scope of protection of the present invention.

Claims

1. A real-time rainfall and rainfall duration prediction method based on Beidou ZWD, characterized in that: The steps include: Step 1: Extract ZWD values, corresponding rainfall and corresponding time series based on historical observations of BeiDou navigation satellites; Step 2: Calculate the hourly change rate of the ZWD value 12 hours before the rainfall time according to different seasons; Step 3: Use the ZWD value and the corresponding hourly change rate and the corresponding rainfall data to establish the rainfall and rainfall duration prediction thresholds; Step 4: Obtain the observation value of the Beidou navigation satellite at the time of prediction, and extract the ZWD value at the time of prediction, and then output the prediction results of rainfall and rainfall duration based on the established rainfall and rainfall duration prediction thresholds.

2. A method for real-time rainfall and rainfall duration prediction based on Beidou ZWD according to claim 1, characterized in that: The calculation method of the ZWD value is: Firstly, the ZTD is solved by the weighted least squares algorithm and GEO satellite weighting strategy based on the historical observations. Then, the Sa empirical model is used to solve the ZHD; Finally, ZWD is calculated using the relationship among ZTD, ZWD and ZHD. The calculation method is as follows: ； Where ZTD is the total tropospheric zenith delay, ZHD is the dry delay, and ZWD is the wet delay.

3. The method for real-time rainfall and rainfall duration prediction based on Beidou ZWD according to claim 1, characterized in that: The historical observation values ​​include pseudoranges, carrier phase observations, precise ephemeris, precise clock errors and corresponding earth nutation.

4. The method for real-time rainfall and rainfall duration prediction based on Beidou ZWD according to claim 1, characterized in that: In step 3, the method for establishing the rainfall prediction threshold is: First, set the initial rainfall prediction threshold; Then, the rainfall events in the historical data are counted, and the change rate of the ZWD value 12 hours before each rainfall event is calculated to obtain 12 values. When at least one of the 12 values ​​is greater than the initial prediction threshold, it is predicted to be a rain event; if no value is greater than the initial prediction threshold, it is predicted to be a no-rain event; and all the prediction results are recorded, including the prediction of rainfall and the actual rainfall, which are recorded as P ture ; Rainfall is predicted, but there is no rainfall in reality, recorded as P false ; No rainfall is predicted, but there is rainfall in reality, recorded as P miss ; Calculate the prediction success rate, prediction error rate and prediction loss rate based on the prediction results; Adjust the initial rainfall prediction threshold and repeat the above steps; The optimal rainfall prediction threshold is selected based on the prediction success rate, prediction error rate and prediction loss rate.

5. The method for real-time rainfall and rainfall duration prediction based on Beidou ZWD according to claim 4 is characterized in that: The calculation method of the prediction success rate, prediction error rate, and prediction loss rate is: ； In the formula, TPR, WPR and MFR represent the prediction accuracy, prediction error rate, and prediction loss rate respectively, and R total is the sum of actual rainfall events; P total is the total number of predicted rainfall events.

6. A method for real-time rainfall and rainfall duration prediction based on Beidou ZWD according to claim 4, characterized in that: The calculation method of the ZWD value change rate is: ； Where, ΔZWD r Indicates the hourly rate of change of ZWD; ZWD p Indicates the ZWD value at the current time, which is also the ZWD value at the predicted time; ZWD i Indicates that the ZWD value at that hour is to be calculated.

7. The method for real-time rainfall and rainfall duration prediction based on Beidou ZWD according to claim 1, characterized in that: In step 3, the method for establishing the rainfall duration prediction threshold is: First, obtain all rainfall events in the historical data and calculate the ZWD values ​​of all rainfall events. Then all ZWD values ​​are classified into intervals, and after the classification is completed, the duration of rain in each category is counted; The average rainfall duration is obtained by averaging all rainfall durations in each category; Determine the rainfall duration prediction reference value based on the average rainfall duration; Finally, the corresponding rainfall duration prediction reference value of each category is used as a data pair, so that the data pair of each category is used as the rainfall duration prediction threshold.

Images