A precipitation estimation method for S-band dual-polarization radar based on historical optimal coefficients

Abstract

The present invention relates to the field of precipitation estimation technology, and in particular to an S-band dual-polarization radar precipitation estimation method based on historical optimal coefficients, comprising obtaining a combined reflectivity feature code in each processing unit, comparing the combined reflectivity feature code of the processing unit with a plurality of historical combined reflectivity feature codes of the corresponding processing unit in a feature database, selecting the one with the smallest deviation as the combined reflectivity feature code in the feature database that is closest to it, and selecting a fitting coefficient corresponding to the combined reflectivity feature code in the feature database. a,i,j,k, The hourly rainfall intensity is calculated according to the formula. The present invention uses a deviation algorithm to quickly obtain the optimal fitting coefficient of the corresponding processing unit, and realizes rapid adjustment of the fitting coefficient, thereby improving the radar estimation precipitation speed in real-time business.

Description

Technical Field

[0001] The present invention relates to the technical field of precipitation estimation, and in particular to an S-band dual-polarization radar precipitation estimation method based on historical optimal coefficients. Background Art

[0002] Quantitative estimation of precipitation by weather radar is an important part of current short-term services. The precipitation estimation accuracy of the relationship is mainly affected by the unstable changes of coefficients A and b. Domestic and foreign researchers have established a ZI relationship suitable for a specific area and time by using actual precipitation and radar reflectivity data from ground observation stations. Limited by changes in weather types, precipitation particle morphology, terrain height, etc., the spatiotemporal uncertainty characteristics of coefficients A and b are obvious. In addition, the change of the vertical profile of reflectivity is objective, and the error increases with the increase of the horizontal distance of radar detection and the vertical height of the target (Joss et al, 1990), especially in complex underground radars using high elevation angle detection to avoid obstruction by ground objects, which will increase this error (Bourrel et al, 1994). It is not the best solution to use fixed coefficients to estimate precipitation in a certain area in general. Therefore, how to design a reasonable dynamic parameter precipitation estimation solution according to local conditions is a problem that needs to be solved in current business. Summary of the invention

[0003] In view of this, the purpose of the present invention is to propose an S-band dual-polarization radar precipitation estimation method based on historical optimal coefficients to solve the problem that the coefficients in the precipitation estimation formula cannot be quickly adjusted according to different locations.

[0004] Based on the above purpose, the present invention provides an S-band dual-polarization radar precipitation estimation method based on historical optimal coefficients, comprising:

[0005] S1. Patch the S-band dual-polarization radar data in the study area to obtain a patch product, which includes a three-dimensional patch result Z of the three-dimensional reflectivity, differential reflectivity, and differential phase shift rate observations. 3D , ZDR 3D、 KDP 3D， and 2D combined reflectivity data;

[0006] S2. Combine the radar dual polarization contour product and high-precision terrain DEM data, output the lowest height of the radar contour product in each grid point greater than the terrain DEM height value, and find the corresponding effective Z according to the height. 3D , ZDR 3D , KDP 3DThe grid point observation values ​​at different heights are recombined to form a new two-dimensional Z, ZDR, and KDP radar observation value field, and the radar observation values ​​are marked as Z higher-dem , ZDR higher-dem 、KDP higher-dem ;

[0007] S3, splitting the radar observation value field into multiple processing units to reflect the precipitation differences caused by different terrains and rain clusters;

[0008] S4, collect historical data to re-analyze radar puzzle products, and process the observed three-dimensional puzzle results in step S1 and historical data into a complete three-dimensional radar puzzle data Z 3D1 , ZDR 3D1 、KDP 3D1 According to the scheme in step S2, the complete Z 3D1 , ZDR 3D1 、KDP 3D1 Dataset processing output history Z higher-dem1 , ZDR higher-dem1 , KDP higher-dem1 Numerical dataset of two-dimensional radar observations;

[0009] S5. Based on the two-dimensional combined reflectivity data in the processing unit, a combined reflectivity feature code in the processing unit is calculated;

[0010] S6. Establish a fitting scheme between historical automatic station precipitation and radar observation values ​​within the processing unit: ;

[0011] Among them, I 1 For automatic station observation precipitation data, a, i, j, k To be the fitting coefficient, the precipitation data observed by the automatic station covered in the processing unit are selected, and the radar observation value closest to the automatic station is selected by the neighbor method in the interpolation method. The parameter-by-parameter traversal method is used to calculate the above formula with a small step size. a, i, j, k , obtained by calculating MAE a, i, j, k The optimal solution makes the radar-estimated rainfall intensity closest to the automatic station observation precipitation data;

[0012] S7. Establish a historical data feature database for each processing unit within the coverage area of ​​the radar puzzle product: the feature database content includes the processing unit serial number, multiple historical combined reflectivity feature codes in each processing unit, and multiple fitting coefficients corresponding to the multiple historical combined reflectivity feature codes. a, i, j, k , which can be used to quickly obtain the fitting coefficients in the later real-time business operation;

[0013] S8, real-time retrieval of the best fitting solution for the processing unit: repeat the steps in S5 during real-time operation to obtain the combined reflectivity feature code in each processing unit, compare the combined reflectivity feature code of the processing unit with multiple historical combined reflectivity feature codes of the corresponding processing unit in the feature database, select the one with the smallest deviation as the combined reflectivity feature code in the feature database that is closest to it, and select the fitting coefficient corresponding to the combined reflectivity feature code in the feature database a, i, j, k , according to the actual observed Z higher-dem , ZDR higher-dem , KDP higher-dem The hourly rainfall intensity is calculated using the formula: ;

[0014] Among them, I2 is the calculated hourly rainfall intensity.

[0015] Optionally, the step S5 specifically includes:

[0016] Based on the two-dimensional combined reflectivity data MCR in the processing unit, the Histgram method is used to calculate the number of grid points contained in each 10dBZ interval of 0-80dBz within the coverage area of ​​the processing unit.

[0017] Calculate the ratio between the number of grid points distributed in each 10 dBZ interval and the total number of grid points in the processing unit to obtain the percentage of each 10 dBz combined reflectivity data in the processing unit;

[0018] The percentage values ​​are arranged in order as the combined reflectivity feature code of the processing unit.

[0019] Optionally, the deviation calculation formula in step S8 is as follows: ;

[0020] Where i represents the dBz level, which corresponds to the distribution interval number from 0-80dBz by 10dBz. represents the error weight. The error weights corresponding to the 8 intervals are (0.8, 0.85, 0.9, 0.95, 1, 1.1, 1.2, 1.3). Represents the percentage of the real-time combined reflectivity feature code. Represents the percentage of combined reflectance feature codes in the feature database, and n represents the number of intervals.

[0021] When matching the feature codes, the features of the stronger echo are selected as important references, so the error weight is higher (greater than 1 above 40dBz), and the feature impact factor of the weaker echo is lower (less than 1 above 40dBz), which is in line with the working habit of paying more attention to strong echoes in meteorology. In this way, when the features of the stronger echoes differ greatly, the deviation value will be amplified to avoid the selection of the combined reflectivity feature code when the features of the stronger echoes differ greatly. Therefore, the combined reflectivity feature code with similar features to the stronger echoes can be better selected, thereby making the fitting coefficient selected by the combined reflectivity feature code more accurate.

[0022] Optionally, the jigsaw puzzle of the S-band dual-polarization radar data in the study area to obtain the jigsaw puzzle product includes: reading the dual-polarization radar single-station observation data, and realizing the jigsaw puzzle of the S-band dual-polarization radar observations in the study area through the steps of single-station data quality control, coordinate conversion, and multi-station fusion.

[0023] Optionally, in step S2, if the bottom-level observation value is invalid, continue searching up two levels to find a valid Z 3D , ZDR 3D , KDP 3D If there is no valid observation value in the two upper layers, it is considered that there is no valid observation value at this point.

[0024] Optionally, in step S3, the processing unit is artificially defined as 0.1°*0.1°, and one processing unit contains observation data of 10*10 grid points.

[0025] Optionally, in step S6, 10-minute automatic station observation precipitation data is selected and fitted with radar observation data.

[0026] Optionally, linear regression technology is used to further optimize the calculated rainfall intensity results. The 10-minute automatic station observation results are linearly fitted with the radar estimation results to obtain a linear fitting relationship, and the linear fitting relationship is reapplied to the estimated precipitation results to obtain an estimate that is closer to the actual situation.

[0027] Optionally, after calculating the rainfall intensity of each processing unit, the method further includes: dividing the rainfall intensity by 6 to obtain the rainfall intensity every 10 minutes, and reassembling the 10-minute rainfall intensity of each processing unit into a two-dimensional matrix rainfall intensity precipitation field consistent with the coverage of the radar puzzle according to the actual distribution of the cells, using a 9-point mean filtering scheme to filter the entire two-dimensional matrix rainfall intensity precipitation field, and after filtering, the final radar 10-minute rainfall intensity precipitation field based on the radar puzzle data is obtained, and then the 10-minute rainfall intensity is accumulated with the previous calculation results to obtain the hourly rainfall intensity.

[0028] Beneficial effects:

[0029] Compared with the traditional radar precipitation estimation scheme, the present invention solves the single radar observation blind spot problem and the standardized output product is convenient for later calculation based on the S-band dual-polarization radar puzzle precipitation estimation scheme; referring to the dual-polarization quantity precipitation estimation relationship, it is proposed to increase the observation information ZDR and KDP, and by splitting the processing unit, the processing range is reduced, and the radar estimation precipitation accuracy is improved; through steps S5-S8, the calculation of the combined reflectivity feature code in the processing unit and the matching scheme of the combined reflectivity feature code and the fitting coefficient are proposed to match the historical fitting relationship, and according to the processing units in different geographical locations, the optimal fitting coefficient of the corresponding processing unit is quickly obtained through the deviation algorithm, and the rapid adjustment of the fitting coefficient is realized, thereby improving the radar estimation precipitation speed in real-time business. Finally, a complete set of S-band dual-polarization radar precipitation estimation business schemes based on historical optimal parameters for real-time business operation is proposed, which provides a basis for forecasters to accurately evaluate the precipitation area. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the present invention or 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 in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0031] Figure 1 A schematic diagram of the relationship between the radar mosaic coverage and a single processing unit according to an embodiment of the present invention;

[0032] Figure 2 An example of the MCR feature calculation method in a processing unit based on the Histgram technology is given. DETAILED DESCRIPTION

[0033] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0034] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0035] A method for estimating precipitation using an S-band dual-polarization radar based on historical optimal coefficients, comprising:

[0036] S1. Patch the S-band dual-polarization radar data in the study area to obtain the patch product, which includes three-dimensional reflectivity, differential reflectivity, differential phase shift rate (hereinafter abbreviated as Z, ZDR, KDP, respectively). 3D , ZDR 3D , KDP 3D， and two-dimensional combined reflectivity data MCR.

[0037] Optionally, by reading the dual-polarization radar single-station observation data and going through the steps of single-station data quality control, coordinate conversion, and multi-station fusion, the S-band dual-polarization radar observation puzzle in the study area can be realized.

[0038] The advantages of the puzzle products obtained through S-band dual-polarization radar data include: 1) Solving the blind spot problem of single radar observation. Compared with single-band radar, dual-polarization radar obtains additional factors such as polarization, differential reflectivity, and differential phase shift rate by adding vertical beam detection. ZDR and KDP are less affected by changes in the spectrum of precipitation raindrops and can better invert precipitation and liquid water content. Therefore, it is more advantageous to use different combinations of Z, ZDR, and KDP to invert precipitation. 2) Processing to form radar contour products is convenient for later calculations. The vertical resolution of the three-dimensional puzzle product is 200m below 3KM, 500 meters below 3-9KM, 1000m between 9KM and 20KM, the horizontal resolution is 0.01°, and the time resolution is 6 minutes.

[0039] S2. Combine the radar dual polarization contour product and high-precision terrain DEM data, output the lowest height of the radar contour product in each grid point greater than the terrain DEM height value in the longitude and latitude direction according to the puzzle range, and find the corresponding effective Z according to the height. 3D , ZDR 3D , KDP 3D The above grid point observation values ​​at different heights are recombined to form a new two-dimensional Z, ZDR, and KDP radar observation value field. The above radar observation values ​​are marked as Z higher-dem , ZDR higher-dem , KDP higher-dem .

[0040] S3, splitting the radar observation value field into multiple processing units (such as Figure 1 ), used to reflect the differences in precipitation caused by different terrains and rain clusters

[0041] S4. Collect historical data (more than 3 years) to re-analyze radar mosaic products, and process the observed 3D mosaic results in step S1 with historical data into a complete 3D radar mosaic data Z. 3D1 , ZDR 3D1 , KDP 3D1 According to the scheme in step S2, the complete Z 3D1 , ZDR 3D1 , KDP 3D1 Dataset processing output history Z higher-dem1 , ZDR higher-dem1 , KDP higher-dem1 Numerical dataset of 2D radar observations.

[0042] S5. Based on the two-dimensional combined reflectivity data MCR in the processing unit, a combined reflectivity feature code in the processing unit is calculated. The combined reflectivity feature code is used to indicate the intensity of the reflectivity and its corresponding coverage size.

[0043] S6. Establish a fitting scheme between historical automatic station precipitation and radar observation values ​​within the processing unit: ;

[0044] Among them, I 1 For automatic station observation precipitation data, a, i, j, k To be the fitting coefficient, the precipitation data observed by the automatic station covered in the processing unit are selected, and the radar observation value closest to the automatic station is selected by the neighbor method in the interpolation method. The parameter-by-parameter traversal method is used to calculate the above formula with a small step size. a, i, j, k , obtained by calculating MAE a, i, j, kThe optimal solution makes the radar-estimated rainfall intensity closest to the automatic station observation precipitation data.

[0045] S7. Establish a historical data feature database for each processing unit within the coverage area of ​​the radar puzzle product: The feature database content includes the processing unit serial number (used to mark the position of the unit in the two-dimensional field), multiple historical MCR combined reflectivity feature codes in each processing unit, and multiple fitting coefficients corresponding to the multiple historical combined reflectivity feature codes. a, i, j, k , which can be used to quickly obtain the fitting coefficients during later real-time business operations.

[0046] Its structure is shown in the following table:

[0047] ;

[0048] S8, real-time retrieval of the best fitting solution for the processing unit: repeat the steps in S5 during real-time operation to obtain the combined reflectivity feature code in each processing unit, compare the combined reflectivity feature code of the processing unit with the combined reflectivity feature code of the corresponding processing unit in the feature database, select the one with the smallest deviation as the combined reflectivity feature code in the feature database that is closest to it, and select the fitting coefficient corresponding to the combined reflectivity feature code in the feature database a, i, j, k , according to the actual observed Z higher-dem , ZDR higher-dem , KDP higher-dem The hourly rainfall intensity is calculated using the formula: ;

[0049] Among them, I2 is the calculated hourly rainfall intensity, and OHP / 6 can be used to get the corresponding 10-minute rainfall intensity.

[0050] As can be seen from the above, the present invention solves the single radar observation blind spot problem and standardized output products for later calculations based on the S-band dual-polarization radar puzzle estimation precipitation scheme compared with the traditional radar estimation precipitation scheme; referring to the dual polarization quantity estimation precipitation relationship, it is proposed to increase the observation information ZDR and KDP, and by splitting the processing unit, the processing range is reduced, and the radar estimation precipitation accuracy is improved; through steps S5-S8, a one-to-one matching scheme of the combined reflectivity feature code calculation and the combined reflectivity feature code and the fitting coefficient in the processing unit is proposed, and the matching scheme is more refined, and a better matching result can be obtained. By matching the historical fitting relationship, according to the processing units in different geographical locations, the algorithm of deviation is used to quickly obtain the optimal fitting coefficient of the corresponding processing unit, and the rapid adjustment of the fitting coefficient is realized, thereby improving the radar estimation precipitation speed in real-time business. Finally, a complete set of S-band dual-polarization radar precipitation estimation business schemes based on historical optimal parameters for real-time business operation is proposed, which provides a basis for forecasters to accurately assess the precipitation area.

[0051] In some embodiments, the step S5 specifically includes:

[0052] Based on the two-dimensional combined reflectivity data MCR in the processing unit, the Histgram method is used to calculate the number of grid points contained in each 10dBZ interval of 0-80dBz in the combined reflectivity data (MCR) within the coverage of the processing unit. Each 10dBZ interval is 0dBz-10dBz, 10dBz-20dBz, 20dBz-30dBz, 30dBz-40dBz, 40dBz-50dBz, 50dBz-60dBz, 60dBz-70dBz, 70dBz-80dBz.

[0053] The ratio between the number of grid points distributed in each 10 dBZ interval and the total number of grid points in the processing unit is calculated to obtain the percentage of each 10 dBz combined reflectivity data in the processing unit.

[0054] The percentage values ​​are arranged in order as the combined reflectivity feature code of the processing unit. Figure 2 The feature code in the example can be compiled as [14,26,30,19,11,10,0,0].

[0055] In some embodiments, the deviation calculation formula in step S8 is as follows: ;

[0056] Where i represents the dBz level, which corresponds to the distribution interval number from 0-80dBz by 10dBz. Represents the error weight, and the error weights corresponding to the 8 intervals are (0.8, 0.85, 0.9, 0.95, 1, 1.1, 1.2, 1.3). The significance of designing this weight is that when matching the feature code, the feature of the stronger echo is selected as an important reference, so the error weight is higher (greater than 1 above 40dBz), and the feature impact factor of the weaker echo is lower (less than 1 above 40dBz), which is in line with the working habit of paying more attention to strong echoes in meteorology. In this way, when the features of the stronger echo differ greatly, the deviation value will be amplified, avoiding the selection of the combined reflectivity feature code when the features of the stronger echo differ greatly. Therefore, the combined reflectivity feature code with similar features of the stronger echo can be better selected, so that the fitting coefficient selected by the combined reflectivity feature code is more accurate. Represents the percentage of the real-time combined reflectivity feature code. represents the percentage of the combined reflectivity feature code in the feature database, n represents the number of intervals, and the root deviation calculation formula is used to calculate the combined reflectivity feature code in the feature database that is closest to the actual situation in the processing unit. In some embodiments, in step S2, if the bottom-level observation value is invalid, continue to search up two levels to find a valid Z 3D , ZDR 3D , KDP 3D If there is no valid observation value in the upper two layers, it is considered that there is no valid observation value at this point. The data obtained in this way is the valid value closest to the ground observation. If there is no valid value near the ground, it means that there may be no precipitation at the grid point, and the upper echo is caused by high-altitude clouds or non-meteorological echoes.

[0057] In actual work, referring to the small-scale range of the weather system, the processing unit in step S3 is artificially defined as 0.1°*0.1° (approximately 10 km*10 km), and one processing unit contains observation data of 10*10 grid points.

[0058] In some embodiments, because radar observation is an instantaneous result, the shorter the automatic station observation data is, the better the correspondence with the radar observation result is. Therefore, in step S6, 10 minutes of automatic station observation precipitation data is selected to fit the radar observation data.

[0059] In some embodiments, a linear regression technique is used to further optimize the calculated rainfall intensity results. The 10-minute automatic station observation results are linearly fitted with the radar estimation results to obtain a linear fitting relationship. The linear fitting relationship is reapplied to the estimated precipitation results to obtain an estimated value that is closer to the actual situation.

[0060] After calculating the rainfall intensity of each processing unit, the method further includes: dividing the rainfall intensity by 6 to obtain the rainfall intensity every 10 minutes, and reassembling the 10-minute rainfall intensity of each processing unit according to the actual distribution of the cells into a two-dimensional matrix rainfall intensity precipitation field consistent with the coverage of the radar puzzle, taking into account the excessive problem caused by inconsistent calculation results at the boundaries of different processing cells. In order to solve this problem, a 9-point mean filtering scheme is used to filter the entire two-dimensional matrix rainfall intensity precipitation field. After filtering, the final radar 10-minute rainfall intensity precipitation field based on the radar puzzle data is obtained, and then the 10-minute rainfall intensity is accumulated with the previous calculation results to obtain a more accurate hourly rainfall intensity precipitation data (OHP).

[0061] It should be understood by those skilled in the art that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present invention (including the claims) is limited to these examples. Under the concept of the present invention, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the above aspects of the present invention, which are not provided in detail for the sake of simplicity.

[0062] The present invention is intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for estimating precipitation using an S-band dual-polarization radar based on historical optimal coefficients, characterized in that: include: S1. Patch the S-band dual-polarization radar data in the study area to obtain a patch product, which includes a three-dimensional patch result Z of the three-dimensional reflectivity, differential reflectivity, and differential phase shift rate observations. 3D , ZDR 3D 、KDP 3D， and 2D combined reflectivity data; S2. Combine the radar dual polarization contour product and high-precision terrain DEM data, output the lowest height of the radar contour product in each grid point greater than the terrain DEM height value, and find the corresponding effective Z according to the height. 3D , ZDR 3D 、KDP 3D The grid observation values ​​at different heights are recombined to form a new two-dimensional Z, ZDR, and KDP radar observation value field, and the above radar observation values ​​are marked as Z higher-dem , ZDR higher-dem 、KDP higher-dem ; S3, splitting the radar observation value field into multiple processing units to reflect the precipitation differences caused by different terrains and rain clusters; S4, collect historical data to re-analyze radar puzzle products, and process the observed three-dimensional puzzle results in step S1 and historical data into a complete three-dimensional radar puzzle data Z 3D1 , ZDR 3D1 、KDP 3D1 According to the scheme in step S2, the complete Z 3D1 , ZDR 3D1 、KDP 3D1 Dataset processing output history Z higher-dem1 , ZDR higher-dem1 、KDP higher-dem1 Numerical dataset of two-dimensional radar observations; S5. Based on the two-dimensional combined reflectivity data in the processing unit, a combined reflectivity feature code in the processing unit is calculated; S6. Establish a fitting scheme between historical automatic station precipitation and radar observation values ​​within the processing unit: ； Among them, I 1 For automatic station observation precipitation data, a,i,j,k To be the fitting coefficient, the precipitation data observed by the automatic station covered in the processing unit are selected, and the radar observation value closest to the automatic station is selected by the neighbor method in the interpolation method. The parameter-by-parameter traversal method is used to calculate the above formula with a small step size. a,i,j,k , obtained by calculating MAE a,i,j,k The optimal solution makes the radar-estimated rainfall intensity closest to the automatic station observation precipitation data; S7. Establish a historical data feature database for each processing unit within the coverage area of ​​the radar puzzle product: the feature database includes the processing unit serial number, multiple historical combined reflectivity feature codes in each processing unit, and multiple fitting coefficients corresponding to the multiple historical combined reflectivity feature codes. a,i,j,k , which can be used to quickly obtain the fitting coefficients in the later real-time business operation; S8, real-time retrieval of the best fitting solution for the processing unit: repeat the steps in S5 during real-time operation to obtain the combined reflectivity feature code in each processing unit, compare the combined reflectivity feature code of the processing unit with multiple historical combined reflectivity feature codes of the corresponding processing unit in the feature database, select the one with the smallest deviation as the combined reflectivity feature code in the feature database that is closest to it, and select the fitting coefficient corresponding to the combined reflectivity feature code in the feature database a,i,j,k , according to the actual observed Z higher-dem , ZDR higher-dem 、KDP higher-dem The hourly rainfall intensity is calculated using the formula: ； Among them, I2 is the calculated hourly rainfall intensity.

2. The S-band dual-polarization radar precipitation estimation method based on historical optimal coefficients according to claim 1 is characterized in that: The S5 step specifically includes: Based on the two-dimensional combined reflectivity data MCR in the processing unit, the Histgram method is used to calculate the number of grid points contained in each 10dBZ interval of 0-80dBz within the coverage area of ​​the processing unit. Calculate the ratio between the number of grid points distributed in each 10 dBZ interval and the total number of grid points in the processing unit to obtain the percentage of each 10 dBz combined reflectivity data in the processing unit; The percentage values ​​are arranged in order as the combined reflectivity feature code of the processing unit.

3. The S-band dual-polarization radar precipitation estimation method based on historical optimal coefficients according to claim 1, characterized in that: The deviation calculation formula in step S8 is as follows: ； Where i represents the dBz level, corresponding to the distribution interval number from 0-80dBz by 10dBz, W i represents the error weight. The error weights corresponding to the 8 intervals are (0.8, 0.85, 0.9, 0.95, 1, 1.1, 1.2, 1.3). Represents the percentage of the real-time combined reflectivity feature code. Represents the percentage of combined reflectance feature codes in the feature database, and n represents the number of intervals.

4. The S-band dual-polarization radar precipitation estimation method based on historical optimal coefficients according to claim 1, characterized in that: The jigsaw puzzle of the S-band dual-polarization radar data in the study area to obtain the jigsaw puzzle product includes: reading the dual-polarization radar single-station observation data, and implementing the jigsaw puzzle of the S-band dual-polarization radar observations in the study area through the steps of single-station data quality control, coordinate conversion, and multi-station fusion.

5. The S-band dual-polarization radar precipitation estimation method based on historical optimal coefficients according to claim 1, characterized in that: In the step S2, if the observation value at the bottom layer of the grid point is invalid, continue to search up two layers to find a valid Z 3D , ZDR 3D 、KDP 3D If there is no valid observation value in the two upper layers, it is considered that there is no valid observation value at this grid point.

6. The S-band dual-polarization radar precipitation estimation method based on historical optimal coefficients according to claim 1, characterized in that: In the step S3, the size of the processing unit is set to 0.1°*0.1°, and one processing unit contains observation data of 10*10 grid points.

7. The S-band dual-polarization radar precipitation estimation method based on historical optimal coefficients according to claim 1, characterized in that: In the step S6, 10-minute automatic station observation precipitation data is selected and fitted with radar observation data.

8. The S-band dual-polarization radar precipitation estimation method based on historical optimal coefficients according to claim 1, characterized in that: The method further comprises: Linear regression technology is used to further optimize the calculated rainfall intensity results. The 10-minute automatic station observation results are linearly fitted with the radar estimation results to obtain a linear fitting relationship. The linear fitting relationship is reapplied to the estimated precipitation results to obtain an estimate that is closer to the actual situation.

9. The S-band dual-polarization radar precipitation estimation method based on historical optimal coefficients according to claim 1, characterized in that: After calculating and obtaining the rainfall intensity of each processing unit, the method further comprises: The rainfall intensity is divided by 6 to obtain the rainfall intensity every 10 minutes. The 10-minute rainfall intensity of each processing unit is reassembled into a two-dimensional matrix rainfall intensity precipitation field consistent with the coverage of the radar puzzle according to the actual distribution of the cells. The 9-point mean filtering scheme is used to filter the entire two-dimensional matrix rainfall intensity precipitation field. After filtering, the final radar 10-minute rainfall intensity precipitation field based on the radar puzzle data is obtained. The 10-minute rainfall intensity is then added to the previous calculation results to obtain the hourly rainfall intensity.

Images