Static meteorological satellite convection primary product evaluation method based on Doppler weather radar

Abstract

The invention discloses a Doppler weather radar-based static meteorological satellite convection primary product evaluation method, relates to the technical field of meteorological satellite convection primary product evaluation, and solves the problems that the existing method does not consider the splitting and merging conditions of cloud clusters, cannot invert and calculate the satellite cloud wind guide when there is less cloud or no cloud, and is inconvenient to use. And the convection primary product cannot be evaluated. According to the method, a convection primary event is extracted from three-dimensional weather radar data to serve as a real condition, a static meteorological satellite convection primary product is evaluated, and the method comprises the following steps of: identifying the convection primary event in the excessive Puler weather radar by taking a reflectivity value greater than or equal to 35dBz as a standard; radar convection inception is taken as a real condition, a neighborhood method oriented to a convection inception area is adopted, the hit rate, the empty report rate, the TS test index and the time advance of satellite convection inception are calculated, and a geostationary satellite convection inception product is objectively evaluated from the time-space perspective, so that the method can be better applied to early warning of severe convection weather.

Description

Technical Field

[0001] The invention relates to the technical field of satellite convective primary product evaluation, and in particular to a geostationary satellite convective primary product evaluation method based on Doppler radar. Background Art

[0002] Severe convective weather is a sudden, highly destructive, localized disastrous weather that can easily cause mass casualties and is one of the most important forecast contents in weather forecasting. Timely monitoring and accurate forecasting of severe convective weather can reduce the huge losses it brings. The onset of convection is a sign that severe convective weather has begun to be active. Accurately detecting the onset of convection is the key to improving the level of local severe convective weather warnings. Due to the wide coverage and high spatial and temporal resolution of satellite data, the onset of convection can be detected earlier than radar, greatly improving the effectiveness of warnings and helping to improve the accuracy and timeliness of forecasts of the onset of convection.

[0003] At present, the inspection and evaluation of the primary products of geostationary satellite convection are still mainly subjective, lacking an objective and comprehensive evaluation method, and cannot meet the actual application needs;

[0004] Due to the suddenness and rapid changes of the initiation of convection and the discontinuity of satellite data, the current evaluation of the initiation of convection products mainly remains in subjective inspection, lacking a feasible, recognized and easy-to-operate objective method. At present, for several individual cases, radar data is used to obtain the actual situation of the initiation of convection, and then the area where the initiation of convection may occur is inferred based on the direction of the cloud-guided wind inverted by satellite. If there is a satellite initiation of convection product in this area, it is judged to be correct;

[0005] This method does not take into account the splitting and merging of clouds, and when there are few or no clouds, the satellite cloud wind cannot be inverted and calculated, and due to lack of data, it is impossible to evaluate the primary convective products. Therefore, it does not meet the existing needs. In this regard, we propose a method for evaluating the primary convective products of geostationary satellites based on Doppler radar. Summary of the invention

[0006] The purpose of the present invention is to provide a method for evaluating primary convective products of geostationary satellites based on Doppler radar, so as to solve the problem that the existing methods proposed in the above background technology do not take into account the splitting and merging of cloud clusters, and when there are few clouds or no clouds, the satellite cloud wind guide cannot be inverted and calculated, and there is a lack of data, so the primary convective products cannot be evaluated.

[0007] To achieve the above object, the present invention provides the following technical solution: a method for evaluating primary products of geostationary satellite convection based on Doppler radar, comprising the following steps:

[0008] Step A: Identify cumulus clouds in Doppler radar data: Obtain high-temporal and spatial resolution radar composite reflectivity data after network quality control from the weather radar mosaic system, and then identify and mark cumulus cloud areas from the radar composite reflectivity data. Cumulus cloud areas are convective areas.

[0009] Step B: Identify radar convection primary events using the area overlap method: The grid points corresponding to the marked convection areas in the preprocessed radar reflectivity data are used as an eight-connected area as a research object. Based on the target-oriented idea, a convection area is used as a sample;

[0010] Step C: Preprocessing of satellite convective primary products: Use Python to extract convective primary product data, perform geometric position correction, data quality control, and finally generate binary data to construct a satellite convective primary data set;

[0011] Step D: Using the neighborhood method, based on the idea of ​​oriented to the target area, calculate the correct forecast, false alarm, and missed alarm events of the satellite convective primary events;

[0012] Step E: Calculate the verification index and time advance of convection initiation: Based on satellite convection initiation products and radar convection initiation events, take radar convection initiation as the actual situation, and use the neighborhood method facing the convection initiation area to calculate the satellite convection initiation hit rate POD, false alarm rate FAR, verification index and time advance TS. The formula is as follows:

[0013]

[0014] Hit is the number of convective initiation areas correctly predicted by the satellite, False is the number of convective initiation areas falsely reported by the satellite, and Failure is the number of convective initiation areas missed by the satellite.

[0015] Preferably, identifying and marking cumulus cloud areas in radar combined reflectivity data comprises the following steps:

[0016] S1: interpolation: using the average method to interpolate the radar combined reflectivity data into longitude and latitude grid data with a resolution of 0.04°×0.04°;

[0017] S2: Perform 0 / 1 conversion: 35 dBz is used as the discrimination threshold of convective intensity. The area with reflectivity value ≥ 35 dBz is the convective area and is replaced by the value 1. The area with reflectivity value < 35 dBz is non-cumulus and is converted to the value 0.

[0018] S3: Perform data quality control to avoid interference from isolated noise and strong echoes. Before identifying the incipient convection, filter out cumulus areas with smaller areas and remove identified cumulus areas with ≤3 pixels.

[0019] Preferably, identifying radar convection incipient events using the area overlap method comprises the following steps:

[0020] E1: Read the radar reflectivity data of the current time and the previous time, and use the neighborhood method to identify the convection areas of the two time periods, that is, the area with a mark value of 1, which is composed of convection grid points connected in eight directions;

[0021] E2: Use the area overlap method to track the cloud clusters, remove the convective events that are generated by the translation or development of the cloud clusters at the previous moment, and only extract the convective events generated at the current moment, which is the convective initiation, and set it as a radar convective initiation live sample;

[0022] E3: Constructing radar convection primary database:

[0023] Z_RADA_C_BABJ_YYYYMMDDHHmmss_P_DOR_ACHN_CREF_YYYYMMDD_HHmmss.da t is a dat format file for data naming, which provides a real-time data set for the subsequent evaluation of satellite convection primary products.

[0024] Preferably, the data naming format is:

[0025] YYYYMMDDhhmmss: product generation time, year, month, day, hour, minute, and second, all in universal time;

[0026] ACHN: area code, the national area is ACHN;

[0027] CREF: abbreviation of radar product type, CREF is radar combination reflectivity;

[0028] YYYYMMDD_hhmmss: data observation time, year, month, day, hour, minute, second;

[0029] dat: file extension.

[0030] Preferably, the geometric position correction comprises the following steps:

[0031] N1: According to the row and column number and longitude and latitude lookup table, if the longitude and latitude data of FY-4B is not integrated into the L1 data set, it is necessary to download the original longitude and latitude file from the website of the National Satellite Meteorological Center;

[0032] N2: For different resolutions corresponding to different longitude and latitude files, the longitude and latitude file corresponding to the 0.04°×0.04° resolution is FY4B-_DISK_1050E_GEO_NOM_LUT_20240227000000_4000M_V0001.raw, with a grid size of 2748×2748. This file is a two-dimensional image file with geographic location coordinate information, which corresponds one-to-one to the pixels of the corrected image.

[0033] N3: The data file in N2 is first filled in from north to south by row or from west to east. Each data stores the corresponding latitude and longitude; each pixel finds the corresponding longitude and latitude information according to the row number and column number in the image to achieve precise positioning.

[0034] Preferably, the data quality control is specifically to eliminate convective incipient events with ≤3 pixels.

[0035] Preferably, calculating the correct forecast, false forecast, and missed forecast of satellite convective primary events comprises the following steps:

[0036] M1: The time resolution of the two is matched by the proximity principle. Based on the satellite convective primary product, the convective primary established by using radar data is taken as the actual situation, and compared with the radar convective primary product of each time before and after 2 hours, and the time advance of the satellite convective primary product is calculated;

[0037] M2: Based on the goal-oriented idea, a convection initiation area is taken as a sample to determine the correct forecast, missed forecast and false alarm, and the false alarm result is finally generated;

[0038] M3: Determine the advance time of satellite convection initiation: within the forecast period, the time difference between the correct forecast of satellite convection initiation and the initial satellite convection initiation is the forecast time advance, and the subsequent times do not need to be checked.

[0039] Preferably, the method for determining correct forecasts, missed forecasts and false alarms is as follows: when the satellite convection primary region overlaps with the radar convection primary region or the number of adjacent grid points is less than 30, it is considered a correct forecast; when the satellite convection primary region is not adjacent to the radar convection primary region at all times, it is considered a missed forecast; when the satellite convection primary region is not adjacent to all radar convection primary regions at all times within the time period, it is considered a false alarm.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] The present invention first extracts the primary convective events from the three-dimensional weather radar data as the actual situation, and then uses the area overlap method and the neighborhood method to objectively evaluate the geostationary satellite primary convective products from the perspective of time and space by calculating the accuracy rate, false alarm rate, missed alarm rate and other indexes, so as to better apply them to the early warning of severe convective weather;

[0042] Using higher-resolution Doppler radar data of convective initiation as the actual situation, the data has high quality and strong continuity, and is not affected by cloudy or cloudless weather. Based on the idea of ​​​​target area, the area overlap method and the neighborhood method are used to calculate the number of correct satellite convective initiation, false reports, and missed protection, and then the inspection index and time advance are obtained, and a set of objective and feasible evaluation methods for satellite convective initiation products are proposed. The invention of this evaluation method provides a reference for improving the accuracy of satellite convective initiation products, has a certain supporting role, and also provides a reference for it in the actual severe convective weather forecast business. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a schematic diagram of the process of the convection primary product evaluation method of the present invention;

[0044] Figure 2 It is a schematic diagram for explaining the distribution of parameters in the test formula of the present invention. DETAILED DESCRIPTION

[0045] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0046] See also Figure 1 and Figure 2 The present invention provides an embodiment: a method for evaluating primary products of geostationary satellite convection based on Doppler radar, comprising the following steps:

[0047] Step A: Identify cumulus clouds in Doppler radar data: Obtain high-temporal and spatial resolution radar composite reflectivity data after network quality control from the weather radar mosaic system;

[0048] The weather radar mosaic system is a new generation weather radar mosaic system V3.0, which obtains high temporal and spatial resolution radar combination reflectivity data after network quality control:

[0049] The format is bin format, the dimension is 4200×6200, the latitude range is 12.2-54.2°N, the longitude range is 73-135°E, and the spatial resolution is 1KM. The product file format is: file header + data block. The file header defines the file volume label (file fixed identifier, file format version code, file byte count), product description (puzzle product number, coordinate type, product code, product description, product data starting position, product data byte count), data time (data clock, observation time year, month, day, etc.), data area information (south, west, north, east boundaries of the data area, etc.);

[0050] Then, cumulus areas are identified and marked from the radar combined reflectivity data. Cumulus areas are convective areas. Interpolation is first performed: the radar combined reflectivity data are interpolated into latitude and longitude grid data with a resolution of 0.04°×0.04° using the average method, and then 0 / 1 conversion is performed: 35dBz is used as the discrimination threshold for convective intensity, and areas with reflectivity values ​​≥35dBz are convective areas, which are replaced with a value of 1, and areas with reflectivity values ​​<35dBz are non-cumulus areas, which are converted to a value of 0; finally, data quality control is performed to avoid interference from isolated noise strong echoes. Before identifying the incipient convection, cumulus areas with smaller areas are filtered out, and cumulus areas that have been identified with ≤3 pixels are eliminated;

[0051] Step B: Use the area overlap method to identify radar convection initiation events: first read the radar reflectivity data of the current time and the previous time, and use the neighborhood method to identify the convection areas of the two time periods, that is, the area with a mark value of 1, which is composed of convection grid points connected in eight directions;

[0052] Then, the area overlap method is used to track the cloud clusters, and the convective events resulting from the translation or development of the cloud clusters at the previous moment are eliminated. Only the convective events generated at the current moment are extracted, which are the convective initiation and set as a radar convective initiation live sample.

[0053] Construct radar convection primary database:

[0054] Z_RADA_C_BABJ_YYYYMMDDHHmmss_P_DOR_ACHN_CREF_YYYYMMDD_HHmmss.da t is a dat format file. The data naming format is:

[0055] YYYYMMDDhhmmss: product generation time, year, month, day, hour, minute, and second, all in universal time;

[0056] ACHN: area code, the national area is ACHN;

[0057] CREF: abbreviation of radar product type, CREF is radar combination reflectivity;

[0058] YYYYMMDD_hhmmss: data observation time, year, month, day, hour, minute, second;

[0059] dat: file extension;

[0060] Provide real-time data sets for subsequent evaluation of primary satellite convection products.

[0061] The eight-connected region composed of grid points corresponding to the marked convection area in the preprocessed radar reflectivity data is taken as a research object. Based on the target-oriented idea, a convection area is taken as a sample.

[0062] Step C: Preprocessing of satellite convective primary products:

[0063] The convective primary products of the Fengyun-4B geostationary satellite (FY-4B) were evaluated. The FY-4B satellite is the first weather forecast satellite of my country's new generation of geostationary meteorological satellites, the Fengyun-4 series. It was successfully launched on April 11, 2022, and was fixed to a geostationary orbit position above the equator at 133° east longitude. On June 1, it entered the weather forecast business trial operation and began to provide observation data and application services to global users. From February 1 to March 5, 2024, it will drift from 133 degrees east longitude to 105 degrees east longitude, replacing the Fengyun-4A satellite to realize the replacement of the main business position observation system. From March 5, business services will be resumed at 105 degrees east longitude.

[0064] The FY-4B satellite geostationary orbit radiation imager has added a water vapor detection channel and adjusted the spectra of some channels, which has improved the level of refined observation. The design of the geostationary orbit interferometric infrared detector has been optimized, and the spatial resolution has been further improved, which can provide more accurate high-spectral atmospheric radiation and temperature and humidity profiles and other products; a new rapid imager has been added, which has the ability to rapidly image with a spatial resolution of up to 250 meters within the regional range, making the monitoring of typhoons, rainstorms and mesoscale disastrous weather more continuous, flexible and precise.

[0065] The FY-4B convective primary product has a temporal resolution of 15 minutes and a spatial resolution of 0.04°×0.04°. It is a secondary product developed using the threshold method and obtained after quality inspection, geolocation, and radiometric calibration. The full disk observable longitude range is 23.82-186.18 and the latitude range is -80.883S-80.883N. It is stored in HDF format. The convective primary product is the data of the full disk nominal projection method, with a dimension of 2748×2748, and valid data is 0 and 1, 1 indicates the convective primary.

[0066] Use Python to extract the primary convection product data, perform geometric position correction, data quality control, and finally generate binary data to construct the satellite primary convection data set;

[0067] The geometric position correction is based on the row and column numbers and longitude and latitude lookup table. If the longitude and latitude data of FY-4B are not integrated into the L1 data set, it is necessary to download the original longitude and latitude file from the website of the National Satellite Meteorological Center. Different resolutions correspond to different longitude and latitude files. The longitude and latitude file corresponding to the resolution of 0.04°×0.04° is FY4B-_DISK_1050E_GEO_NOM_LUT_20240227000000_4000M_V0001.raw, with a grid size of 2748×2748. This file is a two-dimensional image file with geographic location coordinate information, which corresponds to the pixels of the correction image one by one. Then fill in the data file in N2 from north to south by row or from west to east. Each data stores the corresponding latitude and longitude; each pixel finds the corresponding longitude and latitude information according to the row and column number in the image to achieve precise positioning.

[0068] Data quality control: Eliminate convection incipient events with ≤3 pixels.

[0069] Step D: Using the neighborhood method, based on the idea of ​​oriented to the target area, calculate the correct forecast, false alarm, and missed alarm events of the satellite convective primary events;

[0070] Due to the different temporal and spatial resolutions of satellite and radar observations, the two sets of product data have been processed to the same spatial resolution of 0.04°×0.04° in longitude and latitude. The temporal resolutions of the two are matched using the proximity principle, as shown in the following table:

[0071] FY4 satellite observation time in China (minutes) 00 15 30 45 Radar observation time (minutes) 00 12 30 42

[0072] Based on the satellite convective primary product, the convective primary established by using radar data is taken as the actual situation, and compared with the radar convective primary product of each time before and after 2 hours, and the time advance of the satellite convective primary product is calculated;

[0073] Based on the goal-oriented idea, a convection initiation area is taken as a sample to determine the correct forecast, missed report and false report. When the satellite convection initiation area overlaps with the radar convection initiation area or the adjacent grid points are less than 30, it is considered to be a correct forecast; when the satellite convection initiation area is not adjacent to the radar convection initiation area at all times, it is considered to be a missed report; when the satellite convection initiation area is not adjacent to all radar convection initiation areas at all times within the period, it is considered to be a false report, and the false report result is generated at the end;

[0074] Determine the advance time of satellite convection initiation: within the forecast period, the time difference between the correct forecast of satellite convection initiation and the initial satellite convection initiation is the forecast time advance, and the subsequent times do not need to be checked;

[0075] Step E: Calculate the verification index and time advance of convection initiation: Based on satellite convection initiation products and radar convection initiation events, take radar convection initiation as the actual situation, and use the neighborhood method facing the convection initiation area to calculate the satellite convection initiation hit rate POD, false alarm rate FAR, verification index and time advance TS. The formula is as follows:

[0076]

[0077] Hit is the number of convective initiation areas correctly predicted by the satellite, False is the number of convective initiation areas falsely reported by the satellite, and Failure is the number of convective initiation areas missed by the satellite.

[0078] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A method for evaluating primary products of geostationary satellite convection based on Doppler radar, characterized in that: The steps include: Step A: Identify cumulus clouds in Doppler radar data: Obtain high-temporal and spatial resolution radar composite reflectivity data after network quality control from the weather radar mosaic system, and then identify and mark cumulus cloud areas from the radar composite reflectivity data. Cumulus cloud areas are convective areas. Step B: Identify radar convection primary events using the area overlap method: The grid points corresponding to the marked convection areas in the preprocessed radar reflectivity data are used as an eight-connected area as a research object. Based on the target-oriented idea, a convection area is used as a sample; Step C: Preprocessing of satellite convective primary products: Use Python to extract convective primary product data, perform geometric position correction, data quality control, and finally generate binary data to construct a satellite convective primary data set; Step D: Using the neighborhood method, based on the idea of ​​oriented to the target area, calculate the correct forecast, false alarm, and missed alarm events of the satellite convective primary events; Step E: Calculate the verification index and time advance of convection initiation: Based on satellite convection initiation products and radar convection initiation events, take radar convection initiation as the actual situation, and use the neighborhood method facing the convection initiation area to calculate the satellite convection initiation hit rate POD, false alarm rate FAR, verification index and time advance TS. The formula is as follows: Hit is the number of convective initiation areas correctly predicted by the satellite, False is the number of convective initiation areas falsely reported by the satellite, and Failure is the number of convective initiation areas missed by the satellite.

2. The method for evaluating primary products of geostationary satellite convection based on Doppler radar according to claim 1, characterized in that: Identifying and marking cumulus cloud areas in radar composite reflectivity data includes the following steps: S1: interpolation: using the average method to interpolate the radar combined reflectivity data into longitude and latitude grid data with a resolution of 0.04°×0.04°; S2: Perform 0 / 1 conversion: 35 dBz is used as the discrimination threshold of convective intensity. The area with reflectivity value ≥ 35 dBz is the convective area and is replaced by the value 1. The area with reflectivity value < 35 dBz is non-cumulus and is converted to the value 0. S3: Perform data quality control to avoid interference from isolated noise and strong echoes. Before identifying the incipient convection, filter out cumulus areas with smaller areas and remove identified cumulus areas with ≤3 pixels.

3. The method for evaluating primary products of geostationary satellite convection based on Doppler radar according to claim 2, characterized in that: The identification of radar convective incipient events using the area overlap method includes the following steps: E1: Read the radar reflectivity data of the current time and the previous time, and use the neighborhood method to identify the convection areas of the two time periods, that is, the area with a mark value of 1, which is composed of convection grid points connected in eight directions; E2: Use the area overlap method to track the cloud clusters, remove the convective events that are generated by the translation or development of the cloud clusters at the previous moment, and only extract the convective events generated at the current moment, which is the convective initiation, and set it as a radar convective initiation live sample; E3: Constructing radar convection primary database: Z_RADA_C_BABJ_YYYYMMDDHHmmss_P_DOR_ACHN_CREF_YYYYMMDD_HHmmss.dat is a dat format file for data naming, which provides a real-time data set for the subsequent evaluation of satellite convection primary products.

4. The method for evaluating primary products of geostationary satellite convection based on Doppler radar according to claim 3, characterized in that: The data naming format is: YYYYMMDDhhmmss: product generation time, year, month, day, hour, minute, and second, all in universal time; ACHN: area code, the national area is ACHN; CREF: abbreviation of radar product type, CREF is radar combination reflectivity; YYYYMMDD_hhmmss: data observation time, year, month, day, hour, minute, second; dat: file extension.

5. The method for evaluating primary products of geostationary satellite convection based on Doppler radar according to claim 4, characterized in that: The geometric position correction includes the following steps: N1: According to the row and column number and longitude and latitude lookup table, if the longitude and latitude data of FY-4B is not integrated into the L1 data set, it is necessary to download the original longitude and latitude file from the website of the National Satellite Meteorological Center; N2: For different resolutions corresponding to different longitude and latitude files, the longitude and latitude file corresponding to the 0.04°×0.04° resolution is FY4B-_DISK_1050E_GEO_NOM_LUT_20240227000000_4000M_V0001.raw, with a grid size of 2748×2748. This file is a two-dimensional image file with geographic location coordinate information, which corresponds one-to-one to the pixels of the corrected image. N3: The data file in N2 is first filled in from north to south by row or from west to east. Each data stores the corresponding latitude and longitude; each pixel finds the corresponding longitude and latitude information according to the row number and column number in the image to achieve precise positioning.

6. The method for evaluating primary products of geostationary satellite convection based on Doppler radar according to claim 1, characterized in that: The data quality control specifically includes eliminating convective incipient events with ≤3 pixels.

7. The method for evaluating primary products of geostationary satellite convection based on Doppler radar according to claim 1, characterized in that: The calculation of the correct forecast, false forecast, and missed forecast of satellite convective initiation events includes the following steps: M1: The time resolution of the two is matched by the proximity principle. Based on the satellite convective primary product, the convective primary established by using radar data is taken as the actual situation, and compared with the radar convective primary product of each time before and after 2 hours, and the time advance of the satellite convective primary product is calculated; M2: Based on the goal-oriented idea, a convection initiation area is taken as a sample to determine the correct forecast, missed forecast and false alarm, and the false alarm result is finally generated; M3: Determine the advance time of satellite convection initiation: within the forecast period, the time difference between the correct forecast of satellite convection initiation and the initial satellite convection initiation is the forecast time advance, and the subsequent times do not need to be checked.

8. The method for evaluating primary products of geostationary satellite convection based on Doppler radar according to claim 7, characterized in that: The method for determining correct forecasts, missed forecasts and false alarms is as follows: when the satellite convection initiation area overlaps with the radar convection initiation area or the number of adjacent grid points is less than 30, it is considered a correct forecast; when the satellite convection initiation area is not adjacent to the radar convection initiation area at all times, it is considered a missed forecast; when the satellite convection initiation area is not adjacent to all radar convection initiation areas at all times within the time period, it is considered a false forecast.

Images