Thunderstorm gale forecast correction method and system under complex terrain and storage medium

Abstract

The invention discloses a thunderstorm gale forecast correction method and system under a complex terrain and a storage medium, and belongs to the technical field of weather forecast. The method comprises the following steps: S1, generating a thunderstorm and gale occurrence frequency data set to obtain thunderstorm and gale climate background data; s2, acquiring terrain factors, physical quantity factors and time factors to obtain thunderstorm gale characteristic variables; s3, constructing a partitioned thunderstorm and gale forecasting model according to the thunderstorm and gale characteristic variables and the altitude, and obtaining 3h thunderstorm and gale forecasting data through the trained thunderstorm and gale forecasting model; and S4, downscaling the 3h thunderstorm and gale forecast data to 1h to obtain 1h thunderstorm and gale forecast data based on the thunderstorm and gale climate background data, and finally outputting a dichotomy forecast result. The landform, weather and time factors are comprehensively considered, and region division is performed in combination with the altitude, so that thunderstorm and gale forecasting of different geographic regions is more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of meteorological forecasting, and in particular to a method, system and storage medium for correcting thunderstorm and gale forecasts under complex terrain. Background Art

[0002] In recent years, severe convective weather has become more frequent around the world. This type of weather disaster is characterized by a small spatial scale, strong suddenness, and great destructive power. Thunderstorms and gales are usually accompanied by heavy rainfall, hail, tornadoes, etc., which can cause serious damage to buildings, power facilities, transportation, etc. in a short period of time, bringing great challenges and losses to society. Summary of the invention

[0003] The purpose of the present invention is to overcome the deficiencies of the prior art and to provide a method, system and storage medium for correcting thunderstorm and gale forecasts under complex terrain.

[0004] The object of the present invention is achieved through the following technical solutions: The first aspect of the present invention provides: a method for correcting thunderstorm gale forecast under complex terrain, comprising the following steps: S1: Generate thunderstorm frequency data set to obtain thunderstorm climate background data; S2: Collect terrain factors, physical quantity factors and time factors to obtain thunderstorm gale characteristic variables; S3: Based on the altitude, the thunderstorm wind characteristic variables are used to partition the area, and the thunderstorm wind forecast model of the partition is constructed based on the adaptive lifting method. The thunderstorm wind forecast model is trained and corrected according to the thunderstorm wind observation data and thunderstorm wind characteristic variables, and the thunderstorm wind forecast results are tested and evaluated using the TS score; the 3h thunderstorm wind forecast data is obtained through the trained thunderstorm wind forecast model; S4: Based on the thunderstorm and gale climate background data, the 3-hour thunderstorm and gale forecast data is downscaled to 1 hour to obtain 1-hour thunderstorm and gale forecast data, and finally the binary classification forecast results are output.

[0005] Preferably, the S1 further comprises the following steps: According to the hourly maximum wind observation data on the ground and the lightning locator observation data, the thunderstorm wind labels of the target area stations are calibrated. On this basis, the monthly, hourly and annual average occurrence frequency of thunderstorm winds are obtained. The inverse distance weighted interpolation method is used to interpolate the monthly, hourly and annual average occurrence frequency of thunderstorm winds to form a gridded thunderstorm wind frequency dataset with a preset resolution.

[0006] Preferably, the terrain factors include slope, slope direction, surface roughness, and surface undulation; the physical quantity factors include high-altitude variables, single-layer variables, and combined quantities, and the physical quantity factors are obtained by direct output and calculation of model forecasts; and the time factors include month and hour.

[0007] Preferably, the S4 further comprises the following steps: The 3-hour thunderstorm and gale forecast data is input into the time downscaling encoder, and the 3-hour thunderstorm and gale forecast data is downscaled to 1 hour according to the high and low frequency grid points, high and low frequency occurrence periods, and total occurrence frequency to obtain 1 hour thunderstorm and gale forecast data.

[0008] Preferably, the preset resolution is 5km*5km. The second aspect of the present invention provides: a thunderstorm gale forecast correction system under complex terrain, which is used to implement any of the above-mentioned thunderstorm gale forecast correction methods under complex terrain, comprising: A background data generation module is used to generate a thunderstorm and gale frequency data set to obtain thunderstorm and gale climate background data; The characteristic variable generation module is used to collect terrain factors, physical quantity factors and time factors to obtain thunderstorm gale characteristic variables; The model building module is used to divide the area into zones according to the altitude, use thunderstorm and gale characteristic variables, and build a thunderstorm and gale forecast model for the zones based on the adaptive lifting method. The thunderstorm and gale forecast model is trained and corrected according to thunderstorm and gale observation data and thunderstorm and gale characteristic variables, and the thunderstorm and gale forecast results are tested and evaluated using the TS score. The 3-hour thunderstorm and gale forecast data is obtained through the trained thunderstorm and gale forecast model. The time downscaling module is used to downscale the 3-hour thunderstorm and gale forecast data to 1-hour thunderstorm and gale forecast data based on the thunderstorm and gale climate background data, and finally output the binary classification forecast results.

[0009] The third aspect of the present invention provides: a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are loaded and executed by a processor, any of the above-mentioned methods for correcting thunderstorm and gale forecasts under complex terrain is implemented.

[0010] The beneficial effects of the present invention are: 1) Comprehensively consider terrain, weather and time factors, and divide the regions according to altitude, so that the thunderstorm and gale forecast in different geographical areas is more accurate.

[0011] 2) The reliability of the forecast results is greatly improved by optimizing the thunderstorm and gale forecast model using the adaptive lifting method.

[0012] 3) By downscaling the 3-hour forecast time to 1 hour through climate background information, detailed hourly thunderstorm and gale forecasts are achieved, enhancing the operability and accuracy of thunderstorm and gale forecasts on a short-term scale.

[0013] 4) The test results show that the TS score of the hourly thunderstorm and gale forecast product has increased from 1.04% to 5.95%, and the false alarm rate has dropped from 98.8% to 80.8%. It not only shows good forecasting performance in long-term tests, but also outperforms traditional guidance products in individual tests. It can provide more accurate thunderstorm and gale forecasts for practical applications, has high business application value, and can provide support for the meteorological department's warning issuance and disaster prevention. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a flow chart of the correction method for thunderstorm and gale forecast under complex terrain. DETAILED DESCRIPTION

[0015] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0016] See also Figure 1 The first aspect of the present invention provides: a method for correcting thunderstorm and gale forecast under complex terrain, comprising the following steps: S1: Generate thunderstorm frequency data set to obtain thunderstorm climate background data; S2: Collect terrain factors, physical quantity factors and time factors to obtain thunderstorm gale characteristic variables; S3: Based on the altitude, the thunderstorm wind characteristic variables are used to partition the area, and the thunderstorm wind forecast model of the partition is constructed based on the adaptive lifting method. The thunderstorm wind forecast model is trained and corrected according to the thunderstorm wind observation data and thunderstorm wind characteristic variables, and the thunderstorm wind forecast results are tested and evaluated using the TS score; the 3h thunderstorm wind forecast data is obtained through the trained thunderstorm wind forecast model; S4: Based on the thunderstorm and gale climate background data, the 3-hour thunderstorm and gale forecast data is downscaled to 1 hour to obtain 1-hour thunderstorm and gale forecast data, and finally the binary classification forecast results are output.

[0017] In this embodiment, terrain factors, weather factors and time factors are comprehensively considered, and the forecast area is divided into high altitude areas and low altitude areas according to the altitude. The machine learning method is used to build a zoned thunderstorm and gale forecast model to obtain a 3-hour thunderstorm and gale potential forecast, and then the 3-hour forecast time is downscaled to 1 hour using the climate background to form an hourly thunderstorm and gale forecast. The forecast test results show that the adaptive boosting method is the best for the 3-hour thunderstorm and gale forecast. Both long-term tests and individual case tests show that the hourly thunderstorm and gale forecast product obtained based on the adaptive boosting method is better than the guidance product. The TS score is increased from 1.04% to 5.95%, and the false alarm rate is reduced from 98.8% to 80.8%, which has a high business application value.

[0018] In some embodiments, the S1 further comprises the following steps: According to the hourly maximum wind observation data on the ground and the lightning locator observation data, the thunderstorm wind labels of the target area stations are calibrated. On this basis, the monthly, hourly and annual average occurrence frequency of thunderstorm winds are obtained. The inverse distance weighted interpolation method is used to interpolate the monthly, hourly and annual average occurrence frequency of thunderstorm winds to form a gridded thunderstorm wind frequency dataset with a preset resolution.

[0019] In some embodiments, the terrain factors include slope, slope direction, surface roughness, and surface undulation; the physical quantity factors include high-altitude variables, single-layer variables, and combined quantities, and the physical quantity factors are obtained by direct output and calculation of model forecasts; the time factors include months and hours.

[0020] In some embodiments, the S4 further comprises the following steps: The 3-hour thunderstorm and gale forecast data is input into the time downscaling encoder, and the 3-hour thunderstorm and gale forecast data is downscaled to 1 hour according to the high and low frequency grid points, high and low frequency occurrence periods, and total occurrence frequency to obtain 1 hour thunderstorm and gale forecast data.

[0021] In some embodiments, the preset resolution is 5km*5km. The second aspect of the present invention provides: a thunderstorm gale forecast correction system under complex terrain, which is used to implement any of the above-mentioned thunderstorm gale forecast correction methods under complex terrain, comprising: A background data generation module is used to generate a thunderstorm and gale frequency data set to obtain thunderstorm and gale climate background data; The characteristic variable generation module is used to collect terrain factors, physical quantity factors and time factors to obtain thunderstorm gale characteristic variables; The model building module is used to divide the area into zones according to the altitude, use thunderstorm and gale characteristic variables, and build a thunderstorm and gale forecast model for the zones based on the adaptive lifting method. The thunderstorm and gale forecast model is trained and corrected according to thunderstorm and gale observation data and thunderstorm and gale characteristic variables, and the thunderstorm and gale forecast results are tested and evaluated using the TS score. The 3-hour thunderstorm and gale forecast data is obtained through the trained thunderstorm and gale forecast model. The time downscaling module is used to downscale the 3-hour thunderstorm and gale forecast data to 1-hour thunderstorm and gale forecast data based on the thunderstorm and gale climate background data, and finally output the binary classification forecast results.

[0022] The third aspect of the present invention provides: a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are loaded and executed by a processor, any of the above-mentioned methods for correcting thunderstorm and gale forecasts under complex terrain is implemented.

[0023] The above is only a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the concept described herein through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art shall not deviate from the spirit and scope of the present invention, and shall be within the scope of protection of the claims attached to the present invention.

Claims

1. A method for correcting thunderstorm and gale forecasts under complex terrain, characterized by: The following steps are involved: S1: Generate thunderstorm frequency data set to obtain thunderstorm climate background data; S2: Collect terrain factors, physical quantity factors and time factors to obtain thunderstorm gale characteristic variables; S3: Based on the altitude, the thunderstorm wind characteristic variables are used to partition the area, and the thunderstorm wind forecast model of the partition is constructed based on the adaptive lifting method. The thunderstorm wind forecast model is trained and corrected according to the thunderstorm wind observation data and thunderstorm wind characteristic variables, and the thunderstorm wind forecast results are tested and evaluated using the TS score; the 3h thunderstorm wind forecast data is obtained through the trained thunderstorm wind forecast model; S4: Based on the thunderstorm and gale climate background data, the 3-hour thunderstorm and gale forecast data is downscaled to 1 hour to obtain 1-hour thunderstorm and gale forecast data, and finally the binary classification forecast results are output.

2. The method for correcting thunderstorm and gale forecast under complex terrain according to claim 1 is characterized in that: The S1 further comprises the following steps: According to the hourly maximum wind observation data on the ground and the lightning locator observation data, the thunderstorm wind labels of the target area stations are calibrated. On this basis, the monthly, hourly and annual average occurrence frequency of thunderstorm winds are obtained. The inverse distance weighted interpolation method is used to interpolate the monthly, hourly and annual average occurrence frequency of thunderstorm winds to form a gridded thunderstorm wind frequency dataset with a preset resolution.

3. The method for correcting thunderstorm and gale forecast under complex terrain according to claim 1 is characterized in that: The terrain factors include slope, slope direction, surface roughness, and surface undulation; the physical quantity factors include high-altitude variables, single-layer variables, and combined quantities, and the physical quantity factors are directly output and calculated using model forecasts; the time factors include month and hour.

4. The method for correcting thunderstorm and gale forecast under complex terrain according to claim 1 is characterized in that: The S4 further comprises the following steps: The 3-hour thunderstorm and gale forecast data is input into the time downscaling encoder, and the 3-hour thunderstorm and gale forecast data is downscaled to 1 hour according to the high and low frequency grid points, high and low frequency occurrence periods, and total occurrence frequency to obtain 1 hour thunderstorm and gale forecast data.

5. The method for correcting thunderstorm and gale forecast under complex terrain according to claim 2 is characterized in that: The preset resolution is 5km*5km.

6. A thunderstorm gale forecast correction system under complex terrain, characterized by: The method for correcting thunderstorm and gale forecast under complex terrain according to any one of claims 1 to 5 comprises: A background data generation module is used to generate a thunderstorm and gale frequency data set to obtain thunderstorm and gale climate background data; The characteristic variable generation module is used to collect terrain factors, physical quantity factors and time factors to obtain thunderstorm gale characteristic variables; The model building module is used to divide the area into zones according to the altitude, use thunderstorm and gale characteristic variables, and build a thunderstorm and gale forecast model for the zones based on the adaptive lifting method. The thunderstorm and gale forecast model is trained and corrected according to thunderstorm and gale observation data and thunderstorm and gale characteristic variables, and the thunderstorm and gale forecast results are tested and evaluated using the TS score. The 3-hour thunderstorm and gale forecast data is obtained through the trained thunderstorm and gale forecast model. The time downscaling module is used to downscale the 3-hour thunderstorm and gale forecast data to 1-hour thunderstorm and gale forecast data based on the thunderstorm and gale climate background data, and finally output the binary classification forecast results.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are loaded and executed by the processor, the method for correcting thunderstorm and gale forecast under complex terrain as described in any one of claims 1-5 is implemented.

Images