Gust forecasting method based on observation correction and suitable for numerical forecasting mode
By combining historical gust observation data and numerical forecasting mode, the gust forecasting equation is constructed using the Mooning Obkhov boundary layer similarity theory and made observation and correction, the problem of uncertainty and overestimation of empirical parameter terms in the existing gust forecast is solved. Improve forecast accuracy and reliability.
Patent Information
- Application Number
- CN202411456216.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-18
- Publication Date
- 2024-11-15
AI Technical Summary
The existing gust forecasting methods have problems with uncertainty in empirical parameter terms and overestimation, making it difficult to fully consider the gust observation and the characteristics of the mode itself, resulting in poor forecasting results.
The gust forecast equation is constructed using the method based on observation correction, combined with historical gust observation data and numerical forecast mode, and the gust forecast correction value is obtained through frequency matching to optimize the preliminary forecast results.
It improves the accuracy and reliability of gust forecasts, reduces the uncertainty of forecast results, and has higher applicability and reliability than traditional methods.
Smart Images

Figure CN118962854A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a gust forecasting method suitable for a numerical forecasting mode based on observation correction, and belongs to the technical field of weather forecasting. Background Art
[0002] Gusts refer to short-term extreme strong winds that occur over a period of time. Compared with the average wind speed, gusts have greater energy and are more sudden, often causing huge casualties and property losses. In recent years, strong wind disasters caused by severe convective weather in my country have tended to occur repeatedly, and extreme gusts have become one of the main disasters that cause casualties and economic losses. Further improving the level of gust forecasting has important scientific significance for improving meteorological disaster prevention and mitigation capabilities.
[0003] At present, the gust forecasting methods in the world mainly include empirical statistical forecasting methods and forecasting methods based on physical theoretical models. The empirical statistical forecasting method does not need to consider the physical mechanism affecting the gusts. The gust forecast can be carried out through the statistical relationship between the gusts and other factors. However, the empirical statistical model lacks the understanding and description of the physical process of the disaster-causing gusts, and there are certain uncertainties, and it is difficult to obtain a good forecasting effect. The forecasting method based on the physical theoretical model is usually applicable to the numerical forecasting model. The physical mechanism hypothesis of the formation of gusts is considered in the forecasting model. The commonly used method is based on the boundary layer similarity theory. This scheme is applied to the European Center for Numerical Weather Forecasts (ECMWF) and the Beijing Institute of Urban Meteorology, etc., but there are still empirical parameter items in the scheme. It should be noted that the forecast results of the model are greatly affected by the model resolution and the selection of the physical process scheme. How to determine the empirical parameter items in the gust forecast equation according to the model's own characteristics (physical process, resolution, etc.) is the key to affecting the gust forecast effect. With the improvement of computer capabilities, the horizontal resolution of the existing mesoscale model has reached kilometers or smaller scales. Compared with the ECMWF forecast, the resolution can be improved by more than 10 times. Existing studies have directly applied the ECMWF turbulent gust forecast model in the mesoscale forecast model without optimizing the empirical parameters in the scheme, resulting in an overestimation of the gust forecast.
[0004] Therefore, how to fully consider the gust observation and the characteristics of the model itself, including how to determine the empirical parameter items in the gust forecast method of the physical theoretical model and optimize the gust forecast effect, is the key to further develop and optimize the gust forecast method suitable for the model. Summary of the invention
[0005] The present invention aims to address the deficiencies of existing gust forecasting technology, provide a gust forecasting method suitable for numerical forecasting based on observation correction based on historical gust observation data and forecasts of numerical forecast models, combined with boundary layer similarity theory.
[0006] In order to achieve the above technical objectives, the present invention will adopt the following technical solutions:
[0007] A gust forecasting method based on observation correction and applicable to a numerical forecasting model comprises the following steps:
[0008] Step 1: Select an atmospheric numerical prediction model for gust forecasting, and ensure that the atmospheric numerical prediction model for gust forecasting can output a specified meteorological element field according to a specified area and time interval; the output specified meteorological element field includes key parameters for gust forecasting;
[0009] Step 2: Obtain a historical gust observation data set, where the historical gust observation data set includes gust observation data within a specified historical model forecast time period;
[0010] Step 3: Preliminary construction of gust forecast equation:
[0011] Based on Monin Obukhov boundary layer similarity theory, historical gust observation data set and the forecast of the specified meteorological elements in the same time period, the gust forecast equation is obtained by fitting the empirical parameter terms of the gust forecast equation using statistical methods.
[0012] Step 4: Observation correction:
[0013] The preliminary gust forecast result within the specified historical model forecast time period is calculated by the gust forecast equation obtained in step 3 to form a preliminary gust forecast data set;
[0014] The calculated preliminary gust forecast data set is frequency matched with the historical gust observation data set obtained in step 2 to obtain gust forecast correction values corresponding to each given gust forecast threshold, thereby forming a gust forecast correction array to perform observation correction on the preliminary gust forecast result calculated in step 3; the gust forecast correction array includes a plurality of elements, each element including a corresponding given gust forecast threshold and a gust forecast correction value;
[0015] Step 5: Get the gust forecast product within 1 hour of the specified forecast time:
[0016] Based on the atmospheric numerical prediction model selected in step 1, the specified forecast time is output Specify the meteorological element field within 1 hour, and first calculate the corresponding preliminary gust forecast result through the gust forecast equation obtained in step 3 Then, through the gust forecast correction array calculated in step 4, select the one that matches the preliminary gust forecast result. The observation correction value corresponding to the given gust forecast threshold is matched to obtain the specified forecast time. 1-hour gust forecast product.
[0017] Preferably, in step 3, the construction of the gust forecast equation specifically includes the following steps:
[0018] Step 3.1, based on the boundary layer Monin-Obukhov similarity theory and the time interval output by the atmospheric numerical prediction model for gust forecasting, the gust forecast equation is determined as follows:
[0019] (1)
[0020] (2)
[0021] Where: It is the preliminary gust forecast result within 1 hour at the specified forecast time. It is the maximum wind speed value output by the atmospheric numerical prediction model for gust forecast within 1 hour at the specified forecast time. is the estimated value of turbulent gusts based on physical assumptions at the time of the maximum wind speed output by the atmospheric numerical prediction model used for gust forecast within 1 hour of the specified forecast time, is the empirical parameter term, is the remaining estimate; is the boundary layer height, calculated according to the atmospheric numerical prediction model used for gust forecasting, is the length of Monin Obukhov, is the friction speed;
[0022] Step 3.2: Specify the maximum wind speed value within the historical model forecast period , And the corresponding gust observation data in the historical gust observation data set constructed in step 2 Perform time-space matching to obtain a data set M containing n samples. The samples are recorded as { , , };
[0023] Step 3.3: Calculate the empirical parameter term in the gust forecast equation and the remaining estimates :
[0024] The gust forecast equation is transformed into the following one-variable linear equation:
[0025] (3)
[0026] Where: ; ; ;
[0027] Based on the least squares method and the data set M obtained in step 3.2, the empirical parameter term in equation (3) is fitted and solved and the remaining estimates :
[0028] (4)
[0029] (5)
[0030] (6)
[0031] (7)
[0032] (8)
[0033] (9)
[0034] Through the above formula, we can get the empirical parameter and , and then based on the gust forecast equation in step 3.1, the preliminary gust forecast result can be obtained .
[0035] Preferably, in step 4, the observation correction specifically includes the following steps:
[0036] Step 4.1: Based on the preliminary gust forecast dataset and the historical gust observation dataset for the same period, a wind speed threshold set is given as needed. ;in: is the total number of given wind speed thresholds; , satisfy: and , where: is the maximum value of the preliminary gust forecast data set, is the maximum value in the historical gust observation data set;
[0037] Step 4.2: Based on the preliminary gust forecast dataset and the historical gust observation dataset, calculate the wind speed thresholds Corresponding preliminary gust forecast values And gust observation data To construct the corresponding cumulative distribution function CDF, respectively recorded as the predicted cumulative frequency and observed cumulative frequency , and draw the forecast cumulative frequency curve and the observation cumulative frequency curve in the same coordinate system XOY, where: in the coordinate system XOY, the horizontal axis is the wind speed threshold , the vertical axis is the cumulative frequency ;
[0038] In calculating the forecast cumulative frequency When any preliminary gust forecast value satisfy: When Statistical inlet wind speed threshold middle;
[0039] In calculating the observed cumulative frequency When any gust observation data satisfy: When the corresponding gust observation data , statistics of wind speed threshold middle;
[0040] Step 4.3: Based on the forecast cumulative frequency curve and the observed cumulative frequency curve drawn in step 4.2, calculate each wind speed threshold one by one based on the principle of frequency matching. Corresponding gust forecast correction value , to obtain the gust forecast correction value set ; Gust forecast correction value The calculation formula is as follows:
[0041] (10)
[0042] Where: Indicates wind speed threshold The corresponding revised forecast gust wind speed threshold is the forecast cumulative frequency Calculate the observed cumulative frequency by interpolation The corresponding wind speed threshold;
[0043] Gust Forecast Correction Group Corrected values of wind gust forecasts in With a given wind speed threshold group The wind speed thresholds in A gust forecast correction array is constructed in a one-to-one correspondence.
[0044] Preferably, in step 5, for any specified forecast time Preliminary gust forecast for the hour , the corrected gust forecast result can be calculated by the following formula: :
[0045] (11)
[0046] Where: The output of the corrected forecast time The gust forecast result within 1 hour is: is the specified forecast time based on the output of the gust forecast equation Preliminary gust forecast value within 1 hour; Based on the preliminary gust forecast Determine the gust forecast correction value: first change the initial gust forecast value With a given wind speed threshold group The wind speed thresholds in When compared with the initial gust forecast value With a given wind speed threshold group A wind speed threshold in satisfy: When the gust forecast correction value set obtained in step 4.3 is , we can get the wind speed threshold Corresponding gust forecast correction value , the gust forecast correction value This is the value based on the preliminary gust forecast. Gust forecast correction determined .
[0047] Preferably, when the forecast time is specified Preliminary gust forecast for Greater than the gust wind speed limit The forecast time will be specified Preliminary gust forecast for Equivalent to the gust wind speed limit To determine the gust forecast correction ; Gust wind speed limit satisfy:
[0048] and ;
[0049] Where: is the maximum value of the preliminary gust forecast data set, It is the maximum value in the historical gust observation dataset.
[0050] Preferably, in step 4.1, the wind speed interval between two adjacent wind speed thresholds is 1 m / s.
[0051] Preferably, in step 2, obtaining the historical gust observation data set specifically includes the following steps:
[0052] Step 2.1: Obtain the hourly historical gust observation dataset for gust correction, and determine the site range and start and end time of the historical gust observation dataset;
[0053] Step 2.2: Perform quality control on the historical gust observation data set and assign missing values to abnormally large or small values;
[0054] Step 2.3: Further quality control is performed on the historical gust observation dataset to check the continuity of the data and assign missing values to the default time periods.
[0055] Preferably, in step 1, the key parameters for gust forecasting include the zonal wind speed at a height of 10 m, the meridional wind speed at a height of 10 m, the friction speed, the Monin Obukhov length and the boundary layer height.
[0056] Preferably, the atmospheric numerical prediction model for gust forecasting is selected by the following steps:
[0057] Step 1.1: Select any atmospheric numerical forecast model and determine the grid area range, horizontal resolution and time frequency of output meteorological elements for gust forecast;
[0058] Step 1.2: Check whether the standard output file of the selected atmospheric numerical prediction model includes the key parameters for gust calculation. If the standard output file lacks the key parameters for gust calculation, modify the standard output file so that the standard output file includes the key parameters for gust calculation, and the atmospheric numerical prediction model for gust forecasting can be obtained.
[0059] Preferably, in step 1.1, the selected atmospheric numerical forecast model is the WRF meteorological model, the horizontal resolution is 3 km, and the time frequency of outputting meteorological elements is 200 times per hour; in step 1.2, the standard output file of the WRF meteorological model needs to be modified so that the standard output file of the WRF meteorological model includes key parameters for gust calculation.
[0060] Based on the above technical objectives, the present invention has the following advantages over the prior art:
[0061] Based on the existing gust forecast, the present invention provides a gust forecast method suitable for numerical forecast based on observation correction, based on historical gust observation data and numerical forecast model forecast, combined with boundary layer similarity theory. This method fully considers the model's own characteristics, the influence of the model's physical process on gust forecast, and optimizes the gust forecast results using historical observation data. Compared with directly applying the existing gust forecast model, this method is more reasonable, reliable and universal. BRIEF DESCRIPTION OF THE DRAWINGS
[0062] Figure 1 The present invention is a flow chart of a gust forecasting method based on observation correction and applicable to a numerical forecasting model.
[0063] Figure 2The following are example effect diagrams of gust forecast results obtained based on the present invention. In the figure, (a) represents a one-hour gust forecast result with a start time of 00:00 (UTC) on March 25 and a forecast period of 09:00-10:00 on March 25; (b) represents a 24-hour gust forecast result with a start time of 00:00 (UTC) on March 25 and a forecast period of 00:00 on March 25-00:00 on March 26.
[0064] Figure 3 The TS score comparison between the gust forecast result obtained based on the present invention and the gust forecast result obtained based on the traditional gust forecast scheme is shown in the figure. DETAILED DESCRIPTION
[0065] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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. The following description of at least one exemplary embodiment is actually only illustrative and is by no means any limitation to the present invention and its application or use. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Unless otherwise specified, the relative arrangement, expressions and numerical values of the components and steps described in these embodiments do not limit the scope of the present invention. The techniques, methods and equipment known to ordinary technicians in the relevant fields may not be discussed in detail, but in appropriate cases, the techniques, methods and equipment should be regarded as part of the specification. In all examples shown and discussed here, any specific value should be interpreted as merely exemplary, rather than as a limitation. Therefore, other examples of the exemplary embodiments may have different values.
[0066] Taking the WRF meteorological model and Jiangsu region as an example, the present invention provides a gust forecasting method suitable for numerical forecasting based on observation correction, such as Figure 1 As shown, the gust forecasting method of the present invention comprises the following steps:
[0067] Step 1: For the atmospheric numerical prediction model used for gust forecast (WRF meteorological model can be selected), modify the control file or related code of the atmospheric numerical prediction model so that the atmospheric numerical prediction model can output the specified meteorological element field according to the specified area and time interval. The output specified meteorological element field includes the key parameters for gust forecast; specifically, the following steps are implemented:
[0068] Step 1.1: Select any atmospheric numerical forecast model and determine the grid area range, horizontal resolution and time frequency of output meteorological elements for gust forecast;
[0069] In the present invention, the selected atmospheric numerical forecast model is the WRF meteorological model, the horizontal resolution is 3 km, and the time frequency of outputting meteorological elements is 200 times per hour according to the integration time step (18 s);
[0070] Step 1.2: Check whether the standard output file of the selected atmospheric numerical prediction model includes the key parameters for gust calculation. If the standard output file lacks the key parameters for gust calculation, modify the standard output file so that the standard output file includes the key parameters for gust calculation, and the atmospheric numerical prediction model for gust forecasting can be obtained.
[0071] In the present invention, the key parameters for gust calculation include the zonal wind speed at a height of 10 m, the meridional wind speed at a height of 10 m, the friction speed, the Monin Obukhov length and the boundary layer height.
[0072] In the WRF meteorological model, the default output does not include the Monin Obukhov length. Therefore, it is necessary to modify the operation control file of the WRF meteorological model (namelist.input, the standard output file mentioned above), set the stream output, and output the key parameters used for gust calculation; the key parameters used for gust calculation include the zonal wind speed at 10 m, the meridional wind speed at 10 m, the friction speed, the Monin Obukhov length and the boundary layer height. The output time frequency is 200 times per hour (18 s).
[0073] Step 2: Obtain the historical gust observation dataset.
[0074] The historical gust observation dataset includes the gust observation data within the specified historical model forecast period, which is constructed through the following steps:
[0075] Step 2.1: Obtain the hourly historical gust observation dataset for gust correction, and determine the site range and start and end time of the historical gust observation dataset;
[0076] The site range can be selected as all national stations in Jiangsu and Anhui provinces, and the start and end time can be selected from 01:00 on June 1, 2023 to 00:00 on October 1, 2023.
[0077] Step 2.2: Perform quality control on the historical gust observation dataset and assign missing values to abnormally large or small values.
[0078] Step 2.3: Further quality control is performed on the historical gust observation dataset to check the continuity of the data and assign missing values to the default time periods.
[0079] Step 3: Construct the gust forecast equation:
[0080] Based on Monin Obukhov boundary layer similarity theory, historical gust observation data set and the specified meteorological element field forecast in the same time period, the gust forecast equation can be obtained by fitting the empirical parameter terms of the gust forecast equation using statistical methods, which specifically includes the following steps:
[0081] Step 3.1, based on the boundary layer Monin-Obukhov similarity theory and the time interval output by the atmospheric numerical prediction model for gust forecasting, the gust forecast equation is determined as follows:
[0082] (1)
[0083] (2)
[0084] Where: For a specified forecast time The preliminary gust forecast results within 1 hour are: For a specified forecast time The maximum wind speed value output by the atmospheric numerical prediction model for gust forecast within 1 hour, For a specified forecast time The estimated value of turbulent gusts based on physical assumptions at the time of the maximum wind speed output by the atmospheric numerical prediction model for gust forecast within 1 hour, is the empirical parameter term, is the remaining estimate; is the boundary layer height, calculated according to the atmospheric numerical prediction model used for gust forecasting, is the length of Monin Obukhov, is the friction speed.
[0085] Based on the selected WRF forecast model, the forecast element field of the historical time period is obtained and the formula (1) is calculated. , .
[0086] It should be noted that the historical forecast data period is from 01:00 on June 1, 2023 to 00:00 on October 1, 2023 (UTC), and the model forecast is reported from 00:00 on each day, with a forecast of 24 hours.
[0087] Step 3.2: Specify the maximum wind speed value within the historical model forecast period , And the corresponding gust observation data in the historical gust observation data set constructed in step 2 Perform time-space matching to obtain a data set M containing n samples. The samples are recorded as { , , }.
[0088] Step 3.3: Calculate the empirical parameter term in the gust forecast equation and the remaining estimates :
[0089] The gust forecast equation is transformed into the following one-variable linear equation:
[0090] (3)
[0091] Where: ; ; ;
[0092] Based on the least squares method and the data set M obtained in step 3.2, the empirical parameter term in equation (3) is fitted and solved and the remaining estimates :
[0093] (4)
[0094] (5)
[0095] (6)
[0096] (7)
[0097] (8)
[0098] (9)
[0099] Through the above formula, we can get the empirical parameter and , and then based on the gust forecast equation in step 3.1, the preliminary gust forecast result can be obtained .
[0100] Step 4: Observation correction:
[0101] The preliminary gust forecast result within the specified historical model forecast time period is calculated by the gust forecast equation obtained in step 3 to form a preliminary gust forecast data set;
[0102] The calculated preliminary gust forecast data set is frequency matched with the historical gust observation data set obtained in step 2 to obtain gust forecast correction values corresponding to each given gust forecast threshold, thereby forming a gust forecast correction array to perform observation correction on the preliminary gust forecast result calculated in step 3; the gust forecast correction array includes a number of elements, each element includes a corresponding given gust forecast threshold and a gust forecast correction value.
[0103] The observation correction described in this step specifically includes the following steps:
[0104] Step 4.1: Based on the preliminary gust forecast dataset and the historical gust observation dataset for the same period, a wind speed threshold set is given as needed. ;in: is the total number of given wind speed thresholds; , satisfy: and , where: is the maximum value of the preliminary gust forecast data set, It is the maximum value in the historical gust observation dataset.
[0105] In the present invention, the wind speed threshold is set at 0-100 m / s, and the wind speed interval is 1 m / s. Of course, other wind speed intervals can also be selected, such as 2 m / s, 5 m / s, etc.
[0106] Step 4.2: Based on the preliminary gust forecast dataset and the historical gust observation dataset, calculate the wind speed thresholds Corresponding preliminary gust forecast values And gust observation data To construct the corresponding cumulative distribution function CDF, respectively recorded as the predicted cumulative frequency and observed cumulative frequency , and draw the forecast cumulative frequency curve and the observation cumulative frequency curve in the same coordinate system XOY, where: in the coordinate system XOY, the horizontal axis is the wind speed threshold , the vertical axis is the cumulative frequency ;
[0107] In calculating the forecast cumulative frequency When any preliminary gust forecast value satisfy: When Statistical inlet wind speed threshold middle;
[0108] In calculating the observed cumulative frequency When any gust observation data satisfy: When the corresponding gust observation data , statistics of wind speed threshold middle.
[0109] Step 4.3: Based on the forecast cumulative frequency curve and the observed cumulative frequency curve drawn in step 4.2, calculate each wind speed threshold one by one based on the principle of frequency matching. Corresponding gust forecast correction value , to obtain the gust forecast correction value set ; Gust forecast correction value The calculation formula is as follows:
[0110] (10)
[0111] Where: Indicates wind speed threshold The corresponding revised forecast gust wind speed threshold is the forecast cumulative frequency Calculate the observed cumulative frequency by interpolation The corresponding wind speed threshold;
[0112] Gust Forecast Correction Group Corrected values of wind gust forecasts in With a given wind speed threshold group The wind speed thresholds in In other words, any element in the gust forecast correction array can be expressed as .
[0113] Step 5: Get the specified forecast time Gust forecast products:
[0114] Based on the atmospheric numerical prediction model selected in step 1, the specified forecast time is output Specify the meteorological element field within 1 hour, and first calculate the corresponding preliminary gust forecast result through the gust forecast equation obtained in step 3 Then, through the gust forecast correction array calculated in step 4, select the one that matches the preliminary gust forecast result. The observation correction value corresponding to the given gust forecast threshold is matched to obtain the specified forecast time. 1-hour gust forecast product.
[0115] For any given forecast time Preliminary gust forecast for , the corrected gust forecast result can be calculated by the following formula: :
[0116] (11)
[0117] Where: The output of the corrected forecast time The gust forecast results within 1 hour are: is the specified forecast time based on the output of the gust forecast equation Preliminary gust forecast value within 1 hour; Based on the preliminary gust forecast Determine the gust forecast correction value: first change the initial gust forecast value With a given wind speed threshold group The wind speed thresholds in When compared with the initial gust forecast value With a given wind speed threshold group A wind speed threshold in satisfy: When the gust forecast correction value set obtained in step 4.3 is , we can get the wind speed threshold Corresponding gust forecast correction value , the gust forecast correction value This is the value based on the preliminary gust forecast. Gust forecast correction determined .
[0118] In addition, it is important to note that when specifying a forecast time Preliminary gust forecast for the hour Gust wind speed limit The forecast time will be specified Preliminary gust forecast for Equivalent to the gust wind speed limit To determine the gust forecast correction ; Gust wind speed limit satisfy:
[0119] and ; Where: is the maximum value of the preliminary gust forecast data set, is the maximum value in the historical gust observation data set. In the present invention, the gust wind speed limit value It can be 24m / s. At this time, the specified forecast time can be Preliminary gust forecast for the hour The equivalent value is 24m / s to determine the gust forecast correction value .
[0120] In other words, the gust forecast product obtained by the present invention within any forecast time period is , can be calculated by the following gust forecast correction equation:
[0121]
[0122]
[0123]
[0124] The step five is divided into the following steps:
[0125] Step 5.1: Based on the WRF meteorological forecast model, obtain the specified meteorological element field, and further obtain the final gust forecast result based on the gust forecast equation and gust correction value.
[0126] Step 5.2: Based on the selected target area (115 o E-123 o E,30 o N-36 o N) and the determined forecast period to form a gust forecast product. For example, the one-hour gust forecast result with the start time of March 25 00:00 (UTC) and the forecast period of March 25 09:00-10:00 is as follows: Figure 2 As shown in (a) in the figure; the starting time is 00:00 on March 25 (UTC), and the forecast period is 00:00 on March 25-00:00 on March 26. The 24-hour gust forecast results are as follows: Figure 2 As shown in (b) in .
[0127] In order to verify the effect of the gust forecast method of the present invention, the gust forecast product is first obtained based on the forecast model. The forecast model starts at 00:00 every day in March 2024, and the forecast time is 24 hours. The TS scoring method is further used to verify the forecast results. TS score (Threat Score), also known as Skill Score, is a commonly used indicator in meteorology for evaluating forecast accuracy, which can measure the ability of the forecast system to forecast specific weather events. The specific formula of TS score is: , where H is the number of stations with correct forecasts; F is the number of stations with false forecasts; M is the number of stations with missed forecasts. The observation data used the hourly gust observation data of 70 national stations in Jiangsu Province in March 2024 to verify the gust forecast results. In order to compare the forecast effect of the gust forecast algorithm, the gust forecast results of the EC turbulent gust scheme (IUM gust scheme of Beijing Institute of Urban Meteorology) were also calculated for comparison. The TS score results are as follows Figure 3 As shown, from Figure 3 It can be clearly seen in the gust forecast results based on the present invention ( ) in each wind speed range are better than the gust forecast results based on the traditional turbulent gust forecast scheme ( ), the forecast accuracy is improved by more than 10%.
[0128] In summary, the present invention provides a gust forecasting method suitable for numerical forecasting based on observation correction. This technology can be applied to the forecast and warning of disastrous gale, and can effectively provide a reference for disaster prevention and mitigation. This method comprehensively considers the inherent characteristics of the numerical model used for forecasting, the theoretical assumptions of the boundary layer, and optimizes the gust equations and results in combination with historical observation data. Compared with traditional gust forecasting methods, this method has a reasonable theoretical basis and has higher reliability and applicability. Therefore, this technology has a high promotion value.
[0129] The present invention provides a gust forecasting method suitable for numerical forecasting based on observation correction. There are many methods and ways to implement the technical solution. The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several modifications and improvements can be made without departing from the principle of the present invention. These modifications and improvements should also be regarded as the protection scope of the present invention. All components not specified in this embodiment can be implemented by existing technologies.
Claims
1. A gust forecasting method based on observation correction and applicable to numerical forecasting model, characterized in that: The steps include: Step 1: Select an atmospheric numerical prediction model for gust forecasting, and ensure that the atmospheric numerical prediction model for gust forecasting can output a specified meteorological element field according to a specified area and time interval; the output specified meteorological element field includes key parameters for gust forecasting; Step 2: Obtain a historical gust observation data set, where the historical gust observation data set includes gust observation data within a specified historical model forecast time period; Step 3: Preliminary construction of gust forecast equation: Based on Monin Obukhov boundary layer similarity theory, historical gust observation data set and the forecast of the specified meteorological elements in the same time period, the gust forecast equation is obtained by fitting the empirical parameter terms of the gust forecast equation using statistical methods. Step 4: Observation correction: The preliminary gust forecast result within the specified historical model forecast time period is calculated by the gust forecast equation obtained in step 3 to form a preliminary gust forecast data set; The calculated preliminary gust forecast data set is frequency matched with the historical gust observation data set obtained in step 2 to obtain gust forecast correction values corresponding to each given gust forecast threshold, thereby forming a gust forecast correction array to perform observation correction on the preliminary gust forecast result calculated in step 3; the gust forecast correction array includes a plurality of elements, each element including a corresponding given gust forecast threshold and a gust forecast correction value; Step 5: Get the gust forecast product within 1 hour of the specified forecast time: Based on the atmospheric numerical prediction model selected in step 1, the specified forecast time is output Specify the meteorological element field within 1 hour, and first calculate the corresponding preliminary gust forecast result through the gust forecast equation obtained in step 3 Then, through the gust forecast correction array calculated in step 4, select the one that matches the preliminary gust forecast result. The observation correction value corresponding to the given gust forecast threshold is matched to obtain the specified forecast time. 1-hour gust forecast product.
2. The gust forecasting method based on observation correction and applicable to numerical forecasting mode according to claim 1, characterized in that: In step 3, the construction of the gust forecast equation includes the following steps: Step 3.1, based on the boundary layer Monin-Obukhov similarity theory and the time interval output by the atmospheric numerical prediction model for gust forecasting, the gust forecast equation is determined as follows: (1) (2) Where: It is the preliminary gust forecast result within 1 hour at the specified forecast time. It is the maximum wind speed value output by the atmospheric numerical prediction model for gust forecast within 1 hour at the specified forecast time. It is the estimated value of turbulent gusts based on physical assumptions at the time of the maximum wind speed output by the atmospheric numerical prediction model used for gust forecast within 1 hour of the specified forecast time. is the empirical parameter term, is the remaining estimate; is the boundary layer height, calculated according to the atmospheric numerical prediction model used for gust forecasting, is the length of Monin Obukhov, is the friction speed; Step 3.2: Specify the maximum wind speed value within the historical model forecast period , And the corresponding gust observation data in the historical gust observation data set constructed in step 2 Perform time-space matching to obtain a data set M containing n samples. The samples are recorded as { , , }; Step 3.3: Calculate the empirical parameter term in the gust forecast equation and the remaining estimates : The gust forecast equation is transformed into the following linear equation: (3) Where: ; ; ; Based on the least squares method and the data set M obtained in step 3.2, the empirical parameter term in equation (3) is fitted and solved and the remaining estimates : (4) (5) (6) (7) (8) (9) Through the above formula, we can get the empirical parameter and , and then based on the gust forecast equation in step 3.1, the preliminary gust forecast result can be obtained .
3. The gust forecasting method based on observation correction and applicable to numerical forecasting mode according to claim 2 is characterized in that: In step 4, observation correction specifically includes the following steps: Step 4.1: Based on the preliminary gust forecast dataset and the historical gust observation dataset for the same period, a wind speed threshold set is given as needed. ;in: is the total number of given wind speed thresholds; , satisfy: and , where: is the maximum value of the preliminary gust forecast data set, is the maximum value in the historical gust observation data set; Step 4.2: Based on the preliminary gust forecast dataset and the historical gust observation dataset, calculate the wind speed thresholds Corresponding preliminary gust forecast value And gust observation data To construct the corresponding cumulative distribution function CDF, respectively recorded as the predicted cumulative frequency and observed cumulative frequency , and draw the forecast cumulative frequency curve and the observation cumulative frequency curve in the same coordinate system XOY, where: in the coordinate system XOY, the horizontal axis is the wind speed threshold , the vertical axis is the cumulative frequency ; In calculating the forecast cumulative frequency When any preliminary gust forecast value satisfy: The corresponding preliminary gust forecast value Statistical inlet wind speed threshold middle; In calculating the observed cumulative frequency When any gust observation data satisfy: When the corresponding gust observation data , statistics of wind speed threshold middle; Step 4.3: Based on the forecast cumulative frequency curve and the observation cumulative frequency curve drawn in step 4.2, calculate each wind speed threshold one by one based on the principle of frequency matching. Corresponding gust forecast correction value , to obtain the gust forecast correction value set ; Gust forecast correction value The calculation formula is as follows: (10) Where: Indicates wind speed threshold The corresponding revised forecast gust wind speed threshold is the forecast cumulative frequency Calculate the observed cumulative frequency by interpolation The corresponding wind speed threshold; Gust Forecast Correction Group Corrected values of wind gust forecasts in With a given wind speed threshold group The wind speed thresholds in A gust forecast correction array is constructed in a one-to-one correspondence.
4. The gust forecasting method based on observation correction and applicable to numerical forecasting mode according to claim 3 is characterized in that: In step 5, for any given forecast time Preliminary gust forecast for the hour , the corrected gust forecast result can be calculated by the following formula: : (11) Where: The output of the corrected forecast time The gust forecast result within 1 hour is: is the specified forecast time based on the output of the gust forecast equation Preliminary gust forecast value within 1 hour; Based on the preliminary gust forecast Determine the gust forecast correction value: first change the initial gust forecast value With a given wind speed threshold group The wind speed thresholds in When compared with the initial gust forecast value With a given wind speed threshold group A wind speed threshold in satisfy: When the gust forecast correction value set obtained in step 4.3 is , we can get the wind speed threshold Corresponding gust forecast correction value , the gust forecast correction value This is the value based on the preliminary gust forecast. Gust forecast correction determined .
5. The gust forecasting method based on observation correction and applicable to numerical forecasting mode according to claim 4, characterized in that: When the forecast time is specified Preliminary gust forecast for Greater than the gust wind speed limit The forecast time will be specified Preliminary gust forecast for Equivalent to the gust wind speed limit To determine the gust forecast correction ; Gust wind speed limit satisfy: and ; Where: is the maximum value of the preliminary gust forecast data set, It is the maximum value in the historical gust observation dataset.
6. The gust forecasting method based on observation correction and applicable to numerical forecasting mode according to claim 1, characterized in that: In step 4.1, the wind speed interval between two adjacent wind speed thresholds is 1 m / s.
7. The gust forecasting method based on observation correction and applicable to numerical forecasting mode according to claim 1, characterized in that: In step 2, the acquisition of historical gust observation data set includes the following steps: Step 2.1: Obtain the hourly historical gust observation dataset for gust correction, and determine the site range and start and end time of the historical gust observation dataset; Step 2.2: Perform quality control on the historical gust observation data set and assign missing values to abnormally large or small values; Step 2.3: Further quality control is performed on the historical gust observation dataset to check the continuity of the data and assign missing values to the default time periods.
8. The gust forecasting method based on observation correction and applicable to numerical forecasting mode according to claim 1, characterized in that: In step 1, the key parameters used for gust forecasting include the zonal wind speed at 10 m, the meridional wind speed at 10 m, the friction velocity, the Monin-Obukhov length and the boundary layer height.
9. The gust forecasting method based on observation correction and applicable to numerical forecasting mode according to claim 1, characterized in that: The atmospheric numerical prediction model used for gust forecasting is selected through the following steps: Step 1.1: Select any atmospheric numerical forecast model and determine the grid area range, horizontal resolution and time frequency of output meteorological elements for gust forecast; Step 1.2: Check whether the standard output file of the selected atmospheric numerical prediction model includes the key parameters for gust calculation. If the standard output file lacks the key parameters for gust calculation, modify the standard output file so that the standard output file includes the key parameters for gust calculation, and the atmospheric numerical prediction model for gust forecasting can be obtained.
10. The gust forecasting method based on observation correction and applicable to numerical forecasting mode according to claim 9, characterized in that: In step 1.1, the selected atmospheric numerical forecast model is the WRF meteorological model, with a horizontal resolution of 3 km and a time frequency of outputting meteorological elements of 200 times per hour; in step 1.2, the standard output file of the WRF meteorological model needs to be modified so that the standard output file of the WRF meteorological model includes the key parameters for gust calculation.
Citation Information
Patent Citations
Weather warning system and method
CA2853915A1
Offshore wind resource comprehensive evaluation method and device, storage medium and electronic equipment
CN114548634A
Visibility forecasting model based on multi-meteorological-factor intelligent deep learning
CN114880958A
Method and device for objectively forecasting hectometer gust
CN115808727A
Sub-season typhoon generation forecasting method based on multi-channel convolutional neural network
CN115857062A
