Adjustment method for reconstructing large-scale driving field of convection-scale ensemble prediction

Abstract

This invention discloses a method for reconstructing and adjusting large-scale driving fields in convective-scale ensemble forecasts. The method includes obtaining a driving perturbation field and a control forecast of the convective-scale ensemble forecast system by performing dynamic downscaling deterministic forecasts of the driving field of a global ensemble forecast system and an EC fine-grid model; acquiring forecast element values ​​from both models within their forecast lead time, performing multi-sphere screening, and calculating forecast statistics; obtaining perturbation adjustment coefficients for each time step of the convective-scale ensemble forecast system based on the forecast statistics; adjusting the driving perturbation field according to the perturbation adjustment coefficients at each time step to obtain a reconstructed driving perturbation field; and superimposing the reconstructed driving perturbation field onto the control forecast of the convective-scale ensemble forecast system to obtain the reconstructed driving field of the convective-scale ensemble forecast. This method not only improves the efficiency and accuracy of reconstructing and adjusting large-scale driving fields in convective-scale ensemble forecasts but also plays an important role in alleviating the problem of highly dispersed large-scale information in global ensemble forecast systems.

Description

Technical Field

[0001] This invention relates to the field of atmospheric forecasting technology, and in particular to a method for reconstructing and adjusting large-scale driving fields in convective-scale ensemble forecasts. Background Technology

[0002] With the intensification of global climate change, extreme weather events such as torrential rains and severe convection occur frequently, causing serious impacts on human society and the natural environment. Accurate prediction of these extreme weather events is of great significance for disaster prevention and mitigation and for protecting people's lives and property. As an advanced weather forecasting tool, convective-scale ensemble forecasting systems can provide forecast information with high spatiotemporal resolution, which helps to improve the accuracy of extreme weather event forecasts.

[0003] Traditional convective-scale ensemble forecasting systems typically rely on driving fields provided by global ensemble forecasting systems, including initial fields and lateral boundary conditions. While this approach can provide some large-scale information, the uncertainty and resolution limitations of global ensemble forecasts restrict the quality of large-scale forecast uncertainty information in convective-scale ensemble forecasting systems, resulting in large dispersion and an inability to effectively focus on forecast uncertainties related to local convection. Furthermore, traditional dynamic downscaling methods fail to adequately consider the differences in uncertainty between different forecasting models when processing driving perturbation fields, leading to insufficient reliability of forecast results. To overcome these shortcomings, this invention proposes a method for reconstructing and adjusting the large-scale driving field of convective-scale ensemble forecasts. This method combines the global ensemble driving field with higher-quality global deterministic forecasts to reconstruct the background field and lateral boundary conditions of the convective-scale ensemble forecast system. Adjustments are made to the original large-scale ensemble perturbations based on the changes in the quality of the updated background field and lateral boundary forecasts. This further enables the convective-scale ensemble forecast system to focus on the uncertainties in small and medium-scale forecasts. This not only provides a reference for improving the background field and lateral boundary conditions of convective-scale ensemble forecasts but also alleviates the problem of the relatively dispersed large-scale information in the global ensemble forecast system. Summary of the Invention

[0004] The purpose of this invention is to provide a method for reconstructing and adjusting large-scale driving fields in convective-scale ensemble forecasts.

[0005] To achieve the above objectives, the present invention is implemented according to the following technical solution:

[0006] This invention includes the following steps:

[0007] Dynamic downscaling of the driving field of the global ensemble forecast system is performed to determine the driving perturbation field, and dynamic downscaling of the EC fine-grid deterministic model forecast field is performed to obtain the control forecast of the convective-scale ensemble forecast system; the driving field includes the initial field and the lateral boundary.

[0008] Obtain forecast element values ​​from the EC fine-grid deterministic model and the global ensemble forecast system control forecast within the forecast lead time, perform multi-sphere screening on the forecast element values, and calculate forecast statistics;

[0009] Based on the forecast statistics, polynomial fitting is performed to obtain the perturbation adjustment coefficients at the analysis time and the perturbation adjustment coefficients at each forecast time for the convective-scale ensemble forecast system; the perturbation adjustment coefficients at the analysis time are related to the initial field; the perturbation adjustment coefficients at the forecast time are related to the lateral boundary.

[0010] The driving disturbance field is adjusted according to the disturbance adjustment coefficient at the analysis time and the disturbance adjustment coefficient at each forecast time to obtain the reconstructed driving disturbance field. The reconstructed driving disturbance field is then superimposed on the control forecast of the convective-scale ensemble forecast system to obtain the convective-scale ensemble reconstructed driving field.

[0011] Furthermore, the method for determining the driving disturbance field includes:

[0012] The driving disturbance field includes the initial disturbance field and the disturbance side boundary;

[0013] The initial field of the global ensemble forecast system at the analysis time is dynamically downscaled to obtain the initial field of the convective-scale model. The initial field of the convective-scale model is then subtracted from the forecast field of the control forecast at the analysis time in the global ensemble forecast system to obtain the initial field of the perturbation.

[0014] The lateral boundaries of the global ensemble forecast system at each forecast time are obtained by dynamic downscaling the lateral boundaries of the convective-scale model at each forecast time. The perturbation lateral boundaries are obtained by subtracting the lateral boundaries of the convective-scale model at each forecast time from the forecast fields of the control forecast at each forecast time in the global ensemble forecast system.

[0015] Furthermore, the method for obtaining control forecasts for a convective-scale ensemble forecast system includes:

[0016] The control forecast of the convective-scale ensemble forecast system includes the initial value field of the control forecast of the convective-scale ensemble forecast system and the lateral boundary of the control forecast of the convective-scale ensemble forecast system.

[0017] The initial field forecast dynamics at the time of analysis of the EC fine-grid deterministic model are downscaled to the same simulation area and resolution as the time of analysis of the convective-scale ensemble forecast, so as to obtain the initial field of the control forecast of the convective-scale ensemble forecast system.

[0018] The lateral boundary forecast dynamics of the EC fine-grid deterministic model at each forecast time are downscaled to the simulation area and resolution of the convective-scale ensemble forecast at each forecast time, thus obtaining the control forecast lateral boundary of the convective-scale ensemble forecast system.

[0019] Furthermore, the method for calculating the forecast statistics includes:

[0020] Obtain forecast element values ​​from the EC fine-grid deterministic model and the convective-scale ensemble forecast system control forecast within the forecast lead time; the forecast element values ​​include zonal wind U, meridional wind V, temperature T, and geopotential height H; the forecast element values ​​include Each atmospheric pressure level;

[0021] A dynamic adaptive multisphere is constructed based on historical forecast element values. Multisphere filtering is then applied to forecast element values ​​within the forecast lead time. The expression is as follows:

[0022]

[0023]

[0024]

[0025] in As a filtering condition, it means time kind Hierarchical forecast element values With the Forecast element vector of historical samples Euclidean distance Less than time kind Hierarchical forecast element radius And the forecast element values and Vector of forecast elements Euclidean distance Time function mapping With time decay weight The weighted time cumulative value is greater than the empirical threshold. , For time sensitivity, For the first Historical time of each forecast element vector For the dimension of forecast element values, The number of levels for the same forecast element value. For dimension weights, For hierarchical weights, For the first Vector of forecast elements correspond kind Hierarchical forecast element values, Forecast element vector exist Standard deviation of forecast element values ​​at different levels Forecast element vector of Standard deviation of forecast element values ​​at different levels for kind The radius of the standard forecast element at the level, As a regulating factor, for time Laplace operator for atmospheric pressure field;

[0026] The selected forecast element values ​​are replaced by the average values ​​of the forecast elements at the upper and lower levels of the same category. The root mean square error of different forecast elements controlled by the EC fine grid deterministic model and the convective-scale ensemble forecast system is calculated for each forecast period.

[0027] Calculate the root mean square error ratios of different forecast elements for the EC fine-grid deterministic model and the convective-scale ensemble forecast system control forecasts within each forecast period. Sort the root mean square error ratios of the same forecast element at each forecast time in ascending order to obtain the confidence interval and median of the root mean square error at each forecast time for the same forecast element. Take the median of the root mean square error of the same forecast element at different forecast times as the forecast statistics value.

[0028] Furthermore, the method for obtaining the time-perturbation adjustment coefficients of the convective-scale ensemble forecast system and the time-perturbation adjustment coefficients for each forecast time includes:

[0029] Based on the same forecast element Different forecast times Forecast statistics Perform polynomial fitting to determine the adjustment coefficients, expressed as:

[0030]

[0031] in Forecast elements Adjustment coefficient, , , These are the polynomial coefficients;

[0032] Take all forecast elements at the same forecast time The average adjustment factor is the current forecast time. The adjustment factor of the disturbance adjustment factor, when hour, To analyze the time-period disturbance adjustment coefficient, when hour, For each forecast time The disturbance adjustment coefficient.

[0033] Furthermore, the method for obtaining the reconstructed driving field from the convective scale set includes:

[0034] The reconstruction-driven perturbation field includes a reconstruction perturbation initial field and a reconstruction perturbation side boundary; the flow-scale set reconstruction-driven field includes a reconstruction initial field and a reconstruction side boundary;

[0035] The reconstructed initial perturbation field is obtained by using the perturbation adjustment coefficient and the initial perturbation field at the analysis time. This reconstructed initial perturbation field is then superimposed onto the initial control forecast field of the convective-scale ensemble forecast system to obtain the reconstructed initial field, expressed as:

[0036]

[0037] in To analyze the members of the time-scale ensemble forecasting system Reconstruct the initial field, To analyze the initial field of the control forecast of the convective-scale ensemble forecast system at a given time, For analyzing time members The initial value field of the reconstructed perturbation;

[0038] The reconstructed perturbation side boundary is obtained by using the perturbation adjustment coefficients and perturbation side boundaries at each forecast time. This reconstructed perturbation side boundary is then superimposed onto the control forecast side boundary of the convective-scale ensemble forecast system to obtain the reconstructed side boundary, expressed as:

[0039]

[0040] in for Members of the Time-Scale Convection Ensemble Forecasting System The reconstructed side boundary, for The time-scale ensemble forecasting system controls the forecast lateral boundary. for Moment Members The reconstructed perturbation side boundary.

[0041] The beneficial effects of this invention are:

[0042] This invention relates to a method for reconstructing and adjusting large-scale driving fields in traverse-scale ensemble forecasts. Compared with existing technologies, this invention has the following technical advantages:

[0043] This invention, through dynamic downscaling, multi-sphere screening, data statistics, polynomial fitting, and perturbation adjustment and superposition steps, can improve data preprocessing capabilities and enhance model adaptability in the reconstruction and adjustment of large-scale driving fields in convective-scale ensemble forecasts. This improves the efficiency and accuracy of the reconstruction and adjustment of large-scale driving fields in convective-scale ensemble forecasts. Optimizing the reconstruction and adjustment technology of large-scale driving fields in convective-scale ensemble forecasts can greatly save resources and improve work efficiency. It can realize the reconstruction and adjustment of large-scale driving fields in convective-scale ensemble forecasts, providing a reference for improving the background field and lateral boundary conditions of convective-scale ensemble forecasts, thereby alleviating the problem of relatively scattered large-scale information in global ensemble forecast systems. Attached Figure Description

[0044] Figure 1 This is a flowchart illustrating the steps of the method for reconstructing and adjusting large-scale driving fields in convective-scale ensemble forecasting according to the present invention.

[0045] Figure 2 This is a diagram showing the disturbance adjustment coefficients calculated within the forecast period for a certain region using the reconstructed adjustment method for large-scale driving fields in convective-scale ensemble forecasts of this invention.

[0046] Figure 3 This invention provides a precipitation forecast map for a specific region at a specific time, obtained by using the convective-scale ensemble forecasting large-scale driving field reconstruction and adjustment method to adjust the driving field.

[0047] Figure 4 This is a map showing the actual precipitation in a certain region at a certain time. Detailed Implementation

[0048] The present invention will be further described below through specific embodiments. The illustrative embodiments and descriptions herein are used to explain the present invention, but are not intended to limit the present invention.

[0049] The present invention provides a method and system for reconstructing and adjusting large-scale driving fields in convective-scale ensemble forecasting, comprising the following steps:

[0050] like Figure 1 As shown, this embodiment includes the following steps:

[0051] Dynamic downscaling of the driving field of the global ensemble forecast system is performed to determine the driving perturbation field, and dynamic downscaling of the EC fine-grid deterministic model forecast field is performed to obtain the control forecast of the convective-scale ensemble forecast system; the driving field includes the initial field and the lateral boundary.

[0052] Obtain forecast element values ​​from the EC fine-grid deterministic model and the global ensemble forecast system control forecast within the forecast lead time, perform multi-sphere screening on the forecast element values, and calculate forecast statistics;

[0053] Based on the forecast statistics, polynomial fitting is performed to obtain the perturbation adjustment coefficients at the analysis time and the perturbation adjustment coefficients at each forecast time for the convective-scale ensemble forecast system; the perturbation adjustment coefficients at the analysis time are related to the initial field; the perturbation adjustment coefficients at the forecast time are related to the lateral boundary.

[0054] The driving disturbance field is adjusted according to the disturbance adjustment coefficient at the analysis time and the disturbance adjustment coefficient at each forecast time to obtain the reconstructed driving disturbance field. The reconstructed driving disturbance field is then superimposed on the control forecast of the convective-scale ensemble forecast system to obtain the convective-scale ensemble reconstructed driving field.

[0055] In this embodiment, the method for determining the driving disturbance field includes:

[0056] The driving disturbance field includes the initial disturbance field and the disturbance side boundary;

[0057] The initial field of the global ensemble forecast system at the analysis time is dynamically downscaled to obtain the initial field of the convective-scale model. The initial field of the convective-scale model is then subtracted from the forecast field of the control forecast at the analysis time in the global ensemble forecast system to obtain the initial field of the perturbation.

[0058] The lateral boundaries of the global ensemble forecast system at each forecast time are obtained by dynamic downscaling the lateral boundaries of the convective-scale model at each forecast time. The perturbation lateral boundaries are obtained by subtracting the lateral boundaries of the convective-scale model at each forecast time from the forecast fields of the control forecast at each forecast time in the global ensemble forecast system.

[0059] In this embodiment, the method for obtaining control forecasts of a convective-scale ensemble forecast system includes:

[0060] The control forecast of the convective-scale ensemble forecast system includes the initial value field of the control forecast of the convective-scale ensemble forecast system and the lateral boundary of the control forecast of the convective-scale ensemble forecast system.

[0061] The initial field forecast dynamics at the time of analysis of the EC fine-grid deterministic model are downscaled to the same simulation area and resolution as the time of analysis of the convective-scale ensemble forecast, so as to obtain the initial field of the control forecast of the convective-scale ensemble forecast system.

[0062] The lateral boundary forecast dynamics of the EC fine-grid deterministic model at each forecast time are downscaled to the simulation area and resolution of the convective-scale ensemble forecast at each forecast time, thus obtaining the control forecast lateral boundary of the convective-scale ensemble forecast system.

[0063] In this embodiment, the method for calculating forecast statistics includes:

[0064] Obtain forecast element values ​​from the EC fine-grid deterministic model and the convective-scale ensemble forecast system control forecast within the forecast lead time; the forecast element values ​​include zonal wind U, meridional wind V, temperature T, and geopotential height H; the forecast element values ​​include Each atmospheric pressure level;

[0065] A dynamic adaptive multisphere is constructed based on historical forecast element values. Multisphere filtering is then applied to forecast element values ​​within the forecast lead time. The expression is as follows:

[0066]

[0067]

[0068]

[0069] in As a filtering condition, it means time kind Hierarchical forecast element values With the Forecast element vector of historical samples Euclidean distance Less than time kind Hierarchical forecast element radius And the forecast element values and Vector of forecast elements Euclidean distance Time function mapping With time decay weight The weighted time cumulative value is greater than the empirical threshold. , For time sensitivity, For the first Historical time of each forecast element vector For the dimension of forecast element values, The number of levels for the same forecast element value. For dimension weights, For hierarchical weights, For the first Vector of forecast elements correspond kind Hierarchical forecast element values, Forecast element vector exist Standard deviation of forecast element values ​​at different levels Forecast element vector of Standard deviation of forecast element values ​​at different levels for kind The radius of the standard forecast element at the level, As a regulating factor, for time Laplace operator for atmospheric pressure field;

[0070] The selected forecast element values ​​are replaced by the average values ​​of the forecast elements at the upper and lower levels of the same category. The root mean square error of different forecast elements controlled by the EC fine grid deterministic model and the convective-scale ensemble forecast system is calculated for each forecast period.

[0071] Calculate the root mean square error ratios of different forecast elements for the EC fine-grid deterministic model and the convective-scale ensemble forecast system control forecasts within each forecast period. Sort the root mean square error ratios of the same forecast element at each forecast time in ascending order to obtain the confidence interval and median of the root mean square error at each forecast time for the same forecast element. Take the median of the root mean square error of the same forecast element at different forecast times as the forecast statistics value.

[0072] In this embodiment, the method for obtaining the perturbation adjustment coefficients at the analysis time and the perturbation adjustment coefficients at each forecast time of the convective-scale ensemble forecast system includes:

[0073] Based on the same forecast element Different forecast times Forecast statistics Perform polynomial fitting to determine the adjustment coefficients, expressed as:

[0074]

[0075] in Forecast elements Adjustment coefficient, , , These are the polynomial coefficients;

[0076] Take all forecast elements at the same forecast time The average adjustment factor is the current forecast time. The adjustment factor of the disturbance adjustment factor, when hour, To analyze the time-period disturbance adjustment coefficient, when hour, For each forecast time The disturbance adjustment coefficient.

[0077] In this embodiment, the method for obtaining the reconstructed driving field of the convective scale set includes:

[0078] The reconstruction-driven perturbation field includes a reconstruction perturbation initial field and a reconstruction perturbation side boundary; the flow-scale set reconstruction-driven field includes a reconstruction initial field and a reconstruction side boundary;

[0079] The reconstructed initial perturbation field is obtained by using the perturbation adjustment coefficient and the initial perturbation field at the analysis time. This reconstructed initial perturbation field is then superimposed onto the initial control forecast field of the convective-scale ensemble forecast system to obtain the reconstructed initial field, expressed as:

[0080]

[0081] in To analyze the members of the time-scale ensemble forecasting system Reconstruct the initial field, To analyze the initial field of the control forecast of the convective-scale ensemble forecast system at a given time, For analyzing time members The initial value field of the reconstructed perturbation;

[0082] The reconstructed perturbation side boundary is obtained by using the perturbation adjustment coefficients and perturbation side boundaries at each forecast time. This reconstructed perturbation side boundary is then superimposed onto the control forecast side boundary of the convective-scale ensemble forecast system to obtain the reconstructed side boundary, expressed as:

[0083]

[0084] in for Members of the Time-Scale Convection Ensemble Forecasting System The reconstructed side boundary, for The time-scale ensemble forecasting system controls the forecast lateral boundary. for Moment Members The reconstructed perturbation side boundary.

[0085] In actual assessment, taking the reconstruction and adjustment of the large-scale driving field of the convective-scale ensemble forecast in XX region as an example, the convective-scale ensemble forecast system adopts the WRF model XXX version, with the model region set to a horizontal resolution of 3km, the simulation region range of 35.5°~46.3°N, 105.2°~122.4°E, a total of 550×424 grid points covering XX region, and a vertical hierarchy of 59 model surfaces, with the model layer top at 10hPa. The system includes one control forecast and 20 perturbation member forecasts, totaling 21 ensemble members. The system forecasts twice a day, at 00:00 UTC and 12:00 UTC, with a forecast lead time of 48h, and outputs forecast results hourly.

[0086] All member physical process parameterization schemes are set as follows: Thompson microphysics scheme, Mellor-Yamada-Janjic (MYJ) boundary layer scheme and RRTMG long and short wave radiation scheme, and the cumulus convection parameterization scheme is turned off;

[0087] Dynamic downscaling of the driving field of the global ensemble forecast system was performed to determine the driving perturbation field, and dynamic downscaling of the EC fine-grid deterministic model forecast field was performed to obtain the control forecast of the convective-scale ensemble forecast system. The horizontal resolution of the global ensemble forecast system was 0.5°×0.5°, and the horizontal resolution of the EC fine-grid deterministic model was 0.25°×0.25°.

[0088] Forecast element values ​​for the 12-60 h forecast field were obtained from the EC fine-grid deterministic model and the convective-scale ensemble forecast system control forecast. The forecast element categories included zonal wind (U), meridional wind (V), temperature (T), and geopotential height (H). The statistical hierarchy was divided into 10 levels based on atmospheric pressure: 10 / 50 / 100 / 200 / 250 / 500 / 700 / 850 / 925 / 1000 hPa. The statistical period was 6-hourly forecasts starting from 00 UTC on June 21-28, 2022. The number of root mean square error ratio (RMSE) samples (taking the 18-hour forecast field as an example) is 4 (forecast element categories) * 10 (statistical levels) * 8 (statistical cases) = 320. After obtaining 320 RMSE samples, they are arranged in ascending order to obtain the confidence intervals and medians of the RMSE at different forecast times. By performing polynomial fitting on the median of the RMSE, the perturbation adjustment coefficients for the convective-scale ensemble forecast system analysis time (initial conditions) and each forecast time (lateral boundary conditions) are obtained as follows. Figure 2 As shown;

[0089] The initial value of the disturbance field and the adjustment factor for multiplying the disturbance side boundary by the corresponding time step. The adjusted initial value field and reconstructed side boundary of the disturbance can be obtained. These are then superimposed onto the initial value field and side boundary of the convective-scale ensemble forecast system control forecast to obtain the reconstructed initial value field and reconstructed side boundary. This is then achieved through... Figure 3 Precipitation forecast map Figure 4 The comparison of actual precipitation maps demonstrates the accuracy of the precipitation probability forecast after the reconstruction and adjustment of the large-scale driving field in the convective-scale ensemble forecast.

[0090] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for reconstructing and adjusting large-scale driving fields in convective-scale ensemble forecasts, characterized in that, Includes the following steps: S1. Dynamically downscale the driving field of the global ensemble forecast system to determine the driving perturbation field, and dynamically downscale the forecast field of the EC fine-grid deterministic model to obtain the control forecast of the convective-scale ensemble forecast system; the driving field includes the initial field and the lateral boundary. S2. Obtain the forecast element values ​​of the EC fine-grid deterministic model and the global ensemble forecast system control forecast within the forecast lead time, perform multi-sphere screening on the forecast element values ​​and calculate the forecast statistics; S3. Based on the forecast statistics, perform polynomial fitting to obtain the perturbation adjustment coefficients for the analysis time and the perturbation adjustment coefficients for each forecast time in the convective-scale ensemble forecast system; the perturbation adjustment coefficients for the analysis time are related to the initial field; the perturbation adjustment coefficients for the forecast time are related to the lateral boundary. S4. Adjust the driving disturbance field according to the disturbance adjustment coefficient at the analysis time and the disturbance adjustment coefficient at each forecast time to obtain the reconstructed driving disturbance field. Superimpose the reconstructed driving disturbance field onto the control forecast of the convective-scale ensemble forecast system to obtain the convective-scale ensemble reconstructed driving field. The method for obtaining the reconstructed driving field of the convective scale set includes: The reconstructed driving disturbance field includes the reconstructed disturbance initial value field and the reconstructed disturbance side boundary; The flow-scale set reconstruction driving field includes the initial reconstruction field and the reconstruction side boundary; The reconstructed initial perturbation field is obtained by using the perturbation adjustment coefficient and the initial perturbation field at the analysis time. This reconstructed initial perturbation field is then superimposed onto the initial control forecast field of the convective-scale ensemble forecast system to obtain the reconstructed initial field, expressed as: in To analyze the members of the time-scale ensemble forecasting system Reconstruct the initial field, To analyze the initial field of the control forecast of the convective-scale ensemble forecast system at a given time, For analysis of time members The initial value field of the reconstructed perturbation; The reconstructed perturbation side boundary is obtained by using the perturbation adjustment coefficients and perturbation side boundaries at each forecast time. This reconstructed perturbation side boundary is then superimposed onto the control forecast side boundary of the convective-scale ensemble forecast system to obtain the reconstructed side boundary, expressed as: in for Members of the Time-Scale Convection Ensemble Forecasting System The reconstructed side boundary, for The time-scale ensemble forecasting system controls the forecast lateral boundary. for Moment Members The reconstructed perturbation side boundary.

2. The method for reconstructing and adjusting large-scale driving fields based on convective-scale ensemble forecasts according to claim 1, characterized in that, The method for determining the driving disturbance field includes: The driving disturbance field includes the initial disturbance field and the disturbance side boundary; The initial field of the global ensemble forecast system at the analysis time is dynamically downscaled to obtain the initial field of the convective-scale model. The initial field of the convective-scale model is then subtracted from the forecast field of the control forecast at the analysis time in the global ensemble forecast system to obtain the initial field of the perturbation. The lateral boundaries of the global ensemble forecast system at each forecast time are obtained by dynamic downscaling the lateral boundaries of the convective-scale model at each forecast time. The perturbation lateral boundaries are obtained by subtracting the lateral boundaries of the convective-scale model at each forecast time from the forecast fields of the control forecast at each forecast time in the global ensemble forecast system.

3. The method for reconstructing and adjusting large-scale driving fields based on convective-scale ensemble forecasts according to claim 1, characterized in that, The method for obtaining control forecasts for a convective-scale ensemble forecasting system includes: The control forecast of the convective-scale ensemble forecast system includes the initial value field of the control forecast of the convective-scale ensemble forecast system and the lateral boundary of the control forecast of the convective-scale ensemble forecast system. The initial field forecast dynamics at the time of analysis of the EC fine-grid deterministic model are downscaled to the same simulation area and resolution as the time of analysis of the convective-scale ensemble forecast, so as to obtain the initial field of the control forecast of the convective-scale ensemble forecast system. The lateral boundary forecast dynamics of the EC fine-grid deterministic model at each forecast time are downscaled to the simulation area and resolution of the convective-scale ensemble forecast at each forecast time, thus obtaining the control forecast lateral boundary of the convective-scale ensemble forecast system.

4. The method for reconstructing and adjusting large-scale driving fields based on convective-scale ensemble forecasts according to claim 1, characterized in that, The method for calculating the forecast statistics includes: Obtain forecast element values ​​from the EC fine-grid deterministic model and the convective-scale ensemble forecast system control forecast within the forecast lead time; the forecast element values ​​include zonal wind U, meridional wind V, temperature T, and geopotential height H; the forecast element values ​​include Each atmospheric pressure level; A dynamic adaptive multisphere is constructed based on historical forecast element values. Multisphere filtering is then applied to forecast element values ​​within the forecast lead time. The expression is as follows: in As a filtering condition, it means time kind Hierarchical forecast element values With the Forecast element vector of historical samples Euclidean distance Less than time kind Hierarchical forecast element radius And the forecast element values and Vector of forecast elements Euclidean distance Time function mapping With time decay weight The weighted time cumulative value is greater than the empirical threshold. , For time sensitivity, For the first Historical time of each forecast element vector For the dimension of forecast element values, The number of levels for the same forecast element value. For dimension weights, For hierarchical weights, For the first Vector of forecast elements correspond kind Hierarchical forecast element values, Forecast element vector exist Standard deviation of forecast element values ​​at different levels Forecast element vector of Standard deviation of forecast element values ​​at different levels for kind The radius of the standard forecast element at the level, As a regulating factor, for time Laplace operator for atmospheric pressure field; The selected forecast element values ​​are replaced by the average values ​​of the forecast elements at the upper and lower levels of the same category. The root mean square error of different forecast elements controlled by the EC fine grid deterministic model and the convective-scale ensemble forecast system is calculated for each forecast period. Calculate the root mean square error ratios of different forecast elements for the EC fine-grid deterministic model and the convective-scale ensemble forecast system control forecasts within each forecast period. Sort the root mean square error ratios of the same forecast element at each forecast time in ascending order to obtain the confidence interval and median of the root mean square error at each forecast time for the same forecast element. Take the median of the root mean square error of the same forecast element at different forecast times as the forecast statistics value.

5. The method for reconstructing and adjusting large-scale driving fields based on convective-scale ensemble forecasts according to claim 1, characterized in that, The method for obtaining the time-based perturbation adjustment coefficients and the time-based perturbation adjustment coefficients of the convective-scale ensemble forecast system includes: Based on the same forecast element Different forecast times Forecast statistics Perform polynomial fitting to determine the adjustment coefficients, expressed as: in Forecast elements Adjustment coefficient, , , These are the polynomial coefficients; Take all forecast elements at the same forecast time The average adjustment factor is the current forecast time. The adjustment factor of the disturbance adjustment factor, when hour, To analyze the time-period disturbance adjustment coefficient, when hour, For each forecast time The disturbance adjustment coefficient.

Images