Abstract:Based on the characteristics of sudden and local, short life history, serious disasters and so on, the severe convection weather system is difficult to be captured by the conventional meteorological observation network, and is still challenging for catastrophic weather forecasting. In order to improve the service ability in strong weather monitoring and prediction, the following researches have been carried out recently: (1) The new mesocyclone and tornado vortex feature recognition algorithms are developed and proved to be successfully in identifying tornado vortex characteristics in more than a dozen tornado cases. Extracted from Doppler radar volume scan data, a number of parameters (exceed thirty) have been used in the research on the automatic recognition and warning technology of classified severe convective weather (downburst, tornado, hail and short-time strong precipitation). Based on large sample data and results of a variety of analysis methods, a thunderstorm winds Bayes discriminant model has also been established. The testing results show that its Heidke skill score is 0.836, along with the accuracy rate and hit rate are greater than 95%, and the empty rate is below 5%. (2) Rapid update cycle forecast system can effectively improve the quality of model initial values that is very suitable for short time forecast application. For the sake of improving severe thunderstorm prediction, a novel pseudo-observation and assimilation approach involving water vapor mass mixing ratio is proposed to better initialize numerical weather prediction (NWP) at convection-resolving scales. In addition, a new set of simplified and parameterized dual-polarization radar simulators for horizontal reflectivity (ZH), differential reflectivity (ZDR), specific difference phase (KDP), and correlation coefficient (&#961;HV) have been co-developed, and some preliminary data assimilation experiments have shown that the assimilation of dual polarization variables including differential reflectivity and specific difference phase in addition to radar radial velocity and horizontal reflectivity can help improve the accuracy of initial conditions for model hydrometer variables and ensuing model forecasts. (3) Although not yet mature enough for meteorological application, blending technology which is expected to overcome the deficiency of the quantitative precipitation (QPF) by a mesoscale NWP model for the short term at convective scales and the rapidly descending skill of rainfall forecast based on radar extrapolation method beyond the first few hours is under development and debugging, and also has potential in enhancing the ability of rainfall forecast within the nowcasting period. (4) The above methods and systems were applied and provided technical support for meteorological services during the 7th Wuhan World Military Games in 2019, and a good service effect had been achieved.

A Four-Dimensional Variational System for Skillful Operational Prediction of Convective Storms

Potential of 4d-VAR for exigent forecasting of severe weather

Lightning-Fast Convective Outlooks: Predicting Severe Convective Environments with Global AI-based Weather Models

A Multiscale Four-Dimensional Variational Data Assimilation Scheme: A Squall-Line Case Study

Evaluation of an experimental Warn‐on‐Forecast 3DVAR analysis and forecast system on quasi‐real‐time short‐term forecasts of high‐impact weather events

Assimilating FY-4A Lightning and Radar Data for Improving Short-Term Forecasts of a High-Impact Convective Event with a Dual-Resolution Hybrid 3DEnVAR Method

FengWu-4DVar: Coupling the Data-driven Weather Forecasting Model with 4D Variational Assimilation

4D-Var using Hessian approximation and backpropagation applied to automatically-differentiable numerical and machine learning models

Sub-seasonal forecasting with a large ensemble of deep-learning weather prediction models

A Three-dimensional Convolutional-Recurrent Network for Convective Storm Nowcasting

Advances in the Study of Severe Convection Weather Nowcasting in Central China

A Structured Graph Neural Network for Improving the Numerical Weather Prediction of Rainfall

Towards a Self-contained Data-driven Global Weather Forecasting Framework

Convolutional Neural Network for Convective Storm Nowcasting Using 3D Doppler Weather Radar Data

Inferring Thunderstorm Occurrence from Vertical Profiles of Convection-Permitting Simulations: Physical Insights from a Physical Deep Learning Model

Accurate initial field estimation for weather forecasting with a variational constrained neural network

Convolutional Neural Network for Convective Storm Nowcasting Using 3-D Doppler Weather Radar Data

Cool-Season Evaluation of FV3-LAM-Based CONUS-Scale Forecasts with Physics Configurations of Experimental RRFS Ensembles

A Four-Dimensional Ensemble-Variational (4DEnVar) Data Assimilation System Based on GRAPES-GFS: System Description and Primary Tests

Long-range weather forecasts through numerical and empirical methods