Search

Short-term precipitation forecasting is essential for agriculture, transportation, urban management, and tourism. The radar echo extrapolation method is widely used in precipitation forecasting. To address issues like forecast degradation, insufficient capture of spatiotemporal dependencies, and low accuracy in radar echo extrapolation, we propose a new model: MS-DD3D-RSTN. This model employs spatiotemporal convolutional blocks (STCBs) as spatiotemporal feature extractors and uses the spatial-temporal loss (STLoss) function to learn intra-frame and inter-frame changes for end-to-end training, thereby capturing the spatiotemporal dependencies in radar echo signals. Experiments on the Sichuan dataset and the HKO-7 dataset show that the proposed model outperforms advanced models in terms of CSI and POD evaluation metrics. For 2 h forecasts with 20 dBZ and 30 dBZ reflectivity thresholds, the CSI metrics reached 0.538, 0.386, 0.485, and 0.198, respectively, representing the best levels among existing methods. The experiments demonstrate that the MS-DD3D-RSTN model enhances the ability to capture spatiotemporal dependencies, mitigates forecast degradation, and further improves radar echo prediction performance.

The difference between reflectivity factors observed by different radars observing the same target affects the performance of the multi-radar mosaic. For X-band radars without beam blocking, this difference is mainly caused by calibration deviation and signal attenuation due to rain and wet radome. To correct these errors, a new correction algorithm for network phased array weather radars is designed. In this new algorithm, we first regarded the attenuation due to water film on the radome as part of the calibration deviation, and used the network attenuation correction algorithm to make an initial attenuation correction. Then, an observation deviation function of the radar network is designed, which used the gradient descent method to solve the calibration deviation issue between different radars. Finally, a secondary attenuation correction, which used the network attenuation correction algorithm, is applied to the original reflectivity factor after calibration deviation correction. In this study, a network consisting of seven X-band phased array weather radars in the Foshan city of Guangdong province is used to verify the new correction algorithm. For a rainfall event associated with cold air, a correlation coefficient of 0.53 between the reflectivity factor corrected with the traditional path-integrated attenuation (PIA) algorithm and the reflectivity factor observed by an S-band weather radar in Guangzhou was obtained, with a root mean square error of 9.0 dB. While those between the new correction algorithm and the S-band radar are 0.64 and 8.4 dB, respectively, with both being better than PIA. A similar performance was also achieved in a typhoon outer rain-band rainfall event analysis.

Accurate wind speed prediction is significantly important for the full utilization of wind energy resources and the improvement in the economic benefits of wind farms. Because the ensemble forecast takes into account the uncertainty of information about the atmospheric motion, domestic and foreign weather service forecast centers often choose to use the ensemble numerical forecast to achieve the fine forecast of wind speed. However, due to the unavoidable systematic errors of the ensemble numerical forecast model, it is necessary to correct the deviation in the ensemble numerical forecast wind speed. Considering the typical spatio-temporal characteristics of the grid prediction data of the wind field, based on Convolutional Long–Short Term Memory (ConvLSTM) units and attention mechanism, this paper takes the complex and representative North China region as the research area, aiming to reveal the shortcomings of existing deep learning integrated prediction correction models in extracting temporal features of grid prediction data. We propose a new ensemble prediction wind field correction model integrating multi-factor and spatio-temporal characteristics. This model uses reanalyzed land data provided by the European Center for Medium-Range Weather Forecasts as the real data to correct the deviation in the near-surface 10 m wind field data predicted by the regional ensemble numerical prediction model of the China Meteorological Administration. We used the reanalyzed land data provided by the European Centre for Medium-Range Weather Forecasts (ECMWF) as the live data to correct the deviation in the near-surface 10 m wind field data predicted by the regional ensemble numerical forecast model of the China Meteorological Administration (CMA). At the same time, Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) were used as the scoring indicators, and the results of the China Meteorological Administration–Regional Ensemble Prediction System (CMA–REPS) ensemble average, multiple linear regression method correction, Long–Short Term Memory (LSTM) method correction, and U-net (UNET) method correction were compared. Compared with the UNET model method, the experimental results show that when processing the 10 m zonal wind data, 10 m meridional wind data, and 10 m average wind speed data of CMA–REPS 24 h forecasts, the correction results of our model can reduce the RMSE score index by 9.15%, 4.83%, and 7.79%. At the same time, when processing the 48 h and 72 h near-surface 10 m wind field data of the CMA–REPS forecast, our model can improve the prediction accuracy of CMA–REPS near-surface wind forecast data. Therefore, the correction effect of the proposed model in a complex terrain area is evidently better compared to other methods.

During 29–30 UTC June 2013, a severe rainfall event with a long and narrow region of strong precipitation occurred in the central of the Sichuan Basin (SCB). Under the combined influence of a Tibetan Plateau vortex (TPV) and a southwest vortex (SWV), two banded strong radar echoes existed and developed simultaneously over the SCB. The analysis reveals that the vertical wind shear (VWS) caused by the circulations of the TPV and the SWV was the dominant factor of the formation and development of the radar echoes over the SCB. During the coupling period of the two vortices, the SWV provided abundant water vapor at the middle and lower levels over the SCB and the updrafts of the two vortices break through that formed deep convection, which made the precipitation in the SCB reach the maximum intensity. The enhancement of horizontal vorticity caused by the baroclinicity and the secondary circulation related to the two vortices created conditions for the formation of the double-band radar echoes. The matching degree of water vapor and heating conditions accompanying the circulation of the two vortices could affect the developments of convective storms and precipitation.

Global precipitation measurement (GPM) is one of the effective means employed to observe orographic precipitation, and its inversed GPM DPR data can be used to study the microphysical structure of precipitation particles. This study considers statistics on convective precipitation (CP) and stratiform precipitation (SP) events over three types of terrain (plains, mountains, and high mountains) using the DPR onboard the GPM Core Observatory from May to September of 2014–2021 to analyze the vertical structure of heavy CP and SP. In mountain areas and high mountain areas, the updraft rendered by topography or seeder-feeder mechanism is not only conducive to the collision and merger of raindrops into large raindrops, but also increases the concentration of small drops, which is the main reason why the occurrence probability of not only large but also small raindrops increases and the horizontal distribution domain of mass weighted average raindrop diameter (Dm) widens. For heavy SP, the occurrence probability of medium-diameter precipitation particles below the freezing height (FzH) over high mountains is greater than those over plains. The precipitation particles above 10 km altitude of high mountains have characteristics, such as lower droplet number concentration and larger diameter, compared with those over plains. Furthermore, the study also analyzed the correlation between storm top altitude (STA) and Dm, water vapor and STA respectively. This study is helpful to further understand the effect of topography on heavy precipitation through cloud microphysical processes and the vertical structure of precipitation.

In this study, horizontal wind (U and V), air temperature (T), and relative humidity (RH) modelled by the European Centre for Medium-Range Weather Forecasts Reanalysis Interim (ERA-Interim), the National Aeronautics and Space Administration (NASA) Modern Era Retrospective Analysis for Research and Applications (MERRA), the Japanese 55-year Reanalysis (JRA-55), the National Centers for Environmental Prediction (NCEP) Climate Forecast System Version 2 (CFSv2), and the NCEP Final Operational Global Analysis data and the NCEP Final Operational Global Analysis data (NCEP-FNL) products have been compared with observations at 11 radiosonde stations over the eastern and downstream regions of the Tibetan Plateau (TP) from late June until the end of July during 2011 to 2015. The mean bias of all variables for the five gridded datasets (GDs) in the Sichuan Basin (SCB) is larger than that for the TP. The mean values of U, V, and T from each grid dataset are generally consistent with the radiosonde values, whereas considerable bias in the mean RH exists at upper levels. The diurnal variation of the mean bias and root-mean-square (RMS) error in the basin are stronger than those in the TP and the negative/positive peak usually occurs at 06:00 UTC and 18:00 UTC in the basin or at 12:00 UTC in the TP. The inter-annual variations in the basin are significantly stronger, and the maximum values of the variations usually occur at upper levels or near the surface, except for V. The weather conditions have a crucial influence on the performance of the gridded datasets. The mean bias and RMS error of T in the TP on cloudy days are obviously larger than those during sunny conditions. Considerable but unsteady differences occur in the mean bias and RMS error of U and V in different weather conditions. On average, the four variables in the TP are more sensible to the weather conditions.