Search

Data-driven machine learning technology can learn and extract features, a factor which is well recognized to be powerful in the warning and prediction of severe weather. With the large-scale deployment of the radar wind profile (RWP) observational network in China, dynamical variables with higher temporal and spatial resolution in the vertical become strong supports for machine-learning-based severe convection prediction. Based on the RWP mesonet that has been deployed in Beijing, this study uses the measurements from four triangles composed of six RWP stations to determine the profiles of divergence, vorticity, and vertical velocity before rainfall onsets. These dynamic feature variables, combined with cloud properties from Himawari-8 and ERA-5 reanalysis, serve as key input parameters for two rainfall forecast models based on the random forest (RF) classification algorithm. One is for the rainfall/non-rainfall forecast and another for the rainfall grade forecast. The roles of dynamic features such as divergence, vorticity, and vertical velocity are examined from ERA-5 reanalysis data and RWP measurements. The contribution of each feature variable to the performance of the RF model in independent tests is also discussed here. The results show that the usage of RWP observational data as the RF model input tends to result in better performance in rainfall/non-rainfall forecast 30 min in advance of rainfall onset than using the ERA-5 data as inputs. For the rainfall grade forecast, the divergence and vorticity that were estimated from the RWP measurements at 800 hPa show importance in improving the model performance in heavy and moderate rain forecasts. This indicates that the atmospheric dynamic variable measurements from RWP have great potential to improve the prediction skill of convection with the aid of a machine learning model. [ABSTRACT FROM AUTHOR]

Currently, most deep learning (DL)-based models for precipitation forecasting face two conspicuous issues: the smoothing effect in the precipitation field and the degenerate effect of forecasting precipitation intensity. Therefore, this study proposes "time series residual convolution (TSRC)", a DL-based convolutional neural network for precipitation nowcasting over China with a lead time of 3 h. The core idea of TSRC is it compensates the current local cues with previous local cues during convolution processes, so more contextual information and less uncertain features would remain in deep networks. We use four years' radar echo reflectivity data from 2017 to 2020 for model training and one year's data from 2021 for model testing and compare it with two commonly used nowcasting models: optical flow model (OF) and UNet. Results show that TSRC obtains better forecasting performances than OF and UNet with a relatively high probability of detection (POD), low false alarm rate (FAR), small mean absolute error (MAE) and high structural similarity index (SSIM), especially at longer lead times. Meanwhile, the results of two case studies suggest that TSRC still introduces smoothing effects and slightly outperforms UNet at longer lead times. The most considerable result is that our model can forecast high-intensity radar echoes even for typhoon rainfall systems, suggesting that the degenerate effect of forecasting precipitation intensity can be improved by our model. Future works will focus on the combination of multi-source data and the design of the model's architecture to gain further improvements in precipitation short-term forecasting. [ABSTRACT FROM AUTHOR]

This work evaluates and compares the performance of the radar reflectivity and lightning data assimilation schemes implemented in weather research and forecasting-four-dimensional data assimilation (WRF-FDDA) for short-term precipitation and lightning forecasts. All six mesoscale convective systems (MCSs) with a duration greater than seven hours that occurred in the Guangdong Province of China during June 2020 were included in the experiments. The results show that both the radar reflectivity data assimilation and lightning data assimilation improved the analyses and short-term forecasts of the precipitation and lightning of the MCSs. On average, for precipitation forecasts, the experiments with radar reflectivity data assimilation performed better than those with lightning data assimilation; however, for lightning forecasts, the experiments with lightning data assimilation performed better in the analysis period and 1 h forecast, and for some cases, the superiority lasted to three forecast hours. This highlights the potential of lightning data assimilation in short-term lightning forecasting. [ABSTRACT FROM AUTHOR]

This study quantitatively examines the contribution of assimilating observations in the regions with different dynamic instabilities to the analysis and prediction of an extreme rainstorm event in Fujian Province of China. The wind profiling radar (WPR) observations are classified into two groups, i.e., strong and weak instability areas (SIA and WIA), according to their local dynamic instability identified by the ensemble spread. Their performance of assimilation and prediction in terms of the wind and precipitation are evaluated and compared in detail. The results show that the wind analysis error by assimilating all of the WPR observations can be reduced by about 30%. In particular, the wind analysis errors by only assimilating the observations in the SIA are about 12% lower than those in the WIA. They are related to the existence of the low-level horizontal wind shear with strong instability in the SIA. The case study shows that the assimilation of observations in the SIA can effectively correct the wind fields on the two sides of the wind shear line, producing an improved precipitation forecast compared to observation assimilation in the WIA. [ABSTRACT FROM AUTHOR]

Based on the Weather Research and Forecasting (WRF) Model and the three-dimensional variational (3DVAR) data assimilation system, this study investigates the effects of assimilation of radar reflectivity and radial velocity under different momentum control variables on the forecast of Southwest China Vortex precipitation. It is shown that the U−V control variable strengthens the wind speed and vorticity to be better matching the observation, while using ψ−χ as the control variable will produce too large increments which are unphysical. The root mean square errors (RMSE) of radar radial velocity are around 2.4 m/s in the experiment using ψ−χ control variables, while the RMSE are below 2 m/s in the experiment with U−V control variables. The composite reflectivity from the analysis of the U−V control variables matches better with the observation than that from the analysis of the ψ−χ control variables, i.e., the forecast rain band location under U−V control variables is more accurate. ψ−χ control variable enhances the cold high-pressure system in near surface, while the U−V control variable is not significant. The water vapor flux convergence in the lower layers of the ψ−χ control variable is overestimated leading excessive precipitation in the forecast. The Equitable Threat Score (ETS) of the U−V control variable is about 0.1 higher than ψ−χ control variable. In summary, the U−V control variable is superior to the ψ−χ control variable in terms of analysis and forecasting about Southwest China Vortex precipitation. [ABSTRACT FROM AUTHOR]