The grid nudging technique is often used in regional climate dynamical downscaling to make the simulated large-scale fields consistent with the driving fields. In this study, we focused on two specific questions about grid nudging: (1) which nudged variable has a larger impact on the downscaling results; and (2) what is the “optimal” grid nudging strategy for each nudged variable to achieve better downscaling result during summer over the mainland China. To solve these queries, 41 three-month-long simulations for the summer of 2009 and 2010 were performed using the Weather Research and Forecasting model (WRF) to downscale National Centers for Environmental Prediction (NCEP) Final Operational Global Analysis (FNL) data to a 30-km horizontal resolution. The results showed that nudging horizontal wind or temperature had significant influence on the simulation of almost all conventional meteorological elements, while nudging water vapor mainly affected the precipitation, humidity, and 500 hPa temperature. As a whole, the optimal nudging time was one hour or three hours for nudging wind, three hours for nudging temperature, and one hour for nudging water vapor. The optimal nudged level was above the planetary boundary layer for almost every nudged variable. Despite these findings, it should be noted that the optimum nudging scheme varied with simulated regions and layers, and dedicated research for different regions, seasons, and model configuration is advisable.

The microphysical parameterization scheme employed in four-dimensional variational data assimilation (4D-Var) plays an important role in the assimilation of humidity and cloud-sensitive observations. In this study, a newly developed full-hydrometeor assimilation scheme, integrating warm-rain and cold-cloud processes, has been implemented in the Weather Research and Forecasting (WRF) 4D-Var system. This scheme is based on the WRF single-moment 6-class microphysics scheme (WSM6). Its primary objective is to directly assimilate radar reflectivity observations, with the goal of evaluating its effects on model initialization and subsequent forecasting performance. Four assimilation experiments were conducted to assess the performance of the full-hydrometeor assimilation scheme against the warm-rain assimilation scheme. These experiments also investigated reflectivity assimilation using both indirect and direct methods. We found that the nonlinearity of the radar operator in the two direct reflectivity assimilation experiments requires more iterations for cost function reduction than in the indirect assimilation method. The hydrometeor fields were reasonably analyzed using the full-hydrometeor assimilation scheme, particularly improving the simulation of ice-phase hydrometeors and reflectivity above the melting layer. The assimilation of radar reflectivity led to improvements in short-term (0–3 h) precipitation forecasting with the full-hydrometeor assimilation scheme. Assimilation experiments across multiple case studies reaffirmed that assimilating radar reflectivity observations with the full-hydrometeor assimilation scheme accelerated model spinup and yielded enhancements in 0–3-h total accumulate precipitation forecasts for a range of precipitation thresholds.

Northeast China experiences significant rainfall events driven by the Northeast China Cold Vortex (NCCV) during the warm season. However, the microphysical processes associated with this precipitation require further investigation. This study aims to fill this gap by comparing the microphysical characteristics of local stratiform and convective rain under the influence of the NCCV (NCCV precipitation or rain) and by exploring the underlying meteorological mechanisms. The results indicate that both stratiform and convective NCCV precipitation in Northeast China are characterized by sparser raindrops compared to similar rainfall types in other regions of China. In addition, convective NCCV rain consists of much larger raindrops than any other types of rainfall across China, leading to rain rates over 10 times greater than those of stratiform NCCV precipitation. The accelerated generation and growth of raindrops reinforce both stratiform and convective NCCV precipitation, whereas convective NCCV rain exhibits a more complex intensifying mechanism. If the rain rate (R) ≤ 9 mm h⁻¹, raindrops are sparse but large. With the increase in precipitation intensity, raindrops become denser but also collide and break up into smaller particles. When R > 16 mm h⁻¹, dense and large raindrops prevail, deriving from the melting of numerous large solid hydrometeors formed through deep convection. The convection is stimulated by higher mid- to low-level temperatures and greater convective available potential energy, causing clouds to ascend to higher altitudes with faster upward movement, thereby fostering the formation of larger raindrops. Additionally, convective NCCV precipitation is usually enhanced by divergence induced by upper-level jet streams.