Radar data assimilation using WRF-VAR: a comparison of 4DVAR with 3DVAR for an IHOP case

In recent years, the potential of assimilating radar observations for the improvement of short-term prediction of convective weather has been demonstrated using various techniques. In particular, four-dimensional techniques such as 4DVAR have been shown to be able to capture convective-scale dynamics through the retrieval of the unobserved variables, resulting in improved forecasting of convective systems. 4DVAR has been expected by many to be the operational system not only for the large scale but also for the high-resolution cloud-permitting models.

WRF-VAR is a variational data assimilation system developed at NCAR for the WRF (Weather Research and Forecasting) model. WRF-VAR includes both 3DVAR and 4DVAR capabilities. WRF 3DVAR was applied to the assimilation of radar data and demonstrated some benefit for forecasting convective weather systems. A study is being conducted to examine the ability of WRF 4DVAR in assimilating radar observations using a convective storm case observed during IHOP (International H2O Project). The benefit of 4DVAR in convective-scale data assimilation is demonstrated by comparing the two techniques. Detailed analysis is performed to show why the 4DVAR outperforms the 3DVAR. Specifically, the cold pool structure and low-level convergence from the two techniques will be compared to shed some light on the key physical processes that lead to the improved forecasts by the 4DVAR initialization.