Improving Short‐Term Analyses and Forecasts of an EF2 Tornadic Supercell Through EnKF X‐Band Phased Array Radar Assimilation

Abstract This study investigates the impact of assimilating X‐band Phased Array Radar (XPAR) data on the analysis and forecast of an EF2 tornadic supercell in China using a deterministic forecast‐updated Ensemble Kalman Filter method that avoids the potential smoothing of fine‐scale storm details associated with ensemble averaging. Observations indicate that XPAR provides a more continuous spatiotemporal sampling of the low‐level mesocyclone and echo structure than operational S‐band radar. Assimilating reflectivity (RF) and radial velocity (RV) from the S‐band radar fail to reproduce the hook echo and key low‐level rotational signatures, with substantial deviations in the predicted supercell and the low‐level vortex track. Joint assimilation of XPAR RF and RV enables earlier vortex analysis and improves the three‐dimensional storm structure, particularly strengthening the coupling between the mesocyclone and hook echo, resulting in better agreement between the simulated storm and observations. The inclusion of XPAR RV is critical for resolving low‐level rotation, which is poorly represented when assimilating RF alone. Higher‐frequency assimilation of XPAR data further refines echo and wind field analyses, producing a more compact vortex and intense rotating updraft. Consequently, it improves probabilistic predictions of potential tornado paths inferred from updraft helicity—a well‐established proxy for tornado potential, given that tornado‐scale vortices cannot be explicitly resolved at the 0.5‐km grid spacing. In contrast, experiments without XPAR RV or with lower assimilation frequencies exhibit limited skill. These results highlight the value of the XPAR network for both analyzing and forecasting small‐scale convective systems, suggesting its potential to complement operational weather radars.