Assimilation of Radar Radial Velocity and Polarimetric Observations Using LETKF within MPAS-JEDI: Development of Assimilation Capabilities and Test with an Afternoon Thunderstorm in Taiwan

Abstract

A new parameterized polarimetric radar operator (P-PRO) capable of simulating reflectivity ( Z H ), specific differential phase ( K DP ), and differential reflectivity ( Z DR ) is implemented in the Joint Effort for Data assimilation Integration (JEDI), extending its capability to assimilate polarimetric radar observations. The existing radial-velocity ( v r ) observation operator is combined with P-PRO to enable joint assimilation of multiple radar variables. To evaluate the operator and assess the impacts of each variable, data assimilation experiments for a Taiwan afternoon thunderstorm progressively assimilate Z H , v r , K DP , and Z DR using an MPAS-JEDI LETKF configuration. Each assimilated variable primarily improves its corresponding field at analysis time, while the inclusion of v r and K DP leads to more coherent and better-organized convective structures than reflectivity-only assimilation. Statistical diagnostics indicate systematic improvements relative to NoDA across cycles and enhanced short-range forecast skill, with the clearest reductions in forecast error occurring for directly assimilated variables during the first few forecast hours. K DP and Z DR assimilation further contributes to more observation-consistent polarimetric signatures and improved analysis-time storm structure; some forecast improvements are evident, but they are not robustly separable given the single-case design and the limited number of cycles. Overall, these results validate the P-PRO implementation in JEDI and highlight the potential of polarimetric radar assimilation for improving convective-scale analyses, providing a foundation for future multi-case assessments of forecast impacts.