Assessing the Impact of the Valid-Time-Shifting Method on the EnVar Analysis of the Storm Environment for Convection Initiation Cases

Abstract The valid-time-shifting (VTS) method was implemented to increase the background ensemble size for assimilating conventional in situ observations within a convection-allowing ensemble-variational (EnVar) analysis and forecast system. The focus of the study is to examine the impact of VTS on the analysis of the storm environment for convection initiation (CI) cases. The VTS method increases the background ensemble size cost-effectively by shifting the background ensemble initialized at the same time but valid at different times. Three CI cases in May 2019 were chosen. Three sets of experiments were designed to reveal the impact of VTS, all compared to the same 36-member ensemble experiments without VTS (noVTS-36). The first two sets of VTS experiments triples the background ensembles to 108 (VTS36x3) and triples the background members to 36 on a reduced-base 12-member EnVar (VTS12x3). The third VTS experiment is designed to have the same computational cost as noVTS-36 (VTS24x3). Objective results from all three cases show that VTS12x3 achieves similar analysis accuracy and statistically indistinguishable conventional observation variable forecast errors and reflectivity forecast skill for light/medium thresholds compared to noVTS-36, at 73% of the computational cost. Tripling the base ensemble size (VTS36x3) achieves the highest forecast improvement throughout the 9-h evaluation period compared to noVTS-36, at a fraction of the added costs (89%). With a computational cost nearly the same as noVTS-36, VTS24x3 achieves improved environmental analyses, significantly lower forecast error of conventional observation variables, and noteworthy subjective forecast improvements with slight objective skill improvements at light and medium reflectivity thresholds.