Impact of Assimilating High‐Resolution Atmospheric Motion Vectors on Convective Scale Short‐Term Forecasts: 1. Observing System Simulation Experiment (OSSE)

Abstract

AbstractThis research investigates to what extent the high‐spatiotemporal‐resolution Atmospheric Motion Vector (AMV) product derived from the new‐generation Geostationary Operational Environmental Satellites‐Series R can benefit convective‐scale data assimilation (DA) and forecasts of potential high impact weather events. In the first part of this two‐part study, the impact of AMV DA on convective‐scale numerical weather prediction (NWP) is evaluated with an idealized supercell storm. The simulated AMV observations are synthesized from the idealized supercell storm generated by the Weather Research and Forecasting model and then the data are assimilated with a three‐dimensional variational analysis and forecast system. A baseline DA experiment demonstrates that the wind errors within and around the storms are remarkably reduced by assimilating the AMV data, consequently enhancing the storm top‐level divergence and low‐level convergence signatures associated with two strong splitting supercells. Three sets of sensitivity experiments are then performed to test the impact of observation resolution, DA cycling frequency and horizontal correlation length scale respectively. Generally, assimilating higher‐spatial‐resolution AMVs at higher cycling frequency is able to produce reasonable divergence analysis as well as the subsequent 90 min forecasts. However, too frequent DA cycling tends to produce a warm bias and overestimate nonprecipitating hydrometeor and storm‐relative helicity features, resulting in more spurious cells in the short‐term forecasts. It is also found that the correlation length scale of 20 km produces the best results when assimilating the AMV data.