Simulating convective mode of mesoscale phenomena with a WRF-GEFS ensemble

Numerical weather prediction models can often struggle to correctly simulate convective mesoscale phenomena, even at a grid spacing of less than 4 km. A single deterministic simulation of a mesoscale phenomenon can incorrectly capture convective mode and observed features due to imperfect initial conditions or deficiencies in model physics. In particular, bow echoes are poorly reproduced. An ensemble of simulations, each with various permutations of model physics parameterizations, can address the issue of imperfect model physics; however, there has been relatively little research using ensembles with perturbed initial conditions to address incorrectly-forecast convective mode. This study describes how an 11-member Weather Research and Forecasting model ensemble, initialized and constrained with Global Ensemble Forecast System model data, can be used to simulate two different bow-echo cases (an isolated bow echo and line-echo wave-pattern) centered in the Great Plains region, USA. This method's advantages (spread of solutions; potential predictability in different synoptic scenarios) and failings (risk of underdispersion; low-resolution initial-condition data) are discussed.