Correcting Storm Displacement Errors in Ensembles Using the Feature Alignment Technique (FAT)

A goal of Warn-on-Forecast (WoF) is to develop forecasting systems that produce accurate analyses and forecasts of severe weather, such as supercells, for forecasters to utilize in operational warning settings. Recent WoF-related studies have indicated the need to alleviate storm displacement errors in both storm-scale analyses and short-term forecasts. One promising method to reduce these errors is the feature alignment technique (FAT). The FAT mitigates displacement errors between observations and a model field while weakly satisfying user-defined constraints (e.g., smoothness) to form a 2-D field of displacement vectors, which are used to adjust the prior model fields to more closely match the observations. Previous studies merging FAT with variational data assimilation systems have shown substantial improvement of analyses and forecasts, especially at earlier forecast times, by minimizing displacement errors. However, the WoF project mostly employs variants of the ensemble Kalman filter (EnKF) data assimilation technique to produce analyses. Therefore, this study merged the FAT with the CM1-LETKF (local ensemble transform Kalman filter) system and used observation system simulation experiments (OSSEs) to vet the FAT as a potential alleviator of forecast errors arising from storm displacement errors.

An idealized Truth run of a supercell on a 250-m grid is used to generate pseudo-radar observations (i.e., reflectivity and radial velocity). The pseudo-observations are assimilated by the CM1-LETKF system using a 50-member ensemble to produce analyses and forecasts of the supercell on a 2-km grid. The FAT uses the composite reflectivity field to generate a 2-D field of displacement vectors and applies the vectors to the model state variables (at all model levels) at the start of each LETKF analysis cycle. The FAT-LETKF system is tested by displacing the initial model background fields from the observations and/or modifying the environmental wind profile to create a storm motion bias in the forecast cycles. The FAT-LETKF performance is evaluated using the Truth run and compared to that of the LETKF alone (i.e., no correction of displacement errors). The results from these tests reveal the beneficial impact the FAT-LETKF system may have in cases of storm motion bias resulting from mesoscale analysis or model errors. These supercell OSSEs provide the foundation for future FAT-EnKF experiments with real data and/or mesoscale convective systems.