The utility of convection-permitting ensembles for the prediction of stationary convective bands

A.I. Barrett, S.L. Gray, D.J. Kirshbaum, N.M. Roberts, D.M. Schultz, J.G. Fairman

Research output: Contribution to journalArticlepeer-review

Abstract

This study examines convection-permitting numerical simulations of four cases of terrain-locked quasi-stationary convective bands over the United Kingdom. For each case, a 2.2-km-grid-length, 12-member ensemble and a 1.5-km-grid-length deterministic forecast are analyzed, each with two different initialization times. Object-based verification is applied to determine whether the simulations capture the structure, location, timing, intensity, and duration of the observed precipitation. These verification diagnostics reveal that the forecast skill varies greatly between the four cases. Although the deterministic and ensemble simulations captured some aspects of the precipitation correctly in each case, they never simultaneously captured all of them satisfactorily. In general, the models predicted banded precipitation accumulations at approximately the correct time and location, but the precipitating structures were more cellular and less persistent than the coherent quasi-stationary bands that were observed. Ensemble simulations from the two different initialization times were not significantly different, which suggests a potential benefit of time-lagging subsequent ensembles to increase ensemble size. The predictive skill of the upstream larger-scale flow conditions and the simulated precipitation on the convection-permitting grids were strongly correlated, which suggests that more accurate forecasts from the parent ensemble should improve the performance of the convection-permitting ensemble nested within it.
Original languageEnglish
JournalMonthly Weather Review
Early online date26 Feb 2016
DOIs
Publication statusPublished - 2016

Fingerprint

Dive into the research topics of 'The utility of convection-permitting ensembles for the prediction of stationary convective bands'. Together they form a unique fingerprint.

Cite this