A Comparison of Multiscale GSI-Based EnKF and 3DVar Data Assimilation Using Radar and Conventional Observations for Midlatitude Convective-Scale Precipitation Forecasts

• DOI: 10.1175/mwr-d-14-00345.1
• Publication Date: 2015-04-11T00:22:41Z
• Authors: Aaron Johnson; Xuguang Wang; Jacob R. Carley; Louis J. Wicker; Christopher Karstens
• Keywords: 13. Climate action; 01 natural sciences; 0105 earth and related environmental sciences

Abstract A GSI-based data assimilation (DA) system, including three-dimensional variational (3DVar) and ensemble Kalman filter (EnKF), is extended to the multiscale of both meso- synoptic-scale observation networks convective-scale radar reflectivity velocity observations. EnKF 3DVar are systematically compared in this context better understand impacts differences between DA techniques on analyses at multiple scales subsequent precipitation forecasts. Averaged over 10 diverse cases, 8-h forecasts initialized using more skillful than those 3DVar, with without storm-scale DA. The advantage from persists for ~5 h EnKF, but only ~1 3DVar. case study an upscale growing MCS also examined. EnKF-initialized forecast attributed accurate mesoscale environment features. location structure a warm front accurately analyzed Furthermore, storms analysis maintained during period. However, not generate excessive cold pools. Therefore, while remains throughout period, quality degraded by after first hour. Diagnostics revealed that inferior mesoscales storm primarily lack flow dependence cross-variable correlation, respectively, static background error covariance.