Hmt-west Lessons Learned—ii: Use of Met For Evaluation and Diagnostics During Forecast Experiments

Abstract

Evaluation of research models in a real-time forecast environment can be a critical component to the research-to-operations transition. The Developmental Testbed Center (DTC) conducted evaluations of several U.S. operational models and an ESRL- based research regional ensemble modeling system during the 2010-2012 HydroMeteorology Testbed (HMT)-West winter exercises. These evaluations have been largely focused on QPF during extreme precipitation events in the California Sierras and coastal ranges. Results presented in this paper will focus on traditional and spatial verification techniques used during the three seasons of the HMT-West experiment to diagnose individual model performance and to make inter-model comparisons. For example, these methods have been employed to examine the impact of ensemble design decisions including choices for initial conditions, spatial resolution, dynamic core, and microphysics. We will also demonstrate the usefulness of specifically ensemble-related verification methods and display products recently included in the Model Evaluation Tools (MET) developed at the DTC. Among these products are spread-skill and performance diagrams.