Evaluation and Calibration of Short-range PQPF From Multi-model and Time-phased Ensembles During Hmt-west-2006

Abstract

The NOAA Hydrometeorological Testbed (HMT)West 2006 program focuses on short-range (0-12 h) ensemble forecasts over the American River Basin (ARB) in California. Several high-resolution (1 km and 3 km) mesoscale models are operated by the NOAA/ESRL/Global Systems Division during December 2005- March 2006. The model initialization is provided by the Local Analysis and Prediction System (LAPS) and boundary conditions are from coarse resolution of NAM and RUC forecasts. Short-range (1 h, 3 h, 6 h) probabilistic quantitative precipitation forecasts (PQPF) from time-phased (initialized per 3 h) and multi-model (RAMS, MM5, WRF) ensembles are assessed for a number of intensive operation periods (IOP). Uncertainties in PQPF associated with models, initial time, boundary conditions, as well as resolution are analyzed. The forecast skill and quality of PQPF also depends on observation/analysis data, which are from the available HMT quantitative precipitation estimation (QPE) products, such as gauge measurements and the NCEP Stage IV precipitation analyses. Based on the archived model forecasts and bias characteristics, the preliminary bias correction of PQPF is conducted using an artificial neural network and linear regression methods.