Tests of precipitation parameterizations available in latest version of NCAR regional climate model (RegCM) over continental United States

Citation
F. Giorgi et C. Shields, Tests of precipitation parameterizations available in latest version of NCAR regional climate model (RegCM) over continental United States, J GEO RES-A, 104(D6), 1999, pp. 6353-6375
Citations number
31
Categorie Soggetti
Earth Sciences
Volume
104
Issue
D6
Year of publication
1999
Pages
6353 - 6375
Database
ISI
SICI code
Abstract
This paper presents a series of tests of resolvable scale precipitation and cumulus convection schemes available in the latest version of the NCAR reg ional climate model (RegCM). For resolvable scale precipitation we use a ne w simplified explicit moisture (SIMEX) scheme, while for cumulus convection , we test the mass-flux scheme of Grell [1993], the Kuo-type scheme of Anth es [1977], and the newly implemented scheme of Zhang and McFarlane [1995]. The period of simulation is 3 years long, from March 1993 to February 1996, and the domain covers the continental United States at 60 km resolution. W hen using the SIMEX and Grell schemes, the model shows the best performance , with generally good simulation of precipitation, surface air temperature, and large-scale circulations throughout the United States. Both averages a nd anomalies associated with individual seasons are well reproduced at the regional scale. Main model deficiences are an underestimate of cold season precipitation over the eastern United States and a displacement of the summ er precipitation maximum over the central plains. The Kuo scheme has a perf ormance more similar to that of the Grell scheme than that of the Zhang and McFarlane (ZMF) scheme. Use of the ZMF scheme yields a reasonably good sim ulation of regionally averaged precipitation, but deteriorates some aspect of the simulation such as the intensity of the summer low-level jet over th e central plains. We also present sensitivity tests to various parameters i n the schemes which can be of guidance to the model user for optimizing the model performance over different regions.