The Multiyear Surface Climatology of a Regional Atmospheric Model over the Western United States

This paper presents a validation analysis of the climatology of a version of the National Center for Atmospheric Research-Pennsylvania State University limited-area model (MM4) developed for application to regional climate simulation over the western United States. Two continuous multiyear simulations, for the periods 1 January 1982–31 December 1983 and 1 January 1988–25 April 1989, were performed over this region with the MM4 driven by ECMWF analyses of observations and run at a horizontal resolution of 60 km. The model used in these simulations includes horizontal diffusion on terrain-following σ coordinates, a Kuo-type cumulus parameterization, sophisticated radiative transfer and surface physics-soil hydrology packages, and a relaxation boundary- conditions procedure. Model-produced surface air temperatures, precipitation, and snow depths were compared with observations from about 390 stations distributed throughout the western United States. The base-model run reproduced the seasonal cycle of temperature and precipitation well. Monthly and seasonal temperature biases were generally less than a few degrees. The effects of topography on the regional distribution of precipitation were also well reproduced, although local detail was in several instances poorly captured. When regionally averaged, absolute model-precipitation biases were mostly in the range of 10% –50% of observations. The model generally simulated precipitation better in the cold season than in the warm season, and over coastal regions than in the continental interior. The simulated seasonal cycles of snowpack formation and melting were realistic, although modeled and observed snow-depth values differed significantly locally. Over the Rocky Mountain regions the model reproduced wintertime precipitation amounts well but overpredicted summertime precipitation. Because of this overprediction of summertime precipitation, a version of the model was also tested in a simulation from 30 May 1988–26 December 1988 in which the horizontal diffusion coefficients were reduced over topographical gradients and an inflow-outflow lateral boundary condition was implemented for water vapor. These modifications were found to provide an improved simulation of summer precipitation while not substantially altering wintertime precipitation. This work shows that it is feasible to perform good quality, multiyear simulations with current limited-area models and, therefore, that it is feasible to apply such models to climate studies.