WMOD00584 - Laserfiche WebLink

Home Browse Search

events. This, paper: (1) outlines a nested modeling approach to examine precipitation over the Rocky Mountains in Colorado, (2) presents an assessment of a regional modelr s precipitation simulations in complex terrain, and (3) shows an example from a sophisticated local-scale model. The paper, identifies the need for a nested regional/local-scale modeling system to accurately simulate local-scale precipitation processes over complex terrain. Reclamation's GCCRP modeling approach involves two phases. Phase 1 evaluates the nested regional/local-scale modeling method in current climate to determine its capability to accurately simulate existing conditions in well-documented cases. In phase 1, the regional model has been initialized by observed data sets and run in a climate simulation mode for 40 months. As confidence in the large-scale, general circulation model results improves, phase 2 will apply the nested modeling approach to future climate simulations that include effects such as doubled C02. These simulations will use a general circulation model, a regional model, and a local-scale model. In phase 2, the general circulation model will provide synoptic-scale information that initializes the regional model, which will simulate the regional evolution of storm structure. Then a local-scale model will use the reg~onal analyses for its initial conditions and simulate local~scale precipitation with high-resolution (dx=-5 to 10 km) topography. This paper focuses on research in phase 1 - the validation of the regional model and early stages of testing a sophisticated local-scale, t~me- dependent model. THE NESTED MODELING APPROACH - Phase 1 In phase 1, the nested modeling approach uses a regional model to provide the regional-scale forcing that largely controls the evolution of local-scale clouds and precipitation that is simulated by a local- scale model. The regional model is the MM4 (Pennsylvania State University/NCAR Mesoscale Model Version 4). The standard MM4 model is described in Anthes et a1. (1987) ; however, in this study we use results from an augmented version for regional climate studies described by Giorgi and Bates (1989). This version of the MM4 model includes a sophisticated surface physics/soil hydrology package. The model is a compressible and hydrostatic model with equations of motion written in terrain-varying sigma vertical coordinates. Figure 1 shows thethree-dimens ional terrain surface from MM4' s 3000- by 3000-km domain having 60-km grid spacing over the Western United States. Note that MM4' s terrain fails to clearly define the Colorado River Basin, hence the need for higher resolution. The 10-km terrain resolution (see Figure 1) ,in the local-scale Clark model resolves major drainages on the Gunnison River. Generally, increased resolution of terrain is a major factor that controls the detailed structure (spatial distribution and intensity) of local orographic precipitation. Matthews et al. (1991) found that MM4 simulated regional controls of precipitation reasonably well in ]\,rizona during January 1979. Matthews et a1., (1992) found similar good performance in the Gunnison region in the 1982-83 period examined by Medina (1992) using the Rhea model.