WMOD00383

• Title: Simulation of Winter Precipitation in Western Mountain Watersheds with a Local-Scale Model
• Date: 11/1/1992
• Weather Modification - Doc Type: Report

..

-~

estimates at the coarse scale of 200-km or greater spacing (Washington
and Meehl, 1989), a nesting of models has been adopted to resolve
requirements at the local scale. This approach involves use of a GCM's
output to initialize a RSM (regional-scale model) whose output is then
used to initialize a LSPM (local-scale orographic precipitation model)
that finally estimates precipitation at the watershed level. This study
has addressed, primarily, the latter component of the nesting, namely,
the estimation of grid-point precipitation at the watershed scale of 5
to 10 km when the LSPM is initialized with RSM generated soundings.
Eventually, precipitation estimates from the LSPM will serve as input to
hydrologic models to resolve the impacts of watershed characteristics on
runoff.

The RSM role has come about because of the wide gap between the
GCM's resolution and that of the LSPM. Neither the current GCMs nor
. LSPMs can properly simulate at the intermediate resolution. Filling the
gap are the RSMs that use grid-point spacing on ehe order of 50 to 100
km.(Giorgi and Bates, 1989). These models can better simulate effects
of smaller-scale weather perturbations and the terrain than GCMs.

The RSMs generate sounding information at each of their grid
points. For a particular watershed, representative grid points are
selected and their sounding data employed as input to the LSPM.

The GCM and RSM models selected for contribution to the Global
Climate Change Response Program are the NCAR CCM (National Center for
Atmospheric Research; Community Climate Model) currently at about 500-km
resolution (Williamson et al., 1987) and the Pennsylvania State
University/NCAR MM4 model (Anthes and Warner, 1978; Anthes et al., 1987;
Deardorff, 1972; Zhang and Anthes, 1982; 'Anthes, 1977). The MM4 is a
three-dimensional, time-dependent model configured with a grid-point
spacing of 60 km. Both models include a surface physics/soil hydrology
module for modeling soil hydrologic budgets (Dickinson et al.,1986).
Sounding parameter basic information was developed in MM4 runs by NCAR
scientists, Giorgi and Bates, and will be discussed in Giorgi et al.
(1992). Matthews et al. (1992a, 1992b) accessed output data archives
and reconditioned the data for eventual interpolation to LSPM needs.

The LSPM selected is an orographic model particularly adept. at
simulating mountain winter precipitation. The model was originally
developed by Rhea (1977) to model winter orographic precipitation in the
Colorado mountains. Medina (1991) reported preliminary results from the
adaptation and testing of this model for the Salt-Verde watersheds of
Arizona. Winter precipitation over the Atlas Mountains of Morocco was
simulated with the model (Majdoub et al., 1989). These applications
suggested that perhaps the model could be adapted to simulate winter
precipitation in weak orographic conditions. Medina (1992) reported
limited success in modeling Delaware watershed precipitation.

This study covers (1) application of. the LSPM to the Gunnison
watershed employing soundings collected during the 1980s at GJT (Grand
Junction), SLC (Salt Lake City), and INW (Winslow), to tune the model
for that watershed, (2) application to the Gunnison with soundings
derived from MM4, (3) comparison of model estimates of daily and monthly
precipitation with high-elevation gauge measurements, and (4) generation
of simulated daily precipitation distributions for comparison with gauge
value distributions. The fourth element is important to climate studies