Degree Type


Date of Award


Degree Name

Master of Science



First Advisor

Raymond W. Arritt


We examine seasonal forecasts of winter precipitation over the continental United States produced using six regional climate models (RCMs) to downscale a global model. Dynamic downscaling offers higher spatial resolution to better resolve surface features such as topography than the coarser resolution global model. This may allow for a better representation of precipitation over topographically varying terrain such as the western United States and better resolve mesoscale features such as local wind circulations.

The RCMs downscaled the global NCEP Climate Forecast System (CFS) version 1 model from 1982-2003 as part of the Multi-RCM Ensemble Downscaling (MRED) project. An ensemble of forecasts was created by initializing the models with ten different start dates. We assess the models performance in the January-February-March (JFM) and February-March-April (FMA) seasons over the Interior West (IW), Southwest (SW), Gulf Coast (GC) and Ohio Valley (OH). The latter three regions have empirical wintertime El Niño-Southern Oscillation (ENSO) precipitation influences.

We employ the statistical gamma distribution to examine the daily precipitation frequency and intensity over the two seasons. Specifically, our focus is on the mean of the gamma distribution, which is the product of the shape (α) and scale (β) parameters. We find that the downscaled RCMs are able to represent the influence of ENSO by a change in the mean of the gamma distribution between the 22 years of analysis and El Niño and La Niña years over the SW, GC and OH. The change in the mean is generally a result of the gamma distribution parameters α and β both increasing or decreasing. In the IW, the α and β parameters change inversely. Despite no known ENSO influence over this region, we find that there is a difference in the mean of the gamma distribution, particularly in FMA.


Copyright Owner

Andrew James Ansorge



File Format


File Size

96 pages