Climate change and GCM simulation of water resources from mountain snowpac

This research develops a scale translation methodology that statistically relates atmospheric circulation to daily snowfall. Combined with validation of the GENESIS General Circulation Model (GCM), daily synoptic-scale atmospheric circulation is used to predict snowfall over the Colorado Plateau. Linear regression methods are compared with non-linear neural networks in their ability to model snowfall from a synoptic circulation index; the neural nets are chosen to develop transfer functions between circulation and daily snowfall. The neural networks explain over 70% of the daily snowfall variance.;GCMs are notoriously questionable for use in regional climate change impact studies, but this scale translation methodology allows well-modeled GCM features to be used rather than relying on direct GCM output. This research specifically uses the neural network transfer functions to predict daily snowfall amounts for five subregions of the Colorado Plateau in an effort to model the primary water source for the Colorado River system. Improved snowfall modelling and analysis of potential snowfall in a doubled $rm COsb2$ atmosphere are important in the effort to reduce regional vulnerability to climate change. This social vulnerability includes both physical and social relationships within a context of human dimensions of global change.;This study finds that total snowfall on the Plateau should change little in a doubled $rm COsb2$ world based on circulation alone; however, further analysis suggests a 30% reduction in snowfall. Therefore, it is likely that the water resource in the Colorado River system will change under climate change conditions and social vulnerability will increase.