Assessment of Climate Change over Colorado River Basin as predicted by Regional Climate Models

The 2000s western U.S. drought has focused attention on drought susceptibility of the Colorado River Basin (CRB). There is a concern that many climate models predict permanently drier conditions for the next century over the CRB, however interpretation of these projections is complicated by their coarse spatial resolution, which does not resolve the role of the relatively small mountain headwaters area that is the source of much of the basin's runoff. Regional climate models (RCMs) are able to resolve these spatial scales, and for this reason should be a preferred source of information about the future hydrology of the CRB. In this study, we use the RCM simulations from the North American Regional Climate Change Assessment Program (NARCCAP) to evaluate implications for the future discharge of the Colorado River. Within the analysis, the historical simulation of RCMs and their host global General Circulation Models (GCMs) are compared to 1/8-degree historical North American Land Data Assimilation System (NLDAS) data over the CRB. Future climate change scenarios simulated by RCMs from 2040-2069 under the A2 SRES emission scenario are compared to the historical period from 1970-1999. These changes are also compared to changes predicted by the host GCMs over these same periods. We found that RCMs capture the primary features of observations and more closely reproduce the NLDAS estimates than do the host GCMs. The RCMs resolve distributions of surface air temperature, precipitation, evaporation, and runoff variation at different elevations better than their host GCMs. The RCMs have similar climate change signals (although with slightly smaller magnitudes) than their host GCMs. In general, the NARCCAP RCMs project larger decreases in precipitation and evaporation in summer than their host GCMs, which translates to larger reductions in summer runoff. The RCMs have smaller decreases in spring snowmelt and runoff (arising in part from smaller temperature changes, and also from better resolution of topography) than their host GCMs, resulting in smaller decreases in annual runoff. Because winter precipitation is more efficiently converted to runoff than is summer precipitation, the overall effect on annual runoff is that the RCMs generally predict less sensitivity to global warming than do their host GCMs.