The Application of a Statistical Downscaling Process to Derive 21st Century River Flow Predictions Using a Global Climate Simulation

The ability of water managers to maintain adequate supplies in coming decades depends, in part, on future weather conditions, as climate change has the potential to alter river flows from their current values, possibly rendering them unable to meet demand. Reliable climate projections are therefore critical to predicting the future water supply for the United States, but the resolution of the global climate models (GCMs) often used for climate forecasting is too coarse to resolve the changes that can affect hydrology, and hence water supply, at regional to local scales. We therefore apply a statistical downscaling technique that involves a correction of the cumulative distribution functions of the GCM-derived temperature and precipitation for the 20th century, and the application of the same correction to 21st century GCM projections. This is done for three meteorological stations located within the Coosa River basin in northern Georgia, and is used with a surface hydrology model to calculate future river flow statistics for the upper Coosa River. Results are compared to historical Coosa River flow and to flows calculated with the original, unscaled GCM results to determine the impact of potential changes in meteorology on future flows.