A spatial regionalisation approach to reduce uncertainty in climate model bias correction

Abstract:

There are significant biases in precipitation simulations from General Circulation Models and Regional Climate Models. These can be addressed through statistical downscaling or bias correction. Techniques have been developed to address biases at a range of time scales however these are generally applied to individual grid cells. A review of biases in a number of climate models shows that there is a clear spatial structure to the biases. In addition, many hydrological applications introduce additional spatial structure to the analyses. Therefore I propose a spatial model of bias as an extension to existing bias correction approaches. By considering the biases in space, uncertainty in the bias parameters can be reduced. The new method leads to improvements in the bias correction results. The implications for future simulations have been considered.