A Combined Dynamical and Statistical Downscaling Technique to Reduce Biases in Climate Projections: An Example for Winter Precipitation and Snowpack in the Western United States

11 Large biases associated with climate projections are problematic when it comes to 12 their regional application in the assessment of water resources and ecosystems. Here, we 13 demonstrate a method that can reduce systematic biases in regional climate projections. 14 The global and regional climate models employed to demonstrate this technique are the 15 Community Climate System Model (CCSM) and the Weather Research and Forecasting 16 (WRF) model, respectively. The method first utilized a statistical regression technique 17 and a global reanalysis dataset to correct biases in the CCSM-simulated variables (e.g., 18 temperature, geopotential height, specific humidity, and winds) that are subsequently 19 used to drive the WRF model. The WRF simulations were conducted for the western 20 United States and were driven with a) global reanalysis, b) original CCSM, and c) bias21 corrected CCSM data. The bias-corrected CCSM data led to a more realistic regional 22 climate simulation of precipitation and associated atmospheric dynamics, as well as snow 23 water equivalent (SWE) in comparison to the original CCSM-driven WRF simulation. 24 Since most climate applications rely on existing global model output as the forcing data 25 (i.e. they cannot re-run or change the global model), which often contain large biases, this 26 effective and economical method provides a useful tool to reduce biases in regional 27