Assessing uncertainty and predictability of regional climate, with a focus on hurricanes

The inherent uncertainty in - and predictability of - regional climate predictions over limited regions is assessed using a combination of simulations with the NCAR Nested Regional Climate Model (NRCM) and statistics. Our focus is on climate and weather extremes that have high impact on society and how these may be changing with both climate variability and change. This presentation will focus on hurricanes over the Gulf of Mexico. An ensemble of state-of-the-art NRCM simulations is used to assess the uncertainty arising from choice of: various physical processes, initial conditions, domain sizes, boundary conditions and model breeding techniques. The objectives are to: 1) increase our understanding of the ability of regional climate models to capture regional climate information - both at the individual and collective ensemble member levels; 2) investigate the role of different combinations of processes in reproducing regional climate; and 3) estimate the associated uncertainty and inherent predictability. The study is geared towards an improved ability to select optimal combinations of model configurations to maximize the predictive signal for local climate extremes. Initial results from this assessment will be presented. In particular, we include examples specific to potential future changes in hurricane impacts on Gulf of Mexico offshore facilities.