Prediction, Validation, and Calibration of Coastal Storms and Associated High Impact Weather in Ensemble Regional Climate Simulations over the Northeast U.S.

Increases in extreme weather events (heavy precipitation, strong winds, storm surge, etc…) under anthropogenic climate change would have profound consequences for both human society and the natural environment. The Northeast U.S. is particularly vulnerable, since there are frequent coastal storms in this populated region. The ability of atmosphere–ocean general circulation models (AOGCMs) to accurately simulate high impact weather and its trend is of great importance, but unfortunately these AOGCMs can not resolve many of the mesoscale features that cause damaging winds, heavy precipitation, and coastal flooding. A common approach has been to downscale these AOGCMs with regional climate models (RCMs) to better resolve these storms and the underlying surface characteristics; however, past RCM resolutions have been marginal (30-50-km) and few studies have explored the uncertainty of these predictions using ensembles. Furthermore, these raw ensembles are often biased or underdispersed, so more work is needed to calibrate them, especially for high impact weather events.

This project investigates the future changes (up to year 2070) of high impact weather over the Northeast U.S. using an ensemble of Weather Research and Forecasting (WRF) members at 20-km grid spacing nested within an ensemble of climate model simulations using the NCAR Community Climate System Model (CCSM). The first goal is to determine how well this joint ensemble system can predict previous extratropical cyclones and associated high impact weather over the western Atlantic and eastern U.S. during the cool seasons of 1981-2009. The CCSM and WRF members will use different physics perturbations and parameterizations to diversify a 15-member CCSM-WRF ensemble. An important science question is the role of sea surface temperature (SST) gradients in these predictions, so both high and coarse resolution SST will be used. Also, the upscale impact of diabatic heating from the storms will be explored using a new two-way nesting setup of CCSM-WRF being developed at Stony Brook University. The past predictions of cyclone strength, winds, precipitation will be bias corrected using observations and reanalyses and calibrated using an ensemble weighting scheme. The second goal of the project is to use the calibrated ensemble and future simulations to project changes in the statistics of high impact weather (cyclone tracks, heavy precipitation, storm surge, etc…) in the years 2040-2070. The future predictions will use a 14-member CCSM-WRF ensemble with two different emission scenarios and physics (WRF starting at 2040). A new approach to obtain high resolution SSTs in the future via downscaling the CCSM SSTs will be tested. The results will be compared with the NARCCAP ensemble results over the Northeast U.S.

Overall, this project is one of the first to apply high-resolution regional climate ensembles that are calibrated for future predictions of high impact weather. Thus, it will serve as a useful data source for decision makers interested in how heavy precipitation, storm surge, strong winds, and cold air outbreaks may change over the Northeast U.S. during the next several decades.