Future Changes in Precipitation and Associated Physical Processes within Extratropical Cyclones Using CMIP5 and a Downscaled WRF

Future changes in heavy precipitation, storm surge, and high winds associated with extratropical cyclones can potentially have large impacts along the highly populated region of the U.S. East coast. This presentation focuses on the future (later 21^stCentury) changes in precipitation associated with extratropical cyclones during cool season over eastern North America and the western Atlantic, and how the precipitation (e.g., latent heating) and other processes (baroclinicity, stability, moisture, etc…) can impact the cyclone frequency and intensity in this region. The precipitation changes within 10 CMIP5 models and 2 down-scaled Weather Research and Forecasting (WRF-HR) runs at 20-km grid spacing (using GFDL-ESM2M and CCSM4) are evaluated over 3 regions: U.S. East coast land region (ECL), U.S. East coast waters (ECW) and a broader western Atlantic region (WA). Two WRF-LR runs using 100-km grid spacing were also completed to better understand the impact of grid resolution. A cyclone relative approach is employed to quantify the future changes and physical processes. The historical cool season (November-March) precipitation over eastern North America and western Atlantic in the CMIP5 models and WRF is evaluated with respect to the average of two precipitation products (GPCP and CMAP). Those models with a better performance in the historical cyclone climatology generally have a smaller mean absolute error in precipitation, but there are a few exceptions.

During 21^st century there is a significant shift towards the heavy precipitation part for both of the mean and maximum precipitation around cyclone center within the eastern North America and western Atlantic. The precipitation increase for the ECL cyclone centers (> 20%) is larger than the ECW (14-20%) and WA (10-13%) cyclone centers during the late 21^stcentury. The ECL cyclone centers have the largest increase (~150%) in the most extreme maximum precipitation (>56 mm/day). The relatively strong (<990 hPa) cyclone centers have the largest increase (20-30%) in precipitation rate with the ECL, while the increase for the relatively weak (>1005 hPa) is only around 10%. A few reasons for the larger precipitation increase in the ECL region include a larger increase in moisture content than the other two regions, less of a stability increase than the main storm track region offshore, and a slightly stronger upper level jet over ECL favorable for cyclone deepening.  While there is a decrease in low-level baroclinicity during the later 21st century, latent heat release offsets this and thus there is an overall increase in cyclone deepening.

Although the cyclone frequency in the nested WRF runs is dominated by the corresponding GCM solution, the WRF-HR can produce a more realistic cyclone intensity distribution, while the WRF-LR and CMIP5 models underestimate the intense cyclones during the historical period. In contrast, the WRF-HR overestimates the precipitation along the storm track during the historical winters. Similar to the CCSM4, the WRF forced by CCSM4 increases cyclone track density along the East Coast by 10-40%, and the WRF-HR also has a ~10% increase in deep cyclones (< 976 hPa). The increase in latent heat release from heavy precipitation becomes important in the future cyclone changes in the CCSM4 and WRF. In contrast, the GFDL-ESM2M and nested WRF have a decrease (20-40%) in cyclone track density over the western Atlantic, with a more neutral region within the ECL, since the precipitation increase and associated latent heat release in GFDL-ESM2M and WRF are much smaller. Some of the other reasons for this difference between CCSM4 and GFDL-ESM2M will also be discussed.