Comparison of Two Dynamical Downscaling Approaches on the Future Changes in North Atlantic Extratropical Cyclones

Coastal extratropical cyclones can cause heavy precipitation and strong winds, which can lead to extensive damage to populated areas and can greatly impact shipping and fishing routes in the Atlantic Ocean. Understanding how these cyclones will change in a warming climate is important to these populated coastal cities along with the shipping and fishing industries. Due to their relatively course resolution, climate models have difficulties simulating these winter coastal cyclones. As a result, dynamical downscaling techniques have been applied to generate climate model solutions at a higher resolution. It is important to understand the differences between dynamical downscaling techniques, especially how and why they differ from each other in future climate predictions. This project compares two dynamical downscaling techniques: climate model nesting and pseudo-global warming (PGW), and their ability to simulate winter extratropical cyclones in the North Atlantic. In particular, the cyclone storm track densities and intensities will be examined in this study as well as reasons for any differences in future predictions of these storms. Previous studies have investigated North Atlantic cool season extratropical cyclone storm tracks using downscaling techniques; however, there has not been a direct comparison between the techniques using the same model, time periods, and climate model forcing data.

To do this comparison, the Weather Research and Forecasting (WRF) Model (version 3.7) was run over the North Atlantic Ocean at 20-km horizontal grid spacing. The WRF was run over ten historical cool seasons (January- March or JFM) of 1993-2002 using the CFSR reanalysis and 5 models from the Coupled Model Intercomparison Project Phase 5 (CMIP5). The future cool seasons runs were done for JFM of 2090-2099 with the same CMIP5 members for climate model nesting and PGW runs. Both sets of WRF runs used the same physics, namely the Zhang-McFarlane (ZM) scheme, WRF Single-Moment 6 Class (WSM6), and Mellor-Yamada-Janjic (MYJ) schemes were used for the convective scheme, microphysics, and planetary boundary layer, respectively. The Community Atmospheric Model (CAM) scheme was used for the longwave and shortwave radiation, and the Noah land surface model represented the land surface. For the future runs the CAM radiation scheme CO2 concentrations were adjusted to 936 ppm, which is consistent with RCP8.5 2100 concentrations. The PGW approach calculates RCP8.5 scenario temperature changes from the CMIP5 models from 2090-2099. These changes are added to the 1993-2002 data for “future” initial and lateral boundary conditions for the WRF. The other nested approach uses 6-hourly CMIP5 data for initial and boundary condition data, with the WRF restarted at the start of each winter season. To track the cyclones, the modeling analysis, and prediction (MAP) climatology of midlatitude storminess (MCMS) algorithm was used. Storm track densities and cyclone relative composites were then done utilizing the tracks created from the historical and future model runs. This allowed for a direct comparison between the two downscaling techniques on storm track and intensities and allowed for comparison of the nesting approach with the present climate. This presentation will also compare the two approaches in predicting storm track density and intensity in a future climate as well as highlight reasons for any differences in the future climate, such as differences in large-scale flow, baroclinicity and vertical stability, PV generation via latent heating, etc.