Empty nesting: Where do you need high resolution to accurately simulate Atlantic hurricane impacts in climate models?

Wednesday, December 12, 2018 - 16:30
Add to Calendar

As cutting-edge modeling techniques and increases in computing power allow for finer horizontal resolution, the ability of general circulation models (GCMs) to accurately simulate tropical cyclones (TCs) and their climatological distribution improves. Techniques that only simulate a portion of the globe (such as regional climate or variable-resolution global models) can further enhance these analyses by targeting computing resources in a specific region of interest, such as basins where TCs are prevalent. However, configurations of these models generally lean on the maximization of available resources with aspects such as domain size and location, and are historically more of an art than a science.

This study investigates the required spatial extent of a high-resolution nest to accurately simulate TC climatology in the North Atlantic (NATL) basin, with an emphasis on landfalling hurricanes. A nine-member multidecadal ensemble of variable-resolution simulations (25km refinement) is evaluated in the NCAR Community Earth System Model, with varying spatial extents across the NATL and Northern Africa. We evaluate storm climatology and discuss the critical resolution needed for simulating various aspects of the TC lifecycle, including the generation of African Easterly Waves, TC formation in the Main Development Region, intensification across the basin, and large scale steering patterns. Specific focus is paid to storms that strike populated areas in the western NATL, particularly the skill of projected intensity and overland precipitation statistics. Best practices for using limited area and variable-resolution models for climate impact assessments, subseasonal-to-seasonal prediction, and hydrological studies along the eastern coast of North America are also highlighted.

Link for More Information: