Regional Climate Simulations

It is essential to determine the local and regional impacts of global warming.  Global climate models, driven by increasing concentrations of greenhouse gases are useful, but are typically run at grid spacings of 100-200 km and thus lack the resolution to realistically simulate local weather features driven by terrain, land-water contrasts, surface variability, and convective systems.

regio
Terrain maps of a typical global climate model (left) and weather forecast model (12-km grid spacing) on the right.

To deal with this issue, this project is running a high-resolution weather forecasting model (WRF) at 12-km or 4-km grid spacing over a century or more for the western U.S. and other regions (such as southeast China).

Importantly, we are running an ensemble of many of such forecasts to determine the uncertainties and range of possible regional climate projections.  To do this, we have secured all the CMIP-5 climate simulations that have 6-hr output and have run 130-year simulations (1970-2000) for them (12 in all).

Such high-resolution dynamical downscaling of global climate models, has produced a number of surprising results, suggesting increased springtime low clouds over the coastal Pacific Northwest, bands of enhanced warming from melting snowpack, drying zones in the lee of major barriers, and decreasing easterly winds.  Substantial work is now ongoing for varying emission scenarios, more global climate models, and in understanding the local/regional impacts of global warming.