Representing Temperature and Precipitation Extremes by Dynamical Downscaling Using a Global, Multi-resolution Atmospheric Modeling System

Climate change will make extreme weather events more likely to occur and more severe. Changes in heat waves, floods, droughts, wildfires, hurricanes, and snowfall are expected to occur in a changing climate system. These changes will have large consequences on human health, agriculture, energy use, and the global economy and are driven by processes that occur at variable resolutions. Our goal is to explore the ability of a multi-scale climate model to reproduce basic hydroclimatic variability at variable resolutions.

Changing climate extremes can be traced to rising global temperatures, increases in the amount of water vapor in the air, and changes in atmospheric circulation. Global Climate Models (GCMs) are used to understand and simulate how these large-scale changes result from various climate change scenarios. However, the limited spatial resolution of GCMs reduces their ability to estimate local or regional extremes. Extreme weather events are strongly modulated or controlled by local terrain and land surface features that cannot be represented by GCMs. Also, extremes are often mesoscale weather events that are too small to be resolved by a GCM.

For this reason, researchers have turned to dynamical downscaling to estimate local and regional climate extremes. With dynamical downscaling, a regional climate model is used to simulate local climate at high resolution using lateral boundary conditions forced from a GCM. The presence of lateral boundaries in the regional climate model prevents the natural propogation of waves and can lead to significant errors. An interior constraint is often necessary to keep the regional model from drifting from the GCM solution, but this reduces the ability of the regional model to generate extremes.

An alternative approach explored here is to use next-generation multi-resolution modeling systems that can cover the entire earth with a coarse grid and telescope down to much higher resolution in areas of interest to represent local climates. The global grid and absence of lateral boundaries will allow a much more faithful representation of the global climate system than a regional model can provide. Using the multi-resolution Ocean Land Atmosphere Model (OLAM), we will dynamically downscale GCM fields to examine the representation of extreme events in next-generation atmospheric models. Using a global model to dynamically downscale GCM datasets should provide a representation of extreme events that are more faithful in space and time to the driving large-scale fields.