How good is a global climate model in simulating mass changes of mountain glaciers?

Current global climate models (GCMs) do not resolve mountain glaciers in their land surface scheme. Therefore, impact studies of climate change on glaciers still rely on dynamical or statistical downscaling of GCMs. Considering the complexity of a GCM with a full spectrum of parametrizations, it is necessary to identify the key processes on glacier-climate interface that need to be sufficiently well simulated by a GCM if mountain glaciers are to be implemented in the land-scheme of that GCM. In this study we validate current performance of the Community Earth System Model (CESM) in simulating climate processes that drive glacier mass balance on a regional scale. Glacier mass balance models of different complexities are compared: energy balance models which are forced by energy and humidity fluxes simulated by CESM for a glacier surface, and temperature-index models driven only by CESM 2m air temperature at a glacier surface. We also investigate the use of statistical methods in representing glacier dynamics within the land surface scheme, since the physics-based glacier flow models require much higher resolution (<1 km) of surface topography than is available in the state-of-the-art GCMs.