Reliably simulating the local climate as well as the the climate extremes for a region
such as Southeast Asia, which has a complex topography, requires a refine mesh. The
recently developed version of the Community Earth System Model (version 1.2,
CAM5.3) with a variable resolution option (VR-CESM) is assessed for long-term
regional climate modeling over Southeast Asia (SEA). In the model, we implement a
refined mesh with a horizontal resolution of 0.25 degree (~28 km) over Southeast Asia
and a resolution of 1 degree globally. The mean climatology of near-surface
temperature and precipitation in VR-CESM and CESM, with a coarser mesh, are
analyzed and contrasted with reanalysis, gridded observational data sets, weather
station data, and the traditional regional modeling approach—regional climate model
one-way dynamical downscaling. With only prescribed sea surface temperatures
(SSTs), VR-CESM improves the spatial distribution and climatology of precipitation and
temperature compared to simulations at coarser resolution. The ability of CESM at
different resolutions to capture precipitation extremes is also characterized using
Extreme Value Analysis. For the precipitation extremes, we find that VR-CESM
captures well the relation between precipitation extremes and return periods as seen in
the observational data sets, whereas the simulations at coarser resolution either
underestimate the precipitation extremes or overestimate the slope. This study
highlights the value of variable-resolution global climate models in capturing fine-scale
atmospheric process as well as precipitation extremes and emphasizes the importance
of two-way coupling in simulating regional climate over SEA.