Robustness of high-resolution regional climate projections for Greenland: a method for uncertainty distillation

ABSTRACT: Managing adaptation to climate changes in Greenland will depend, to a large degree, on high-resolution climate simulations and associated uncertainty estimates. A single high-resolution climate simulation is generally insufficient to quantify the uncertainty of a given scenario projection. For Greenland, this becomes a critical issue because of a lack of high-resolution climate experiments for this region. Therefore, we introduce and test a new method to solve this uncertainty assessment problem. Using the regional climate model (RCM) HIRHAM5 over Greenland in combination with an ensemble of RCM simulations from a different geographical setting, (i.e. EURO-CORDEX), we investigate to what extent the uncertainty of projected climate change at high resolution can be evaluated from corresponding temperature spreads in a wider set of global climate models (GCMs), that is, CMIP5. The study is based on a set of time-slice simulations downscaled with HIRHAM5 at 5.5 km resolution for the RCP4.5 and RCP8.5 scenarios for Greenland with boundary information from the GCM EC-Earth. Our proposed uncertainty assessment method establishes a foundation on which high-resolution and relatively costly regional climate projections can be assessed as well as when using only a single RCM without the presence of analogous downscaling experiments with other RCMs and GCMs, and instead relying on existing information from CMIP5. Thus, the uncertainty of a wide range of climate indices that scales with temperature can be evaluated and quantified through the inter-model temperature spread within CMIP5.