An Introduction to Trusting Global Climate Models and Bias Correction

The Great Lakes Integrated Sciences and Assessments (GLISA) program strives to provide usable climate information for the Great Lakes region that is plausible, defensible, and actionable for regional partners. To better recommend the ways in which practitioners can better trust and utilize Global Climate Models (GCMs), we have written a white paper investigating bias and bias correction impacting a GCM’s performance. Through this paper, we suggested small bias be defined as less than 10% model error compared to observations, and the larger biases defined as being greater than 10%. To test these definitions, we evaluated GCMs’ temperature and precipitation simulations from the Coupled Model Intercomparison Project 5 (CMIP5), the North American Coordinated Regional Downscaling Experiment (NA-CORDEX), and the University of Wisconsin-Madison Regional Climate Model Version 4 (UW-RegCM4). The results suggested many of the models were too wet and mostly cold during the winter and spring months, but the summer months had a more equal distribution of wet/dry or cold/hot biases in the simulations. The fall months resulted in the CMIP5 models being too dry while the NA-CORDEX and UW-RegCM4 models being considered too wet. The overall recommendations from the paper and model evaluation include additional investigations of the CMIP5, NA-CORDEX, and UW-RegCM4 model errors and encouraging practitioners to utilize local information to fill in the knowledge gaps resulting from model bias.