Influence of the Great Lakes on Hydro-Climatic Variables in CMIP5 Models and Implications for Weather-Driven Applications

Previous studies in the Great Lakes region of North America indicate that statistically significant trends have been observed over the past several decades in both overland and over-lake measurements of several hydro-climatic variables. The potential for the Great Lakes themselves to modify regional climate, particularly in leeward areas, is well known. However, given the relatively coarse spatial resolution of Global Climate Models (GCMs), the physical representation of the Great Lakes is often quite different from reality and in many cases are missing entirely from model processes and land/atmosphere/ocean gridding schemes. This study evaluates a subset of GCMs from the Coupled Model Intercomparison Project Phase 5 (CMIP5), selected for their representation (or absence) of the Great Lakes in their underlying processes and sub-daily output of the hydro-climatic variables: near-surface air temperature, near-surface specific humidity, and precipitation. Historical values (1979-2005) are compared to output from the North American Regional Reanalysis (NARR), with particular attention paid to representation of regions windward and leeward of Lake Michigan. Preliminary analysis of projected late-21st century (~2071-2100) output indicates substantial changes from the historical climate across the region in terms of increased near-surface air temperature and near-surface specific-humidity, while the precipitation signal is mixed. Annual, seasonal, monthly mean values and diurnal cycles were found to be modified by the presence/absence of lakes in the GCMs. Finally, we illustrate the potential utility of this information in a plant disease-related application.