Quantification of Temperature and Precipitation Variability Over the Great Lakes

The North American Great Lakes (the largest system of lakes on Earth) have historically responded to regional climate perturbations through changes in seasonal and inter-annual precipitation (both over-lake and over-land), changes in lake water levels, and gradual (but persistent) changes in air and surface water temperatures. The 15 years following the 1997-1998 El Niņo underscore the sensitivity of the Great Lakes to climate change, and are marked by an extreme shift in the regional water budget dominated by a rapid increase in over-lake evaporation rates, and subsequent record-low water levels on Lakes Superior, Michigan, and Huron. Interestingly, this period has also included years with unprecedented seasonal water level increases as well, indicating changes in both the variability and frequency of extreme events within this massive freshwater system. Collectively, these changes represent a challenge to operational water resource management planning and a need to improve water budget forecasts through a more thorough understanding of the Great Lakes climate system, and its connection to broader scale climate patterns.

Research-oriented and operational forecasting of the Great Lakes water budget and water levels is currently based on a suite of products ranging from conventional rainfall-runoff and lake thermodynamic models to empirically-based decision making tools, all of which rely on basin-scale projections of seasonal temperature and precipitation (T and P). Assessing relationships between observed Great Lakes basin T and P variability and projections from readily-available climate outlooks therefore represents a potential stepping stone towards improvements in regional model-based operational water management protocols. Here, we conduct this assessment, with a particular emphasis on comparing NOAA Climate Prediction Center (CPC) outlooks (of T and P) with observed values of Great Lakes basin-scale T and P (both over the land and lake surface, a distinction that is not often considered in many coarse-scale climate models, but one that has a profound impact on the regional water budget). We then propagate differences between CPC outlooks and observed T and P values through existing Great Lakes water budget forecasting products to quantify impacts of CPC biases on seasonal water level projections. We find that explicit acknowledgment (and quantification) of relationships between basin T and P and large-scale climate indices (i.e., El Niņo-Southern Oscillation and Pacific Decadal Oscillation) has the potential to improve regional climate projections above and beyond what is currently provided by the CPC outlooks. This finding serves as a basis for modification of existing CPC outlook maps to more explicitly address the needs of the international Great Lakes basin water resources management community.