Impacts of Intra-Annual Climate Variability and Change on Phosphorus Loads in the Great Lakes Basin

Riverine phosphorous loads vary as a function of source strength, discharge, and landscape characteristics. In this work, we consider how seasonal variability in river discharges control phosphorous loads and how climate change will impact discharges and corresponding phosphorus loads. We focus on 14 watersheds in the U.S. Great Lakes Basin, given the potential for riverine phosphorus exports to contribute to ecosystem impacts in the Great Lakes. The 14 watersheds vary in terms of land use and hydrologic regimes. Seasonal load-discharge relationships were estimated using a linearized form of a power law functions for each watershed, based on historical data from 1961-1999. Best-fit seasonal periods ranged from monthly to three months. Estimates of the leading coefficient ranged over more than 11 orders of magnitude, over the seasons and study watersheds. Estimates of the power ranged slightly less than unity to greater than two, reflecting wide differences in the linearity of the load-discharge relationship.

For the future climate periods (2046-2065 and 2081-2100), a suite of 9 bias corrected projections were made using the CMIP3 database, generating precipitation and temperature inputs to the Large Basin Runoff Model (LBRM). The LBRM is a calibrated, lumped parameter model that predicts discharges for large watersheds in the Great Lakes. In this case, LBRM discharge predictions were used as inputs to the phosphorus load discharge relationships. In general, median flows are predicted to change little, but low flows (characterized by Q₅) are predicted to decrease on average by 12% and 19%, and high flows (characterized by Q₉₅) are predicted to increase on average by 9% and 12% over the near- and far-future periods, respectively. For most watersheds, median phosphorous loads change linearly with respect to changes in median discharges. However, for a few watersheds, median phosphorous loads are predicted to increase proportionally greater than increases in discharge. This outcome is explained by concurrence between increases in high seasonal flows and seasonally-high values of the power law coefficient. These results illustrate the importance of considering seasonality in water quality modeling, especially when assessing the impacts of climate change.