Combined Temperature and Precipitation Variability May Increase the Frequency of Harmful Algal Blooms in Lake Champlain, 1992-2100

The frequency and severity of harmful algal blooms (HABs) in fresh water lakes are driven by many variables, ranging from nitrogen and phosphorus loading, water depth, formation and evolution of zooplankton species, to vertical and horizontal profiles of temperature gradients. In this paper, we assess the combined effects of precipitation and temperature variability in inducing HABs in shallow freshwater lake systems. We use 0.8KMx0.8KM downscaled ensembles from 6 GCMs to drive a watershed scale Dynamic Bayesian Network (DBN) that predicts two-weekly magnitude of ChlA in two shallow segments of Lake Champlain Basin from 2015-2100 for spring, summer and fall months (April-November). The DBN is trained and tested on hydrologic, biogeochemical and ecological monitoring data from 1992 to 2015. While temperature increases in the worst-case climate change scenarios are widely expected to lead to more frequent and intense HABs in shallow lake segments, we find that the combined effects of temperature and precipitation variability at watershed scales could significantly increase the frequency and duration of HABs in the lake segments than higher temperatures alone. We examine the sensitivity of modeled bloom conditions to nutrient flux variability under different climate change ensembles and find that the DBN projections for testing the combined effects of temperature and precipitation variability are robust across two freshwater lake systems. We discuss the generalizability of the estimated DBN to other shallow lake water systems and identify cyber-infrastructure and information requirements from field-based sensors to train and test the DBNs in specific environmental conditions at watershed scales.