Coastal Upwelling Influences Hypoxia Spatial Patterns and Nearshore Dynamics in Lake Erie

Hypoxia, defined as dissolved oxygen (DO) < 2 mg/L, in the central basin of Lake Erie has been studied since the mid-1900s. Even so, spatial patterns of hypoxia, and episodic hypoxia in nearshore areas where drinking water plant intakes are located, are not well characterized owing to limited observations and short-term dynamics. We evaluated a physically based, DO model with respect to patterns of hypoxia observed in Lake Erie. The DO model used assigned rates of sediment and water column oxygen demand that were temperature dependent but otherwise spatially and temporally uniform. The DO model was linked to National Oceanic and Atmospheric Administration's (NOAA) Lake Erie Operational Forecasting System hydrodynamic model, an application of the Finite Volume Community Ocean Model (FVCOM). Model temperature and DO were compared with observations from ship-based studies, real-time sensor networks and an array of moored sensors that we deployed in 2017. In years with dominant southwesterly winds, persistent downwelling occurred along the south shore, which resulted in a thinner thermocline and earlier initiation of hypoxia along the south shore than the north. Occasional northeast winds temporarily reversed this pattern, causing upwelling along the south shore that brought hypoxic water to nearshore locations and water intakes. The DO model reproduced observed spatial and temporal patterns of hypoxia and revealed locations subject to episodes of hypoxia, including nearshore Ohio, north of Pelee Island, and near the Bass Islands. Model skill was limited in some respects, highlighting the importance of accurate simulation of the thermal structure and spatial patterns of oxygen demand rates. Plain Language Summary Hypoxia, defined as dissolved oxygen (DO) < 2 mg/L, results from a combination of physical and biological processes. Hypoxia in Lake Erie reduces habitat and food supply for fish and complicates drinking water treatment. We applied a mathematical model that was driven by weather information (wind and air temperature) and simulated water temperature, currents, and DO. By evaluating whether a model that focused on physical processes could predict hypoxia, we investigated the relative importance of physical versus biological drivers. We compared model temperature and DO with measurements conducted by us and by others. In years with dominant southwesterly winds, persistent downwelling occurred along the south shore, which resulted in a thinner thermocline and earlier initiation of hypoxia along the south shore than the north. Occasional northeast winds temporarily reversed this pattern, causing upwelling along the south shore that brought hypoxic water to nearshore locations and water intakes. The DO model reproduced observed spatial and temporal patterns of hypoxia, indicating the importance of physical drivers, and revealed locations subject to episodes of hypoxia, including nearshore Ohio, north of Pelee Island, and near the Bass Islands.