Three-dimensional modeling of sediment resuspension in a large shallow lake

Abstract

We simulated bottom resuspension events in Lake Erie, using a coupled three-dimensional hydrodynamic and water quality model. Key parameters in the model, including critical bottom shear stress (τcr) and resuspension rate (α) were calibrated and validated by comparing the model output to observations. These included total suspended solid (TSS) concentrations in the bottom boundary layer (RMSE = 0.74 mgL-1) and water column (RMSE = 0.81 mgL-1), and to time series of acoustic backscatter signal (R2 > 0.8) and turbidity (R2 ≈ 0.4) from long-term moorings near the lakebed in 2008–09 and 2013. Signals from phytoplankton, in spring and summer, caused discrepancies between modeled TSS and the observed turbidity data. Although common practice, we show that literature-based or field-observed critical shear stress should not be directly applied in large-scale Reynolds-averaged sediment model as this will likely underestimate resuspension. In agreement with the literature, the model reproduced more frequent and intensive surface-wave driven resuspension in the shallow regions (< ~20 m), particularly in the western basin, compared to the deeper central and eastern basins, where bottom stresses induced by mean currents (τc) were comparable with those due to surface waves (τw). However, on the north-shore of the eastern basin, τc often predominated over τw. We simulated thermocline motion, including up- and down-welling events and swashing of the internal Poincaré wave, to contribute to τc in the central basin and form nepheloid layers.