Orographic and Land Surface Impacts on Numerical Weather Forecasts When Simulating a Sudden Downwelling Event in a Medium-Sized Lake

The increasingly urgent problem of harmful algal blooms (HABs) in lake and near-shore coastal environments has placed renewed focus on the coupled simulation of weather, hydrology and hydrodynamics. To date, modeling efforts at the Great Lakes and elsewhere have provided results generally consistent with observations, however, hydrodynamics closer to the shoreline – where HABs commonly originate – are often not well simulated. The poor near-shore hydrodynamic simulation can often be attributed to a weather model running at a resolution too coarse to properly resolve airflow dynamics at the land-water interface. To interrogate this issue, the present study uses Lake George, NY, a medium sized lake (55 x 3 km) in the eastern Adirondacks, as a testbed to quantify and understand the impact of weather model resolution on near-shore hydrodynamics during a sudden downwelling event in July 2017. Such downwelling events are of importance to HABs-related research because immediately after the relaxation of the downwelling, the observed chlorophyll-a profiles show an increased concentration within the mixed layer. This well-observed event is shown to be better captured by a hydrodynamics model (SUNTANS Community Model) when driven by a weather model (WRF) at its highest horizontal resolution (333 m), primarily because of stronger winds across the lake. The hydrodynamics model is configured with a variable horizontal resolution from ~27m to ~70m. Additional model experiments reveal how better representation of the surrounding orography and the land-water boundary in WRF at coarser resolution can improve the over-lake wind field, and thereby the strength of the downwelling. The results underline that if a weather model must be executed at coarse resolution (although not too coarse) in a region of complex topography, special attention should be given to how static fields are represented.