Simulation of advection and vertical distribution of buoyant cyanobacterial colonies in Lake Erie with a Lagrangian particle model for short-term forecasts of harmful algal blooms

Mark D Rowe1, Eric J Anderson2, Timothy Wynne3, Richard P Stumpf3, David L. Fanslow2 and Henry A. Vanderploeg2, (1)Cooperative Institute for Great Lakes Research, University of Michigan, Ann Arbor, MI, United States, (2)National Oceanic and Atmospheric Administration, Great Lakes Environmental Research Laboratory, Ann Arbor, MI, United States, (3)National Oceanic and Atmospheric Administration, National Centers for Coastal Ocean Science, Silver Spring, MD, United States
Abstract:
Cyanobacterial harmful algal blooms (CHABs), primarily Microcystis, are a recurring problem in western Lake Erie. Short-term forecasts of CHABs are important to water treatment plant operators, anglers, recreational boaters, and beach users. NOAA NCCOS and NOAA GLERL have developed experimental forecast products that indicate the present location and extent of CHABs from satellite imagery, then predict the movement of the CHAB five days into the future using forecast meteorology. These products use Lagrangian particle tracking models to forecast CHAB transport, forced by currents from a hydrodynamic model. Western Lake Erie is a shallow (< 11 m) freshwater system. CHABs occur during the months of July to October, a time when the water column is intermittently stratified or mixed to the bottom in response to wind and surface heating or cooling. Microcystis colonies may be mixed through the water column or concentrated in surface scums, depending on the balance between buoyancy and turbulent mixing. Present forecast models do not simulate the vertical distribution of buoyant cyanobacterial colonies, which detracts from model skill when mixing events cause large changes in surface concentration. We evaluated model skill in a 2011 hindcast using a Lagrangian particle model that included vertical mixing and buoyancy by comparison to satellite imagery and in-situ observations. Turbulent diffusivity and 3D currents were provided by the Finite Volume Coastal Ocean Model (FVCOM). Random-walk vertical mixing schemes may produce artifacts in simulated concentration profiles in the presence of diffusivity gradients. We evaluated several random-walk numerical schemes and time step criteria to determine conditions under which artifact formation could be avoided. Inclusion of vertical mixing with buoyancy enabled the model to simulate observed changes in surface chlorophyll concentration in response to variable mixing conditions, which improved model skill statistics.
Back to: Comparing Physical Processes in Shallow Seas, Large Lakes, and Semi-Enclosed Basins Posters