Extending the forecast model: Predicting the spatial distribution of HABs in western Lake Erie

Session: 37. - Harmful Algal Blooms (HABs) and their Toxicity: Remote Sensing and Modeling Approaches

Nathan Manning, University of Michigan, manningn@umich.edu
Yu-Chen Wang, University of Michigan, yuchenw@umich.edu
Colleen Long, University of Michigan, longcm@umich.edu
Mike Sayers, Michigan Tech. Research Inst., mjsayers@mtu.edu
Karl Bosse, Michigan Tech Research Inst., krbosse@mtu.edu
Robert Shuchman, Michigan Technological University, shuchman@mtu.edu
Donald Scavia, University of Michigan, scavia@umich.edu

Abstract

The western basin of Lake Erie is a spatially,hydrodynamically and politically diverse system, with a significant economic component that can be severly negatively impacted by a Harmful Algal Bloom. Models predicting the severity of the HABs in Lake Erie are useful tools for a wide range of scientists, managers and public stakeholders. However, most current models are not spatially explicit in thier forecasting, and thus the utility of these models is diminished.  We used Michigan Tech Research Intitute's MODIS derived chlorophyll estimates for the years 2002-2016 in conjunction with a recalibration of University of Michigan's bloom forecast model to develop spatially explicit bloom estimations for the near-shore region of the western basin. The western basin was divided into a series of nested sub-divisions that represented hydrologic and political boundaries. For each of these subdivisions a non-linear regression model was developed to describe the relationship between bloom intensity in the subdivisions, and what is observed in the whole basin. The relationships presented here are robust, and temporally stable. By linking these models with the existing forecasting model it will be possible to make regionally explicit estimates of bloom size and impact prior to the onset of blooms in mid-to-late summer. 

1. Keyword
Lake Erie

2. Keyword
harmful algal blooms

3. Keyword
modeling