Applications of agent-based modeling to nutrient movement Lake Michigan
Citation:
Brown, T., J. Pauer, AND T. Hollenhorst. Applications of agent-based modeling to nutrient movement Lake Michigan. State of Lake Michigan Conference, Green Bay, WI, November 07 - 10, 2017.
Impact/Purpose:
The nearshore region of the Great Lakes where most human interaction with this unique resource occurs. Excessive eutrophication (algal growth driven by nutrient inputs) causing enhanced levels of nuisance algae including harmful algal blooms has plagued parts of the nearshore. Mathematical models are powerful tools to identify the main drivers of nearshore eutrophication and to assist stakeholders with management options to improve or maintain water quality. Here we examine new modeling methodologies to further our understanding of these complex system. Our preliminary results show the potential for hybrid models combining grid and agent based approaches to give a richer, more realistic view of the nearshore system to managers and researchers.
Description:
As part of an ongoing project aiming to provide useful information for nearshore management (harmful algal blooms, nutrient loading), we explore the value of agent-based models in Lake Michigan. Agent-based models follow many individual “agents” moving through a simulated system. Individually simulated fish in a modeled stream are a common example of agent-based modeling, but modern computing power allows such large numbers of agents to be processed they can approximate continuous phenomena like nutrient distribution in lakes. They complement fixed-cell (or grid) models, and hybrid models combining both representations are possible. By tracking identity and state (time and place of origin, internal state, interaction with environment, and interaction other agents) over time, agent-based models make it easy to ask and answer complex questions. What's the residence time of tributary water in the nearshore? How do lake wide flows differ seasonally and inter-annually? How is water exchanged between Green Bay and Lake Michigan? We include examples of state of the art visualization techniques - combining model results with advanced visualization techniques helps user interpret and ultimately apply data to their problems.