Multiple Models for Management in the Chesapeake Bay. STAC Publication Number 14-004

Abstract

In early 2012, the Director of the Chesapeake Bay Program (CBP) asked the Program s Scientific and Technical Advisory Committee to consider how the CBP might use multiple models. That request resulted in two workshops. The first detailed a pilot multiple modeling project for the Bay s shallow waters (Friedrichs et al. 2012). The second workshop, summarized here, sought to develop more general recommendations on how the CBP could utilize multiple models in management decisions. The workshop considered case studies of multiple models in environmental regulation, the perception of multiple models by the public and decision makers, and legal issues associated with multiple modeling in regulatory settings. This report presents the Findings (factual conclusions or the consensus of workshop experts about the use and benefits of multiple modeling) and four Recommendations that the CBP could initiate to begin realizing those benefits: Findings 1. Using multiple models offers many documented advantages over analyzing one model of an environmental system. 2. There are different ways to implement multiple models (multi-model ensembles, using other models to assess a decision model, modular community modeling, and model comparisons in pilot studies or testbed areas). The common principle is that findings are stronger when multiple lines of evidence, multiple data sets, or multiple algorithms agree. 3. Analyzing multiple models increases knowledge and understanding of underlying processes. 4. Average predictions from a set of models typically perform better than those from any single model. 5. Information from multiple models can help quantify model uncertainty, which is critical to sound science and rational decision-making. 6. Modeling is inexpensive compared to the costs of monitoring, implementation, and poor decisions. 7. Properly framed multiple models can be a legal asset rather than a liability. 8. Managers and the public are poorly informed about modeling, model uncertainty, multiple models, and the value of investments in modeling. 9. Multiple modeling can expand opportunities for additional technical experts and non-technical stakeholders to participate in modeling, fostering greater understanding and acceptance of models and the decisions based upon them. 10. Multiple models of the Chesapeake Bay and its watershed already exist and they could be integrated into CBP modeling to improve knowledge and decision-making. Recommendations 1. The CBP should implement a multiple modeling strategy for each major decision-making model of the Bay (airshed, land use, watershed, and estuary) and analyze the output to quantify skill, advance knowledge, and inform adaptive management. 2. The CBP should exercise the multiple model systems developed under Recommendation 1 to quantify model uncertainty and confidence in key predictions used in decision-making. 3. The CBP should estimate and better communicate the appropriate levels of spending on monitoring, modeling, and research relative to the costs of implementation and the cost of poor decision-making. 4. The CBP should implement ways to better communicate modeling, uncertainty, and multiple model results to partners, decision makers, and the public.