Using participatory agent-based models to measure flood managers' decision thresholds in extreme event response

Extreme flooding impacts to coastal cities are not only a function of storm characteristics, but are heavily influenced by decision-making and preparedness in event-level response. While recent advances in climate and hydrological modeling make it possible to predict the influence of climate change on storm and flooding patterns, flood managers still face a great deal of uncertainty related to adapting organizational responses and decision thresholds to these changing conditions. Some decision thresholds related to mitigation of extreme flood impacts are well-understood and defined by organizational protocol, but others are difficult to quantify due to reliance on contextual expert knowledge, experience, and complexity of information necessary to make certain decisions. Our research attempts to address this issue by demonstrating participatory modeling methods designed to help flood managers (1) better understand and parameterize local decision thresholds in extreme flood management situations, (2) collectively learn about scaling management decision thresholds to future local flooding scenarios and (3) identify effective strategies for adaptating flood mitigation actions and organizational response to climate change-intensified flooding. Our agent-based system dynamic models rely on expert knowledge from local flood managers and sophisticated, climate change-informed hydrological models to simulate current and future flood scenarios. Local flood managers from interact with these models by receiving dynamic information and making management decisions as a flood scenario progresses, allowing parametrization of decision thresholds under different scenarios. Flooding impacts are calculated in each iteration as a means of discussing effectiveness of responses and prioritizing response alternatives. We discuss the findings of this participatory modeling and educational process from a case study of Boston, MA, and discuss transferability of these methods to other types of climate-change influenced systems.