The Design of Constructed Salt Marshes under Deep Uncertainty

Using nature in engineering design is a long-standing goal of water resource managers. Landscape features like salt marshes serve critical ecological functions in coastal environments delivering important water quality, erosion control and flood inundation benefits. Despite their benefits, natural coastal features like salt marsh have been decimated by development and pollution, leaving much of our coastline unprotected. The diverse benefits of these nature-based features (NBF) make them attractive components of traditional water resource project designs. However, the reality of including these features in engineered designs is complicated by physical and biological uncertainties that limit their use. In this presentation, the development of a new framework methodology for incorporating NBF into engineering designs is described. A San Francisco (SF) Bay demonstration study evaluating constructed salt marsh designs is also presented. Under the proposed framework, uncertain variables are evaluated across ranges of plausible values and decomposed into constituent parts where possible. For instance, in the SF Bay case study, paleo-data informed stochastic weather and water level generators are used to evaluate the range of plausible climate scenarios and decompose total water level uncertainty into sea level rise, tidal variability, seasonal-to-interannual still water level anomalies, surge, etc. Scenario discovery is presented as a component of the framework used to evaluate the robustness of candidate designs. For instance, the framework can explore the ability of a salt marsh to maintain its wave attenuation performance across a range of plausible future climate and sea level rise scenarios. A unique aspect of this study is that biological uncertainties such as vegetation composition, growth rate, density, and mortality, are evaluated as uncertain parameters, and thus allowed to co-evolve with physical and design parameters such as salt marsh size, shape, elevation, and orientation, in a series of computational experiments.