Ecosystem Modeling of Coastal Acidification and Hypoxia and Structural Uncertainties in the Representation of Sediment-Water Exchanges
John C Lehrter1, Lisa Lowe2, Dong S Ko3, Arnaud Laurent4, Katja Fennel5, Dubravko Justic6, Lixia Wang6, James J Pauer7, Brandon Jarvis8, David L Beddick Jr.8, Richard Devereux9 and Wei-Jun Cai10, (1)Environmental Protection Agency Gulf Breeze, Gulf Breeze, FL, United States, (2)CSRA, Research Triangle Park, NC, United States, (3)Naval Research Laboratory, Stennis Space Center, MS, United States, (4)Dalhousie University, Halifax, NS, Canada, (5)Dalhousie University, Department of Oceanography, Halifax, NS, Canada, (6)Louisiana State University, Oceanography and Coastal Sciences, Baton Rouge, LA, United States, (7)US Environmental Protection Agency, ORD/NHEERL/MED, Grosse Ile, MI, United States, (8)US EPA, Gulf Breeze, FL, United States, (9)US Environmental Protection Agency Research Triangle Park, Durham, NC, United States, (10)University of Delaware, School of Marine Science and Policy, Newark, DE, United States
Abstract:
Numerical ecosystem models of coastal acidification (CA) and hypoxia have been developed to synthesize current scientific understanding and provide predictions for nutrient management. However, there is not a scientific consensus about the structure of these models nor is structural uncertainty generally characterized. Here, we examine model structural uncertainties and knowledge gaps related to the role of sediment biogeochemistry in the occurrence of CA and hypoxia. We evaluate five sediment-water exchange model structures that range in complexity. From most to least complex, these include 1) a high vertical resolution (400 layers) method-of-lines sediment diagenesis model that represents the electron-accepting processes and calculates sediment-water exchanges, carbonate chemistry, and sediment burial, 2) a single-layer sediment model of sediment organic matter, O2, and nutrient processes, 3) a sediment boundary flux meta-model derived from a sediment-diagenesis model that relates sediment fluxes to modeled organic matter sedimentation rates and bottom-water concentrations of O2 and nutrients; 4) a bottom-water instantaneous remineralization model based on modeled sedimentation of organic matter and reaction stoichiometry and 5) an empirical model of sediment-water exchanges of O2, DIC, and nutrients developed from observed relationships with bottom-water O2 and nutrient concentration. We evaluate the five sediment model structures in an application of the Coastal General Ecosystem Model (CGEM) to the northern Gulf of Mexico, which encompasses the major influence of the Mississippi River and where a relationship between nutrients, CA, and hypoxia has been observed. The results highlight tradeoffs between complex and simple models where the complex model structures have more realistic lags and feedback responses with the water-column, but at a greater computational cost. Ultimately, we seek to understand tradeoffs between model complexity and prediction skill.
