Tuesday, June 5, 2007

An Agent-Based Model of Floodplain and Aquifer Biogeochemistry

Alison P. Appling1, Emily S. Bernhardt1, Geoffrey C. Poole2, John S. Kimball3, and Jack A. Stanford3. (1) University Program in Ecology, Duke University, PO Box 90338, Durham, NC 27708, (2) Eco-metrics, Inc.; Institute of Ecology, University of Georgia; & Flathead Lake Biological Station, University of Montana, 2520 Pine Lake Road, Tucker, GA 30084, (3) Flathead Lake Biological Station, The University of Montana, 32125 Bio Station Lane, Polson, MT 59860-9659

We are modeling floodplain biogeochemistry by linking an ecosystem process model, Biome-BGC, to a spatially-explicit, agent-based model of aquifer resources (water, heat, carbon, and nitrogen), the Waterborne Resource Exchange Network (WREN). Nutrient transfers between floodplain aquifers and riparian plant communities mediate the quantity and quality of river C and N entering floodplains and rivers; Biome-BGC describes effects of hydrology on vegetation community structure and nutrient cycling at the floodplain surface. Especially in deep aquifers, thermodynamic conditions along hyporheic flowpaths may significantly transform dissolved nutrients as they are transported through an aquifer. WREN models the hydrologic transport and microbial transformation of nutrients, representing the floodplain ecosystem as a network of linked agents that store, transform, and exchange resources (eg. C and N) according to Michaelis-Menten uptake dynamics. We are integrating and adapting these two models to describe nutrient cycling in the Nyack floodplain of the Middle Fork Flathead River in northwestern Montana. Here we present our modeling methods and preliminary results. The integrated model will be evaluated and refined based on predicted vs. observed variations in aquifer patch C and N concentrations over diel and seasonal timescales, ultimately improving our understanding of floodplain biogeochemical dynamics and complexity.