95th ESA Annual Meeting (August 1 -- 6, 2010)

OOS 43-1 - Data assimilation and ecological forecasting in a data-rich era

Thursday, August 5, 2010: 8:00 AM
401-402, David L Lawrence Convention Center
Yiqi Luo, Department of Microbiology and Plant Biology, University of Oklahoma, Norman, OK, Kiona Ogle, School of Life Sciences, Arizona State University, Tempe, AZ, Colin Tucker, Botany and Program in Ecology, University of Wyoming, Laramie, WY, Shenfeng Fei, Department of Botany and Microbiology, University of Oklahoma, Norman, OK, Shannon L. LaDeau, Cary Insitute of Ecosystem Studies, Millbrook, NY, James Clark, Nicholas School of the Environment, Duke University, Durham, NC and David S. Schimel, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA
Background/Question/Methods Several forces are converging to transform ecological research and increase its emphasis on quantitative forecasting. These forces include: (i) dramatically increased volumes of data from observational and experimental networks, (ii) increases in available computational power, (iii) advances in ecological models and related statistical and optimization methodologies, and, most importantly, (iv) societal needs to develop better strategies for natural resource management in a world of ongoing global change. Traditionally, ecological forecasting has been based on process-based models, informed by data in largely ad hoc ways. Although most ecological models incorporate some representation of mechanistic processes, today’s ecological models are generally not adequate to quantify real-world dynamics and provide reliable forecasts with accompanying estimates of uncertainty.  A key tool to improve ecological forecasting is data assimilation, which uses data to inform initial conditions and to help constrain a model during simulation to yield results that approximate reality as closely as possible.

Results/Conclusions Data assimilation can advance ecological forecasting by providing (i) improved estimates of model parameters and state variables, (ii) quantification of uncertainties arising from observations, models and their interactions, (iii) selection of alternative model structures, and (iv) evaluation of alternative data observing or experimental strategies for future implementation.  In an era with dramatically increased availability of data from observational and experimental networks, data assimilation is a key technique that helps convert the raw data into ecologically meaningful products so as to accelerate our understanding of ecological processes, test ecological theory, forecast changes in ecological services, and better serve the society.