PARENT SESSION
Thursday, August 10, 1:30-5:00 pm
COS 97 - Statistical ecology
L-4, Lobby Level, Cook Convention Center
Presiders: D Ford

Bayesian uncertainty analysis in modeling net ecosystem CO₂ exchange with an ecosystem process model.

Mitchell, Stephen^*,1, Beven, Keith², Dean, Sarah², Freer, Jim², ¹ Oregon State University, Corvallis, OR, USA² Lancaster University, Lancaster, United Kingdom

ABSTRACT- Net ecosystem CO₂ exchange (NEE) is typically measured directly by eddy covariance towers or is estimated by ecosystem process models, yet comparisons between the data obtained by these two methods often show poor correlation. There are two potential explanations for this discrepancy. First, there could be fundamental problems in model structure that prevent an accurate simulation of NEE. Second, ecosystem process models are dependent on ecophysiological parameter sets derived from field measurements in which a single parameter for a given species can vary by as much as 500%. The latter problem suggests that with such broad variation among multiple inputs, any ecosystem modeling scheme must account for the possibility that there is an infinite number of parameter sets that could preclude the model from emulating the observed NEE dynamics of a terrestrial ecosystem, as well as the possibility that there may be many parameter sets within a particular model structure that can successfully reproduce the observed data. We address this issue of multidimensional parameter variability by adapting the generalized likelihood uncertainty estimation (GLUE) methodology for a widely used ecosystem process model, BIOME-BGC. This procedure involved 50,000 model runs, each with randomly generated parameter values from a uniform distribution of published parameter ranges, resulting in estimates of NEE that were compared to observed daily NEE data from an old-growth Ameriflux site in Metolius, Oregon, USA. While BIOME-BGC was able to provide reasonable estimates of additional outputs such as net primary production (NPP) on an annual time scale, there was a uniform failure of BIOME-BGC to emulate daily NEE data, indicating fundamental shortcomings in the ability of this model to produce realistic carbon flux data. Such results suggest a reevaluation of ecosystem process model assumptions, and possibly a rethinking of eddy covariance methodologies.

Key words: net ecosystem exchange, Ameriflux, BIOME-BGC