Analysis of Uncertainty in Model Predictions for Lake Lanier
Osidele, Olufemi O.
Beck, M. B.
MetadataShow full item record
In this paper we summarize a comprehensive analysis of model-based predictions that accounts for uncertainties due to: (i) the internal process parameters, (ii) the temporal pattern of variation of external inputs and forcing functions, and (iii) the initial state of the system. We apply the Regionalized Sensitivity Analysis (RSA) and Tree-Structured Density Estimation (TSDE) procedures to a generalized ecosystem model of Lake Lanier, Georgia. Our results suggest that the seasonal variation in watershed inputs, water temperature, and the timing of life-history stages of top predators, are as critical to predicting lake eutrophication as the internal ecosystem processes. These procedures are useful in ranking priorities for policy actions and directing the focus of scientific research on environmental systems.