(a) Small watershed with thin soil. (b) Large watershed with thick soil.
How much confidence can be attributed to predictions from any single model in environmental assessments?
In this study, investigator bias was eliminated and Monte Carlo uncertainty analysis applied to three commonly used watershed acidification models: the Integrated Lake-Watershed Acidification Study (ILWAS), the Model of Acidification of Groundwater in Catchment (MAGIC), and the Enhanced Trickle-Down (ETD). Model forecasts of the acid-neutralizing capacity (ANC) of lakes under increased acidification scenarios were similar when predictions were viewed on a relative scale (e.g., changes in ANC), but differed when viewed on an absolute scale (e.g., number of lakes with ANC less than some value).
Confidence levels in model forecasts can be improved by comparing predictions from multiple models. If the models are based on sufficiently different hypotheses but predictions agree, then there may be a higher confidence than in predictions from any single model.
Rose, K. A., R. B. Cook, A. L. Brenkert, R. H. Gardner, and J. P. Hettelingh. 1991. Systematic comparison of ILWAS, MAGIC, and ETD watershed acidification models: 1. Mapping among model inputs and deterministic results. Water Resources Research 27:2577- 2603.
Rose, K. A., A. L. Brenkert, R. B. Cook, R.H. Gardner, and J. P. Hettelingh. 1991. Systematic comparison of ILWAS, MAGIC, and ETD watershed acidification models: 2. Monte Carlo analysis under regional variability. Water Resources Research 27: 2591-2603.
Cook, R. B., K. A. Rose, A. L. Brenkert, and P. F. Ryan. 1992. Systematic comparison of ILWAS, MAGIC, and ETD watershed acidification models: 3. ANC mass balance budgets. Environmental Pollution 77:235-242.
Integrated Assessment Briefs. 1995. ORNL/M-4227. Oak Ridge National Laboratory, Oak Ridge, TN.