Assessing the Potential Extent of Damage to Inland Lakes in Eastern Canada due to Acidic Deposition. IV. Uncertainty Analysis of a Regional Model

Publisher Website
Google Scholar
Abstract
In this paper, we consider the significance of uncertainty in inputs to the regional surface water acidification model of Marmorek et al. (1990. Can. J. Fish. Aquat. Sci. 47: 55–66). Monte Carlo methods are used to propagate input uncertainty through to measures of uncertainty in the model's overall predictions. We consider four key inputs which represent primary areas of input uncertainty: acidic sulphate deposition (Da), original lake sulphate concentrations ([SO4]0), a watershed neutralization parameter (FW), and an in-lake neutralization parameter (SS). The analysis is repeated for secondary watersheds from north-central Ontario, the Quebec Laurentians, and insular Newfoundland. Uncertainty in Da and [SO4]0 consistently overwhelms the effect of uncertainty in FW and SS on overall model uncertainty. The relative importance of Da decreases as one moves east from the high-deposition regions of central Ontario. We discuss the implications of our results for priority setting for future acidification research and monitoring.