Uncertainty in Modeling Regional Environmental Systems: The Generalization of a Watershed Acidification Model for Predicting Broad Scale Effects

Policies aimed at the alleviation of negative environmental impacts have increasingly been based on predictions made with models of the environmental systems involved. The usefulness of these models is limited by many uncertainties, e.g., is the complexity of a system properly reflected in the model structure and chosen aggregation-level? Is the data used in the model representative of the system? Is the temporal and spatial scale used in the model appropriate for understanding the system's behavior? The combination of these uncertainties may lead to unexpected results. This study identifies key factors determining watershed responses to acid deposition in different regions, using the RAINS Lake Acidification Model. The study aims at: -- Providing a method for defining a region such that environmental policy directed at alleviating watershed acidification will become more suitable for the large scale management of surface water quality. -- Depicting regional characteristics that allow for the usage of a less detailed and thus more aggregated model. -- Providing an operational concept of critical loads for policy insight into watershed quality as a function of the spatial zoning of watersheds. The results of the study are: -- The 5-step method of flexible zoning introduced in this study, allows for a probabilistic investigation of the compatibility between models and available spatial data. -- The calibration of models to previously defined regions may be of limited use for policy purposes because predictions of environmental effects (i.e., watershed acidification) as a result of changing deposition patterns over large regions may be error prone. -- Cumulative distributions of model predictions about the acidification of watersheds should be used to assess critical and target loads for broad regions