A Model for Analyzing Lake Water Acidification on a Large Regional Scale. Part 2: Regional Application

The International Institute for Applied Systems Analysis is developing a computer model which can be used by decision makers to evaluate policies for controlling the impact of acid rain in Europe. As part of this task, a dynamic model has been developed for describing the processes leading to acidification of surface waters. The modeling philosophy is to use a simplified approach, which is warranted for a broad geographical scope. The simulation model is constructed of several modules, each of them providing an overview of a particular aspect of lake acidification. Because of the very sparse input data available on a large regional scale, a new method is applied for estimating unknown inputs. The model is calibrated to presentday conditions by selecting input combinations from feasible ranges. Monte Carlo techniques are used to determine those combinations of inputs that produce the observed present-day lake acidity distribution, when the model is driven by a specified deposition. The ensembles obtained in the calibration procedure for each lake region are used for the scenario analysis. The usefulness of the method is compared with respect to the traditional a prtori parameter estimation technique. Results of sensitivity and uncertainty analysis are used to compare model predictions with observed values and to indicate where changes in the distributions of model parameters will affect predictions the most