Modelling the recovery of acidified lake catchments in central Ontario incorporating Bayesian calibration and Monte Carlo uncertainty assessment

Acidification of soils and surface waters remains an environmental threat to ecosystems in sensitive regions of Canada. Environmental models contribute to acid emissions management by predicting the impact of proposed strategies on affected ecosystems, although the utility and credibility of model predictions can be improved with increased clarity around the accuracy, reliability and sensitivity to inputs of the models. The objective of this work was to assess the recovery of acidified lake catchments in well studied areas of central Ontario using the Model of Acidification of Groundwater In Catchments within a framework that performs Bayesian calibration and Monte Carlo uncertainty assessment. A single site assessment was conducted at Plastic Lake to assess the suitability of the model framework for use at a regional level. This was followed by a multi-site application in central Ontario under two future acid deposition scenarios current legislation and maximum feasible reduction. Despite high uncertainty in some of the model inputs, by calibrating to many years of surface water chemistry simultaneously, the framework calibration performance and model predictions compared well with previous studies in the region. When applied to 20 catchments with at least 5 years of surface water chemistry, limited recovery from acidification is expected under current legislation while more rapid and complete surface water chemistry recovery was shown to be possible under a maximum feasible reduction scenario. All model forecasts included a propagated assessment of uncertainty due to the model parameters and input data enabling additional and valuable insights to be drawn from the model simulations.