Probabilistic predictions of hydrologic services in ungauged basins and under global change

Many ecosystem services can be referred to as hydrologic services because their availability depends on the actions of the watershed on the precipitation it captures. Assessing the impact of climate change on hydrologic services requires the application of hydrological models to obtain streamflow predictions to derive indicators of ecosystem health. Successful assessment of the future health of ecosystem therefore depends upon reliable predictions of streamflow along with quantification of the uncertainty of these predictions. We use a probabilistic framework to derive streamflow predictions in a changing climate wherein we quantify the impact of climate on hydrologic model parameters, and hence the watershed. The strategy extends a regionalization approach (used for predictions in ungauged basins) by using spatial variability as a first order approximation for temporal variability of catchment behavior. A spatial model of hydrologic signatures is used to condition a watershed model in a Bayesian framework. Using the resulting probabilistic streamflow predictions, indicators of hydrologic alteration (IHA) and other ecological indices are derived. These indicators convert the information embedded in the probabilistic streamflow predictions into probabilistic predictions of indicators of ecosystem health. The framework is tested against observed ecological indicators for Montana. We further explore the coupling of the hydrologic model with an economic model to assess changes in the economic value of hydrologic services with time. The sensitivity of the coupled model to different future management scenarios is demonstrated.