Understanding and modeling the impacts of climate change on source water quality and utility planning

Climate variability and change affect the hydrologic cycle, which impacts drinking source water quality. Shifts in water quality can limit the usage of source waters, influencing treatment and management decisions. While significant efforts are underway to incorporate climate information into water availability planning, little has been done to extend this to water quality. To address this deficiency, tools are developed to help inform process and planning decisions based on the effects of climate on water quality.;Water quality is often evaluated in terms of thresholds that have been set by regulatory agencies or identified as limits to particular treatment options. Thus, a probabilistic framework is presented in which the risk of a water quality exceedance can be obtained. Wherein, a local logistic regression is used to model and subsequently evaluate the threshold exceedances conditional on a streamflow ensemble. Streamflow ensembles are constructed in two ways: (i) using probabilistic seasonal climate forecasts in an average streamflow resampling scheme, and (ii) incorporating climate change projections into a nonstationary generalized extreme value (GEV) model to estimate maximum streamflow. The approach is applied to a drinking water source in Oregon, where elevated turbidity values are correlated with high streamflows in their primary supply, forcing them to switch to a more expensive back-up source to maintain regulatory compliance.;New source development also motivates consideration of the effects of climate on water quality. Hence, an integrated stochastic framework is offered to simulate paired streamflow and water quality under a range of climate scenarios for use in a decision context. Outcomes from the main steps include: (i) a functional relationship between streamflow and water quality, (ii) streamflow ensembles derived from nonparametric resampling, and (iii) consequently, water quality ensembles resulting from the evaluation of the streamflow ensembles in conjunction with the historic relationship. This is demonstrated on a Colorado municipal water provider that is developing a new water source. Here, the salinity concentrations vary with streamflow and pose limits to its use. Resulting simulations characterize both natural variability and climate change uncertainty, and are considered in a cost assessment application.