Transit times link pollution sources to drinking water quality in a "One Water" system

Abstract

Innovative approaches are needed to manage chronic and emerging water quality challenges in communities that rely on treated wastewater and urban stormwater as sources of raw water for drinking water treatment, or “One Water” systems. When amended to account explicitly for upstream versus distributed inflow to the reservoir, we show that unsteady transit time theory links pollution sources to water quality in the Occoquan Reservoir (Virginia, USA), one of the largest and oldest One Water systems in the United States. Using 11 years of hydrologic and water quality data, the model identified distinct sources and transformation rates for reactive (nitrate) and relatively non-reactive (sodium, chloride) solutes. High predictive skill was achieved with a strikingly small number of parameters: two for sodium and chloride (one for the upstream storage selection function, one for solute input from distributed sources; Nash–Sutcliffe Efficiency (NSE) = 0.65 and 0.76) and two additional for nitrate (capturing seasonal denitrification linked to summer stratification and hypolimnetic processes; NSE = 0.55). The simplicity of unsteady transit time theory supports rigorous parameter estimation (Bayesian Markov Chain Monte Carlo) and model structure evaluation (Bayesian Information Criterion). It also opens the door to real-time interactive simulations with stakeholders, supporting collaborative solutions to cascading water quality challenges.