Classic indicators and diel dissolved oxygen versus trend analysis in assessing eutrophication of potable‐water reservoirs

Abstract Potable source‐water reservoirs are the main water supplies in many urbanizing regions, yet their long‐term responses to cultural eutrophication are poorly documented in comparison with natural lakes, creating major management uncertainties. Here, long‐term discrete data (June 2006–June 2018) for classical eutrophication water quality indicators, continuous depth‐profile data for dissolved oxygen (DO), and an enhanced hybrid statistical trend analysis model were used to evaluate the eutrophication status of a potable source‐water reservoir. Based on classical indicators (nitrogen, N and phosphorus, P concentrations and ratios; phytoplankton biomass as chlorophyll a , chl a ; and trophic state indices), the reservoir was eutrophic to hypereutrophic and stoichiometrically imbalanced. Anoxia/hypoxia occurred for 7–8 months annually systemwide, even throughout the water column for days to weeks in some years; and elevated total ammonia (up to ~900 μg tNH 3 L −1 ) in surface waters from late summer/fall through late winter/early spring suggested substantial internal legacy nutrient loading. These surprising DO and tNH 3 phenomena may characterize many reservoirs in urbanizing areas, and the associated cascade of negative impacts may increasingly affect them under global warming. Total organic carbon (TOC), seasonally influenced by phytoplankton biomass, commonly exceeded 6 mg L −1 , which is problematic for potable‐water treatment, and significantly trended up over time. Wet‐year inflow dilution influenced an apparent decreasing trend in nutrients within the hypereutrophic upper reservoir, which receives most tributary inputs. Nevertheless, significant reservoirwide trends (increasing total phosphorus [TP], phytoplankton chl a , TOC) and mid‐ and/or lower region trends (increasing total nitrogen [TN], tNH 3 , decreasing TN:TP ratios) suggest that water quality degradation from eutrophication has worsened over time. These findings support broadly applicable recommendations to strengthen protection of potable source‐water reservoirs in urbanizing watersheds: (1) protective numeric water quality criteria are needed for TOC as well as TN, TP, and chl a ; (2) continuous diel data capture more realistic DO conditions than traditional sampling, and can provide important insights for water treatment managers; and (3) assessment of reservoir eutrophication status to track management progress over time should emphasize classic indicators equally as statistical trends, which are highly sensitive to short‐term meteorological forcing