Comparing drinking water treatment costs to source water protection costs using time series analysis. Issue 11 (5th November 2015)

• Record Type: Journal Article
• Authors: Heberling, Matthew T.; Nietch, Christopher T.; Thurston, Hale W.; Elovitz, Michael; Birkenhauer, Kelly H.; Panguluri, Srinivas; Ramakrishnan, Balaji; Heiser, Eric; Neyer, Tim
• Is Part Of: Water resources research. Volume 51:Issue 11(2015:Nov.)
• Page Start: 8741
• Page End: 8756
• Publication Date: 2015-11-05
• Subjects: water treatment costs -- source water protection -- time series analysis -- error correction model -- Ohio -- turbidity; Hydrology -- Periodicals; 333.91
• DOI: 10.1002/2014WR016422
• Languages: English
• ISSNs: 0043-1397

Abstract

Abstract: We present a framework to compare water treatment costs to source water protection costs, an important knowledge gap for drinking water treatment plants (DWTPs). This trade‐off helps to determine what incentives a DWTP has to invest in natural infrastructure or pollution reduction in the watershed rather than pay for treatment on site. To illustrate, we use daily observations from 2007 to 2011 for the Bob McEwen Water Treatment Plant, Clermont County, Ohio, to understand the relationship between treatment costs and water quality and operational variables (e.g., turbidity, total organic carbon [TOC], pool elevation, and production volume). Part of our contribution to understanding drinking water treatment costs is examining both long‐run and short‐run relationships using error correction models (ECMs). Treatment costs per 1000 gallons (per 3.79 m 3 ) were based on chemical, pumping, and granular activated carbon costs. Results from the ECM suggest that a 1% decrease in turbidity decreases treatment costs by 0.02% immediately and an additional 0.1% over future days. Using mean values for the plant, a 1% decrease in turbidity leads to $1123/year decrease in treatment costs. To compare these costs with source water protection costs, we use a polynomial distributed lag model to link total phosphorus loads, a source water quality parameter affected by land use changes, to turbidity at the plant. We find the costs for source water protection to reduce loads much greater than the reduction in treatment costs during these years. Although we find no incentive to protect source water in our case study, this framework can help DWTPs quantify the trade‐offs. Key Points: Determine how treatment costs change due to changes in source water quality Error correction model estimates short and long‐run effects on treatment costs Polynomial distributed lag model translates watershed load to treatment costs