Journal American Water Works Association February 2017 Volume 109 Number 2

JournalFebruary 2017 Volume109 Number2 - American Water Works Association Brackish Water Reverse Osmosis: A Cost-Effective and Renewable Supply State Approaches for Cyanotoxins in US Drinking Water Protecting Drinking Water Utilities From Cyberthreats Innovative Reuse Strategy in Florida’s Unique Hydrogeology The Effects of Preozonation on Organic Fouling in a Membrane Process Estimating the National Costs of Regulating Perchlorate in Drinking Water Creating a Data-Driven Turnover Response Plan Total Intensity of Odor: A New Method to Evaluate Odors

one sourceThe best in Smart Water flows fromplan monitor Flood Forecasting & Simulation design Multi-Species Analysismanage Green Infrastructure OptimizationIntegrated Catchment ModelingProject Collaboration Asset Condition Assessment& Data Management & Prioritization Demand Analysis & Forecasting

Real-Time OperationsReal-Time Modeling Real-Time Performance Monitoring Unidirectional Flushing operateSurge/Transient Analysis sustainWhole Life Cycle Costing optimize protectFor two decades, Innovyze has continued to boldly set and raise the bar for water and wastewater infrastructure softwareengineered for smart water innovation. Today, we deliver the world’s most complete, open smart water modeling solutions —packed with the speed and power you need to competitively plan, manage, design, protect, operate and sustain highly efficientand resilient infrastructure systems in the face of unprecedented challenges. Factor in our legendary stellar customer care,and it’s clear why Innovyze is the premier choice for top utilities, government agencies, and engineering organizations worldwide. To learn more about our advanced water modeling solutions, call 626 568 6868, e-mail [email protected], or visit us at www.innovyze.comCopyright © 2017 Innovyze, Inc. DemandAnalyst, DemandWatch, FloodWorks, InfoMaster, InfoNet, InfoSurge,InfoSWMM, InfoWater, InfoWorks, ICMLive, IWLive, SWMMLive, Model360 and SCADAWatch are registeredtrademarks of Innovyze, Inc. All other marks are the property of their respective owners.

On Water KENNETH L. MERCER FEBRUARY 2017 • Vol. 109, No. 2 & WorksOn Water and Works Editor-in-Chief EDITORIAL Kenneth L. Mercer, PhDL ast month, I described two ways change occurs in the safety-focused/risk-averse water industry: namely, in response to cataclysmic, mostly negative events with Senior Editorial Manager Kimberly J. Retzlaff broad external visibility, and less noticeably, through continuous improvements at the margins. Regardless of the driver, what’s ultimately pursued is not change SeniorTechnical Editor Maureen Peckfor change’s sake, but improvement for the greater good. Contributing Editors Maripat Murphy Carina Stanton The changes you may have already noticed with this month’s Journal were not driven Holly E.Trippby a negative shock, but through a concerted effort to improve AWWA’s flagship publi- Jenifer F. Walkercation. To this end, the cover has assumed more of a table-of-contents view that should Publishing Coordinator Cindy Ubaallow you to quickly find what you want to read first. It’s also a nod to the past whenthe Journal cover was a direct read of what was “In this Issue.” PRODUCTION We have moved the peer-reviewed articles forward in the printed version to increase the Senior Graphic Designer Daniel Feldmanprominence of original research pertinent to the water industry. Research has always been Senior Production Editor Linda Yeazelthe foundation of the Journal, and moving it forward may foster further communication Production Editors Jan Baileyand potentially collaboration between Journal AWWA’s readers who consist largely of Sandra Lankenauutility members, consultants, regulators, researchers, and manufacturers. While it’s impor- Contributing Artists Megan McCarthytant to promote the exchange of information between researchers (academic impact), the Gillian WinkJournal also serves as a bridge between the various groups who make up the bulk of our Melanie Yamamotoaudience who need guidance to tackle pressing issues (industry impact). In addition,although most of the peer-reviewed original research will appear as one-page summaries MARKETINGin print, at least one article will be presented in its entirety each month. Sales Project Manager Karen Pacyga Journal AWWA strives to be an internationally acknowledged authority in the water Advertising Coordinator Connor Larsonindustry, and we will continue to focus on the science, engineering, and management ofwater supply, treatment, and distribution, as well as how these are impacted by reuse, TERRITORY SALES MANAGERSstormwater, and wastewater. Submissions are received from across the globe, but edito-rially, the Journal steers toward issues that mostly affect developed water systems with a Northeast focus on North America. A big challenge in this area is making sure the peer reviewerscan focus on the technical material and not English grammar. To this end, we are mak-   Ryan Fugler: 303-347-6238ing structural and procedural changes to improve the experience so reviewers’ time andefforts are used most effectively.                      [email protected] Southeast We’re also striving to deepen the technical review of feature articles (i.e., those notreceiving full peer review) and will be looking to potentially expand volunteer opportu-   Pam Fithian:        303-347-6138nities to address this. Non-peer-reviewed—or “feature”—articles provide commentaryon relevant issues and trends to the water industry, but they often communicate techni-                      [email protected] information as well, so our ultimate goal is to ensure these are fair in presentation Midwest and accurate in content.   Nancy Mortvedt:      303-734-3442 Finally, I have revamped the nature of “Inside Insight” from providing a brief reviewof what’s in the issue to presenting a topical exploration relevant to the water industry.                       [email protected] this, the name of this column is now “On Water & Works.” The views presented West here are mine, and I welcome any further discussions they may trigger. Of note, guesteditors will be afforded the opportunity to pen this editorial, too, so look for Stuart   Kathy Smith:       303-347-6237Krasner’s thoughts this June.                       [email protected] Many of the changes you will see in this and coming issues were developed in consul-tation with the Journal Editorial Board (JEB), and I’d like to thank them specifically for Journal - American Water Works Association (ISSN 0003-their efforts to continue the good work of the past while identifying areas for improve- 150X) is published monthly by the American Water Worksment and working to implement these changes. Their names are listed a few pages from Association, 6666 W. Quincy Ave., Denver, CO 80235 USA;this one on page 8 (in print, that is). tele­phone: (303) 794-7711; fax (303) 794-7310; e-mail [email protected]. AWWA assumes no responsibility for Journal staff will ensure this publication remains the voice of the water industry opinions or statements of facts expressed by contributorswhile striving to improve its value to authors, readers, and advertisers. This month, or advertisers, and editorials do not necessarily representJournal AWWA explores topics such as the effect of preozonation on organic fouling, official policies of the association.the cost of perchlorate regulations, and much more. SUBSCRIPTIONS: Periodicals postage paid at Denver, Colo., and additional mailing offices. Subscription to the https://dx.doi.org/10.5942/jawwa.2017.109.0033 Journal is a member benefit; additional subscriptions, sold only to those not eligible for AWWA membership:2 ON WATER & WORKS | 109:2  •  FEBRUARY 2017  |  JOURNAL AWWA $279.00 ($361.00 outside North America). Additional print copy for members (organizational members sent to the same organizational address or members entitled to e-periodicals only) are domestic: $73.00, foreign: $142.00. Single-issue copies are $34.00, plus shipping. MISSING ISSUES: Contact AWWA Customer Service Group concerning any problems with re­ceipt of issues. Claims for missing issues must be submitted upon receipt of the following issue. Allow 90 days for change-of- address notification. INDEXING: Indexed regularly by Applied Science & Technology Index, Biological Abstracts, Chemical Abstracts, Compendex, Pollution Abstracts, Water Resources Abstracts, Environmental Science & Pollution Management, and Waternet. CODEN: JAWWA5 ISSN, print: 0003-150X ISSN, electronic: 1551-8833 POSTMASTER: Send address changes to Journal AWWA, American Water Works Association, 6666 W. Quincy Ave., Denver, CO 80235. Canadian Publications Mail #40612608— Return undeliverable Canadian addresses to Bleuchip International, P.O. Box 25542, London, ON N6C 6B2. Copyright © 2017 by American Water Works Association. All rights reserved. PRODUCED IN USA

PROUDLY AMERICAN. The strongest, longest-lasting hydrant on the market. Founded in 1878, Clow hydrants have been manufactured and assembled in America for more than a century. From the classic Eddy to the cutting-edge Medallion, each of our hydrants is built to meet AWWA, NSF, UL and FM specifications. Because at Clow, we believe something worth doing is worth doing right. Clow Valve Co. | www.clowvalve.com Clow Valve Co. is a division of McWane, Inc. | McWane. For Generations.

FEBRUARY 2017 VOLUME 109 NUMBER 2 JournalFIGURE 4 EEM data for raw (A) and ozonated (B) surface water FIGURE 1 Estimated compliance cost burdena FIGURE 8 Comparison of TIO versus FPA methods for 10 plant ef uent samples A Raw water B Ozonated water400 400 Very small350 (<500) Small 350 (501–3,300) Medium 300 (3,301–10,000)Excitation—nm Fluorescence Fluorescence 12 Excitation—nmintensity—AUintensity—AU System Size—population served 10 1.0 1.0 TIO Value 8 0.8 0.8 6300 0.6 0.6 4 0.4 Large 0.4 (10,001–100,000) 0.2 2250 250 0.2 0 Very large 0 2 4 6 8 10 (>100,000) FPA Value 12 200 0.0 200 300 350 400 04.0500 5001.00550 602.000 0.0 4.00 FPA—flavor profile analysis, SMCL—secondary maximum 300 350 400 450 500 550 600 3.00 Emission—nm 27contaminant level, TIO—total intensity of odor Emission—nm Costs—$/1,000 gal15AU—arbitrary units, EEM—excitation–emission matrix TIO value >4 fails proposed SMCL; FPA value ≥6 is considered 25MCL—maximum contaminant level, O&M—operations and unacceptable. maintenance aamortized capital and O&M costs to individual systems if federal perchlorate MCL is 4 µg/LPeer Reviewed 25 2615 Estimating the National Costs of Data-Driven HR: Creating a Turnover Regulating Perchlorate in Response Plan Based on thePreozonation Effects on Organic Drinking Water Kaplan-Meier EstimatorFoulants in a Coagulation–Ultra ltration Membrane Process A 2009 assessment estimating the The Kaplan-Meier estimator (KM curve) national compliance costs associated is a statistical tool that human resourcesThe low-pressure membranes used in with perchlorate was updated to (HR) professionals can use to makeultrafiltration, which removes turbidity account for utility actions taken in data-driven decisions regarding employeeand suspended solids from drinking water, response to regulations for perchlorate turnover. This article provides backgroundare prone to fouling by organic matter in California and Massachusetts. The on how the KM curve was applied at theand are thus rendered less effective; updated cost assessment reinforces City of Houston’s Public Utilities Divisionvarious pretreatment processes have been concern that the potential impacts on and describes the resulting workforceexplored to control membrane fouling. individual systems are significant, plan. The authors also demonstrate how aThe research discussed in this article particularly for small water systems. KM curve is constructed.examined integrating preozonation with This study demonstrates thatcoagulation and ultrafiltration, comparing opportunity costs for lost water (source Stanley Lam and Aaron Chanthe quality of water that had gone abandonment) are comparable tothrough conventional pretreatment with treatment costs. 27and without preozonation. Caroline G. Russell Total Intensity of Odor: A NewPaul G. Biscardi and Steven J. Duranceau and Kevin M. Morley Method to Evaluate Odors Public Access Peer-Reviewed Articles In light of the disadvantages of two methods—threshold odor number (TON) Each month, the full version of at least one peer-reviewed article and flavor profile analysis (FPA)—for appears in print, along with the expanded summaries of additional testing odors in drinking water, the total peer-reviewed articles that appear in their entirety on the Journal intensity of odor (TIO) method was AWWA website (www.awwa.org/journal). All peer-reviewed articles developed. This article illustrates the need from 1990 to present are available to view online free of charge. for a more accurate odor measurement tool and demonstrates how TIO can eliminate unnecessary testing and could be a workable alternative to TON and FPA in meeting compliance requirements for odor aesthetics. Allison Jacobsen-Garcia, Melissa Dale, Roy Desrochers, and Stuart Krasner

RELIABILITYwhen it’s needed most.350psi A-2361 Resilient Wedge Gate Valve:Easy to Open. Easy to Close. Tough to Replicate.Utilities are concerned with the safe operation and stopping abilityof their vehicles. Without properly selected, installed and maintainedbrakes, the risk of life and property loss greatly increases. Gatevalves in water distribution systems are no different; only the mostreliable gate valves should be used. That’s why water utilities preferMueller’s 350psi A-2361 all-ductile iron resilient wedge gate valve.This triple-listed gate valve is AIS-compliant, easy to handle, and hasa pressure-assist wedge geometry. These unique features combine toassure a rapid seal when it’s needed most. Dual purpose lifting lugs • Improved site safety; aligns valve box and eliminates adaptor Internal components interchangeable with installed A-2300 series valves • No additional inventory required 350psi AWWA/UL/FM working pressure • Meets the increasing demands of higher water main pressures T-head bolt retention • Eliminates the need for anti-rotation bolts Pressure-assist wedge geometry • Less torque required to sealFor more information about Mueller or to learn more about the 350psi ductile iron gate valve,call 1.800.423.1323 or visit www.muellercompany.com.Download the 350psi Download ourproduct brochure! Free App Today!Copyright © 2015 Mueller Co., LLC. All Rights Reserved.The trademarks, logos and service marks displayed in this document herein are the property of Mueller Co., LLC, its affiliates or other third parties.

FEBRUARY 2017 VOLUME 109 NUMBER 2 Journal 28 40 50 Feature Articles 40 50 28 Occurrence and State Approaches Protecting Drinking Water Utilities for Addressing Cyanotoxins From Cyberthreats Brackish Water Reverse Osmosis: in US Drinking Water A Proven Cost-Effective Renewable The US Department of Homeland Security, Water Supply Because cyanotoxins are considered US Environmental Protection Agency, potential drinking water contaminants and AWWA have formed a partnership Faced with continued decreases in but are not regulated under the Safe to enhance cybersecurity for the water in- nonrenewable groundwater levels and the Drinking Water Act, the US dustry. This article summarizes approaches increasing cost to pump the water, the Environmental Protection Agency and common cybersecurity and water East Cherry Creek Valley Water and released health advisories in 2015 to system vulnerabilities and presents a new Sanitation District in Aurora, Colo., provide guidance on microcystin and approach for protecting drinking water turned to brackish water as a renewable cylindrospermopsin levels. A study was systems against hacking and cyberattacks. water source. The authors of this article conducted to investigate observed describe the design of the facility’s cyanotoxin levels and resulting actions Robert M. Clark, Srinivas Panguluri, multistage reverse osmosis treatment by state primacy agencies during the Trent D. Nelson, and Richard P. Wyman system, how management challenges have 2015 cyanotoxin-producing bloom been addressed, and how operation of the season, with the objective of collecting 59 system has demonstrated cost savings. and analyzing data to help direct state guidelines and lay the groundwork for Advanced Oxidation Process, Douglas Brown, Tim Rynders, future data collection. Unique Hydrogeology Allow Chris Stillwell, Chris Douglass, for Innovative Reuse Strategy and Scott Niebur Tarrah Henrie, Sarah Plummer, and J. Alan Roberson To address new regulatory requirements— eliminating ocean outfall and imple-FEBRUARY 2 0 1 7Journal2 February 2017 Volume109 Number2 - American Water On the cover: A menting 20.4 mgd/year of additional Works Association water district in reuse—the Southern Regional Wastewater Colorado has Write for the Journal Treatment Plant in Hollywood, Fla., Brackish Water secured a reliable, underwent a 10-month study to test Reverse Osmosis: renewable water Journal AWWA advanced oxidation processes. The re- A Cost-Effective supply by using is currently sult was an alternative reuse treatment and Renewable reverse osmosis seeking peer- option with roughly half the cost and Supply to treat brackish reviewed and carbon emissions of traditional treat- groundwater. feature articles. Find submission ment methods. State Approaches for Cyanotoxins in guidelines at www.awwa.org/submit. US Drinking Water Benjamin D. Stanford Protecting Drinking Water Utilities and Abigail Antolovich From CyberthreatsVOLUME 109 NUMBER 2 Innovative Reuse Strategy in Florida’s Unique Hydrogeology The Effects of Preozonation on Organic Fouling in a Membrane Process Estimating the National Costs of Regulating Perchlorate in Drinking Water Creating a Data-Driven Turnover Response Plan Total Intensity of Odor: A New Method to Evaluate Odors

Accuracy manager.Revenue generator.Intelligence gatherer.Low-flow pioneer.That also happens to be the provenleader in smart water meters.The Sensus iPERL® smart water meter does all that and more. Deliveringunparalleled low-flow accuracy with high-flow durability. Its innovativemagnetic technology captures previously unmeasured water to drive additionalrevenue. And because iPERL has no moving parts, it operates without frictionor wear. Maintaining accuracy over a 20-year lifespan, your iPERL will continueto measure just as accurately as the day it was installed. It’s also equipped withsmart alarms to detect issues like leaks, tampering and empty pipes. Whenintegrated with our FlexNet® communication network, remotely gathering andtransmitting data has never been more reliable or profitable.Nothing’s out of reach.To drive more revenue from your water network,visit www.sensus.com/iperl-na.

Columns and JOURNAL EDITORIAL BOARD Departments Andrew D. Eaton (chair)64 2 On Water & Works Dulcy M. Abraham67 On Water and Works Joseph J. Bernosky71 Dominic Boccelli 10 Open Channel David E. Bracciano EUM: It’s About Continual David Cornwell Improvement Joseph A. Cotruvo 12 People in the News Christopher S. Crockett 64 Researcher to Researcher Steven Duranceau Richard W. Gullick Impactful and Impact Factor Charles D. Hertz 67 Eco Logic Karl G. Linden Darren A. Lytle Investing in a Water-Secure Future Joan A. Oppenheimer 71 Industry News Christine A. Owen 77 Media Pulse Theresa R. Slifko 79 AWWA Section Meetings John E. Tobiason 80 Product Spotlight 80 Future AWWA Events BACK ISSUES: For any Journal AWWA article from 1914 81 Buyers’ Resource Guide to the present or to order back issues up to 24 months 108 List of Advertisers old, call the AWWA Customer Service Group, 6666 W. Quincy Ave., Denver, CO 80235; 1-800-926-7337. Mission: Journal AWWA communicates the scholarship of the water industry through peer-reviewed original research and practical REPRINTS & PERMISSIONS: To request reprints and perspectives to professionals that manage and treat water. digital permissions, please e-mail [email protected]. Vision: Journal AWWA strives to be an internationally For permission to reproduce data or material from acknowledged authority on the science, engineering, and Journal AWWA, e-mail [email protected]. management of water supply, treatment, and distribution. PHOTOCOPYING CONDITIONS: Authorization to photo- copy items for internal or personal use or for the use of specific clients is granted by AWWA for libraries and other users registered with the Copyright Clearance Center (CCC) Transactional Reporting Service, provid- ed that the base fee of $2 per request is paid directly to CCC, 222 Rosewood Dr., Danvers, MA 01923. Provide the following code with payment: 003-150X/00 $2.This consent does not extend to other kinds of copying, such as copying for general distribution, advertising, creating new collective works, or resale. Statements and opinions expressed herein are those of the authors or contributors and do not necessarily re ect the policies or positions of the association or Journal AWWA. EXECUTIVE, EDITORIAL, PRODUCTION, & ADVERTISING OFFICES 6666 W. Quincy Ave., Denver, CO 80235 303-794-7711 e-mail: [email protected] fax: 303-794-7310 www.awwa.org A PUBLICATION OF THE AMERICAN WATER WORKS ASSOCIATION Dedicated to the world’s most important resource, AWWA sets the standard for water knowledge, management, and informed public policy. David B. LaFrance, Chief Executive Of cer Zsolt G. Silberer, Director of Publishing

Performance MattersWith a time-tested, simple and up-sized chamber design, the Recordall® Model 25 nutating disc water meter continues to set the standard for accuracy and durability. Built on a foundation of more than 110 years of water metering expertise, Recordall disc meters are an ideal solution for measuring flow.To discover how Recordall disc meters can help you generate more revenue for your water utility than other mechanical meter types, visit: www.badgermeter.com/recordall-disc-meters. www.badgermeter.com

Open DAVID B. LAFRANCE, CHIEF EXECUTIVE OFFICER ChannelEUM: It’s About Continual ImprovementI n 2007, the US Environmental Protection Agency (USEPA) and six collaborating organizations sat down utility leaders unlock the potential of their utility by using the and looked at the broad and common challenges fac- keys to effectively manage the utility operational attributes. ing the water and wastewater sector. Things like rising The 2008 EUM framework and guidance document is evergreen—meaning the principles and practices can with-costs, aging infrastructure, workforce issues, and regula- stand the test of time. However, because one of EUM’s vetions were universal concerns. Faced with the magnitude of management keys is “continual improvement,” the collabo-these concerns, these organizations were motivated to respond rating organizations decided it was best to practice what theyand develop a framework that could guide the entire water preached, and in 2016 a slightly larger group of collaborat-sector—water and wastewater, big and small, public and ing organizations (see the sidebar that lists them) met andprivately owned, and geographically dispersed—in navigating examine demographic, environmental, social, technological,the complexities. and other trends and changes that had occurred over theA year later, the result was a guidance document called past decade. The group’s conclusion was that EUM needed a“Effective Utility Management” (released in June 2008). At minor but meaningful tune-up.its core, EUM, as it is commonly called, recognizes that a util- Among the changes that tipped the decision to updateity’s success is a function of managing and addressing several EUM was the growth in “smart” water systems integratingcomplicated and critical operational functions. And although their data across multiple operational and business func-this can seem overwhelming, the collaborating organizations tions, the inescapable complications relating to weather andboiled the list down to 10 comprehensive operational attri- climate variability and its impact on utility operations, thebutes. They then coupled the attributes with ve keys to man- growing awareness of water in the media and by the pub-agement success. These keys are all best management practices lic, and increased and rigorous regulatory conditions. Afterfor leaders and include, for example, leadership, strategic examining the signi cance of these and other factors, the col-business planning, and measurement. The magic comes when laborating organizations were compelled to modify aspects of the 10 attributes (the updated list is shown in the sidebar) while at the same time maintaining the concise yet compre-2016 Collaborating Organizations hensive framework of the original EUM. All utility leaders have the challenge of keeping their long-• American Public Works Association term vision in focus while managing day-to-day issues. This is• American Water Works Association especially true when the external factors that affect the opera-• Association of Clean Water Administrators tions are changing. EUM, now more than ever, is an essential• Association of Metropolitan Water Agencies reference for all utility leaders during these challenging times—• Association of State Drinking Water Administrators its power comes from the balance of succinct guidance coupled• National Association of Clean Water Agencies with the all-inclusive how-to structure, framework, and direc-• National Association of Water Companies tion that can quickly cut through the complicated morass and• US Environmental Protection Agency remind leaders of the 10 most critical things to focus on.• Water Environment Federation AWWA is proud to be one of the original collaborating organizations of EUM and equally as proud to be part of its EUM’s 10 Updated Attributes (2016) 2016 update. We feel that EUM has such value that during 2017, AWWA will launch three two-day seminars on EUM• Product quality fundamentals and will also work with USEPA and other EUM• Customer satisfaction partners to hold a series of three webinars for anyone who• Employee and leadership development wants to learn more about EUM’s framework. If you are• Operational optimization interested in these programs, you can check our website,• Financial viability awwa.org, for dates, times, and locations. Or if you want• Infrastructure strategy and performance more background on EUM, the website watereum.org has• Enterprise resiliency what you need. As a nal note, AWWA members recently pre-• Community sustainability sented the EUM framework to their colleagues in India. It was• Water resource sustainability a huge success and, with some luck, helps set a framework for• Stakeholder understanding and support India’s water transformation. https://dx.doi.org/10.5942/jawwa.2017.109.003210 OPEN CHANNEL | 109:2 • FEBRUARY 2017 | JOURNAL AWWA

Mike WesselsMachining17 Years of Service Mark Wolf Machining 35 Years of ServiceWe do it all. And we do it in the name of customer service. Whether helping More than a brand. We’re a family.you build a new neighborhood or making a one-unit production run for a customapplication, making water work for you is our job. And from the production floorto the president’s office, we know who our boss is. You—the customer. That’s amindset that matters. It’s how our family has been doing business since 1856.And it’s how we make products that make water work.Visit AYMcDonald.com or call 1.800.292.2737 to learn more.

PeinotphleeNews   TRANSITIONS Dewberry, Heintzman served in the US Navy and retired as a captain. He then went on to Roberson The board of directors of the Association of work for the US Public Health Service, Heintzman Safe Drinking Water Advisors (ASDWA) where he helped develop policy for the US Carson announced that J. Alan Roberson was selected Environmental Protection Agency for the Casado as its new executive director; his employ- Safe Drinking Water Act. Heintzman retired Smith ment took effect on Jan. 11, 2017. Roberson after 30 years of government service. has more than 25 years of experience in drinking water legislation, regulations, and Dublin San Ramon Services District policies on a variety of issues. He served (DSRSD; Dublin, Calif.) has hired Jeff most recently as director of policy for Carson as its new operations manager. As Corona Environmental Consulting, where he DSRSD’s senior executive responsible for developed policy positions for utilities and wastewater, drinking water, and recycled government agencies on drinking water reg- water operations, Carson oversees a ulations, served as principal investigator on $14.3 million annual budget and 62 several Water Research Foundation projects, em­ployees. For four years prior, Carson was and collaborated with numerous stakehold- operations and maintenance manager for the ers. Previously he was director of federal City of Hayward’s water pollution control relations at AWWA, where he provided tech- facility, where he helped launch renewable nical and policy input on all aspects of energy and water recycling projects. Carson drinking water regulations. He also worked has 19 years of wastewater industry experi- closely with the US Environmental ence in the Bay Area. Before working for the Protection Agency and the Department of City of Hayward, he was with the Sewerage Homeland Security on implementation of Agency of Southern Marin for four years, the requirements for vulnerability assess- serving as interim general manager and chief ments and emergency response plans. operator. He also worked at Oro Loma Roberson has published more than 35 peer- Sanitary District for 11 years. Carson replaces reviewed works covering a broad range of DSRSD’s former operations manager, drinking water topics. Dan Gallagher, who retired in May 2016. Arcadis North America has appointed Alan Luis Casado has joined Gannett Fleming as its Baxter as the company’s new city executive for southeast region director and senior vice- Chicago, Ill. In his role at Arcadis, Baxter is president. In this role, he is responsible for the responsible for expanding the company’s mar- overall financial, operational, management, ket presence in the Chicago region and leading and administrative performance of eight the firm’s delivery of engineering solutions southeast region offices, including locations and capabilities. He brings more than 30 years in Jacksonville, Miami, Orlando, Tallahassee, of program and project management experi- Tampa, and West Palm Beach, Fla.; New ence to the role. His professional expertise Orleans, L.A.; and Memphis, Tenn. Casado includes environmental consulting, construc- brings to the position more than 24 years of tion management, program development, con- experience in which he has led complex infra- tract negotiations, building assessments, and structure projects across the southern United architectural design. For the past nine years, States and in Central and South America. His Baxter has served as a program executive in professional portfolio includes projects com- the Arcadis pharmaceutical program. pleted for clients in the water, earth science, construction services, and geographic infor- Dewberry announced that Daniel Heintzman mation systems markets. has joined the firm’s Denver, Colo., office as a senior project manager, where he will sup- WSP | Parsons Brinckerhoff has appointed Chip port the company’s water-related services. As Smith as area manager of the firm’s water prac- an environmental engineer with more than tice in the Carolinas. In his new position, Smith, 36 years of experience, he has extensive who will be based in the firm’s Charlotte, knowledge of public health policies and will N.C., office, will focus on the planning, be responsible for managing water-related design, and construction of water, wastewater, projects in Colorado. Before joining and stormwater infrastructure improvement Continued on page 7812 PEOPLE IN THE NEWS | 109:2  •  FEBRUARY 2017  |  JOURNAL AWWA

TO CREATEWE’VE JOINED FORCES SMARTER WATER SOLUTIONS.Xylem and Sensus are now one.As one company, we can now help you navigate the full breadth of challenges across the water cyclewith industry-leading solutions today and a world of possibilities tomorrow. We’re embarking on anexciting journey of innovation in advanced sensing, measurement, communications and analytics tohelp you optimize your operations, with greater efficiency and reliability and give you more controlover your network than ever before.Starting today, we’re one company, with a strong flow of innovation and ideas for smarter solutions in water,electric, gas and lighting networks. Innovations that cross traditional utility boundaries, building on each otherto deliver on the promise of the Internet of Things for utilities, municipalities and smart cities.Join us as we chart a new and exciting course ahead.xylem.com

Journal- American Water Works Association PEER-REVIEWED ARTICLESThe following section contains this issue’s peer-reviewed, original research content. Each month weprint the full version of at least one article, along with the expanded summaries of additional peer-reviewed articles that appear in their entirety on the Journal AWWA website (www.awwa.org/journal).All peer-reviewed articles from 1990 to present are available free of charge for anyone to view online.There are several advantages to publishing your research in Journal AWWA:WATER INDUSTRY LEADER IN PUBLIC ACCESS LONGER REVIEW ARTICLESJournal AWWA is the rst journal in the water Journal AWWA will consider articles that exceedindustry to provide public access to its standard limits for text length andvital research and guidance from number of graphical elements toits peer-reviewed articles at no suf ciently present comprehensivecharge to readers or authors. reviews of subject areas.IN PRESS ARTICLES COOPERATION WITH AWWA CONFERENCESTo hasten the dissemination of peer-reviewed AND EVENTSinformation, Journal AWWA posts unedited If you’ve made a presentation at an AWWAmanuscripts online soon after conference and would like tothey have been accepted for publish your ndings in Journalpublication. In Press articles can AWWA, there are no copyrightbe found at www.awwa.org/journal barriers to doing so. Material thaton the In Press Articles page. has been presented at an AWWA conference may be reused as partWATER EXPRESS of AWWA publications.Articles of immediate interest to Journal readersare put through the Water Express process. Journal AWWA submission guidelines can beReviewers are preselected and accessed online at www.awwa.org/submit.review times are shortened to Questions regarding manuscript submissionsexpedite the time to acceptance, can be directed to the editor-in-chief atthereby reducing the time to [email protected]. Join the A.P. Black Award Winners in the Pages of Journal AWWAPrévost LeChevallier Clancy Summers Hrudey Schock Trussell14 109:2 • FEBRUARY 2017 | JOURNAL AWWA

Peer Reviewed Research FocusPreozonation Effects on Organic Foulants in aCoagulation–Ultrafiltration Membrane ProcessPAUL G. BISCARDI1 AND STEVEN J. DURANCEAU11Department of Civil, Environmental, and Construction Engineering, University of Central Florida, Orlando, Fla.The effect of integrating ozone ahead of coagulation, filtrate true color by 40%, ultraviolet absorbance (UVA)flocculation, and sedimentation (CFS) as pretreatment at 254 nm by 10%, and specific UVA by 30%, relativeto ultrafiltration (UF) membranes was investigated at to the control, indicating that while ozone had impairedthe bench scale for treatment of a surface water turbidity removal during CFS pretreatment, it hadcontaining organic foulants. Ozone was applied before improved removal of aromatic-rich organics.a CFS–UF process and compared with a CFS–UF Fluorescent excitation–emission matrixes confirmedcondition without ozone as the control. While CFS that humic acid-like and fulvic acid-like substancesalone reduced turbidity by 27%, CFS increased known to cause irreversible fouling were retained onturbidity by 61% while applying ozone. When the control membrane but were absent on the membraneintegrated with CFS and UF, however, ozone reduced when ozone was integrated with CFS pretreatment.Keywords: coagulation, fluorescence, preozonation, ultrafiltration Ultrafiltration (UF) is a low-pressure membrane filtra- research on pretreatment processes such as coagulationtion process primarily used to remove turbidity and sus- (Kimura et al. 2014), biofiltration (Netcher & Duranceaupended solids during the production of safe drinking 2016), and preoxidation with ozone (Van Geluwe et al.water. As reported by Furukawa (2008), research, coupled 2011) continue to be of vital importance.with advancements in membrane technology and increas-ingly stringent regulations, has led to further adoption of Few studies have attempted to investigate how vari-low-pressure membranes for treatment of surface water. ous pretreatment processes can be integrated with eachIn many cases, adoption of membrane technology has other to minimize membrane fouling. In a review of UFcome in the form of retrofits to existing conventional fouling control, Gao and colleagues (2011) identifiedsurface water treatment plants, in which traditional gran- only a single study that included integrated pretreat-ular media filters are replaced with UF membranes. ment, namely an ozone–adsorption–coagulation pre- treatment system investigated by Mozia et al. (2006).FOULING AND ITS MITIGATION In their research, Mozia and colleagues studied the Literature review. A major challenge associated with integration of powdered activated carbon adsorption following ozonation but found that this integratedmembrane filtration is fouling. Membrane fouling pretreatment approach resulted in increased membraneoccurs as materials either deposit and form a cake layer fouling. Alternatively, the integration of ozone andor adsorb directly to the surface of a membrane (Jermann coagulation pretreatment before UF has not been eval-et al. 2007, Zularisam et al. 2006). Organic fouling of uated extensively in the literature. Preoxidation withmembranes results in a loss of permeability and an ozone (preozonation) has been shown to reduce mem-increase in the energy required to filter water (Jacangelo brane fouling, and independently, to act as a coagulantet al. 1989). Effective operation of UF membranes for aid during conventional treatment under certain condi-surface water treatment often requires management of tions (Sam et al. 2010, Bose & Reckhow 2007). How-membrane fouling and optimization of conventional ever, there is a gap in knowledge regarding how preozo-pretreatment processes for fouling minimization. In nation can be utilized for removal of organic foulantsparticular, organic fouling is often mitigated through a when used in conjunction with conventional surfacevariety of pretreatment processes. As a result, extensive water treatment processes. BISCARDI & DURANCEAU  |  109:2 • FEBRUARY 2017  |  JOURNAL AWWA 15

Ozone historically has been applied in water treat- investigated preozonation to reduce fouling of a poly-ment for a variety of purposes (Camel & Bermond vinylidene difluoride membrane. That work demon-1998). As a strong oxidant, ozone can assist in the strated that preozonation could improve flux by 10%treatment of iron, manganese, color, taste, and odor. but could not eliminate fouling entirely. Schlichter andOzone can also be used as a disinfectant for inactiva- colleagues (2003) suggested that the increase in fluxtion of chlorine-resistant protozoa, bacteria, and from ozone was attributable to the effect of ozone onviruses (von Gunten 2003). Additionally, chlorinated the organic foulants. In their work, model solutionsdisinfection by-product formation potential can be with humic acid were treated at the bench scale. Whenreduced by substituting chlorine with ozone for disin- applying preozonation, they observed a reduction infection (Farahbakhsh et al. 2004, Camel & Bermond fouling along with an increase in flux (Schlichter et al.1998). Several researchers have also demonstrated that 2003). Lee and colleagues (2004) studied the mecha-ozone can also be used as a coagulant aid (Bose & nism by which ozone was reducing fouling. Namely,Reckhow 2007, Schneider & Tobiason 2000, Camel & they observed that humic substances were being brokenBermond 1998, Jekel 1998). The results of these stud- down to lower-molecular-weight organic compounds,ies demonstrated that preozonation can either enhance which were passing through the membrane and lead-or have an adverse effect on the coagulation process, ing to elevated organic matter in the filtered water.depending on various conditions that include pH, alka- This effect had previously been observed in severallinity, hardness, and natural organic matter (NOM) other studies that used media filtration (Schneider &content. However, the fundamental mechanisms dictat- Tobiason 2000, Becker & O’Melia 1996, Edwards &ing the optimal conditions for the use of ozone as a Benjamin 1992, Jekel 1986).coagulant aid are not entirely clear. Also of note, thevast majority of these studies were conducted under Objective of current research. Although recent work hasthe assumption that coagulation would be followed by investigated the use of ozone to directly reduce organictraditional media filtration or were focused primarily fouling of membranes, and other researchers, indepen-on NOM removal by coagulation and did not mention dently, have considered the use of preozonation as athe filtration technique. coagulant aid, very few studies have evaluated the inte- gration of ozone, coagulation, and membrane filtration. Preozonation has also been proposed as a pretreat- Specifically, there is a lack of knowledge regarding thement to directly reduce organic fouling of membranes downstream impact of ozone–coagulation treatment on(Fujioka & Nghiem 2015, Barry et al. 2014, Moslemi organic matter known to cause irreversible membraneet al. 2014, Szymanska et al. 2014, Orta de Velásquez et fouling. As conventional water treatment plants continueal. 2013, van Geluwe et al. 2011). It is well known that to replace media filtration systems with membranes, theozone will break down organic compounds, in particu- need for such information regarding fouling control withlar aromatic-rich and humic-like NOM. This is signifi- ozone has become increasingly important.cant because aromatic-rich and humic-like NOM havebeen linked to irreversible fouling of UF membranes The aim of the current bench-scale research was to(Peiris et al. 2010, Jones & O’Melia 2001, Jucker & take the first steps in evaluating the integration ofClark 1994). Jucker and Clark (1994) showed that preozonation with coagulation and UF. Specifically,humic acids were prone to adsorb to UF membranes to the quantity and characteristics of organic foulantsa greater degree than were fulvic acids. Jones and were tracked in a bench-scale evaluation that com-O’Melia (2001) further studied the concept of humic- pared integrated conventional pretreatment both withacid adsorption onto UF membranes and observed a and without preozonation. Organic characterizationdirect effect on flux decline. More recently, Peiris and was achieved through the application of fluorescencecolleagues (2010) used fluorescence spectroscopy to spectroscopy, ultraviolet absorbance (UVA), and size-identify the major foulant components in NOM. Their exclusion chromatography in order to track the aromatic-work specifically identified aromatic proteins and rich components of NOM known to cause chemicallyhumic-like fluorescing substances as major organic fou- irreversible fouling of polymeric membranes.lants in surface water. Given that ozone is known totransform these compounds, it is not surprising that MATERIALS AND METHODSozone has shown potential for reducing organic fouling. Source water. Raw surface water was collected from theHowever, most studies have only considered using ozonedirectly ahead of the membrane with no intermediate Lake Manatee Water Treatment Plant in Manatee County,processes. Also, previous work has been largely limited Fla., and tested within 48 h. The water was stored in ato fouling of ceramic membranes because ozone-resistant cooler set to 4°C before testing. The source water con-polymeric membranes were not widely available before tained approximately 20 mg/L dissolved organic carbon2004 (Lee et al. 2004). In one of the few published (DOC). This organic-rich surface water was chosenstudies with an ozone-tolerant membrane, Park (2002) because it is currently treated through conventional sur- face water treatment with media filtration but is slated to transition to UF membranes (Sethi et al. 2015). Lake16 BISCARDI & DURANCEAU  |  109:2 • FEBRUARY 2017  |  JOURNAL AWWA

Manatee raw surface water was subsequently treated for 60 s was conducted after approximately 45 min ofunder two scenarios: one with preozonation and a control filtration. Samples were collected in duplicate through-study without preozonation. out each filtration cycle. Preozonation. For the evaluation with preozonation, a Analytical methods. Each sample collected during thetransferred ozone dose of approximately 14 mg/L was bench-scale experiments was tested for water qualityachieved using a bench-scale ozone generator.1 The parameters including pH, turbidity, temperature, dissolvedtransferred ozone dose was chosen to provide approxi- metals, and conductivity in accordance with standardmately 0.7 mg O3/mg DOC, which had been recom- methods (Standard Methods 2005). Additional parametersmended for optimal coagulation performance in previ- that relate to organic characterization were measuredous research by Schneider and Tobiason (2000). Raw including DOC, ultraviolet-visible (UV-vis) spectra, truesurface water was dosed in batches of 20 L at a time. color, and fluorescence excitation–emission spectra.Ozone residual was not detectable after dosing the sur-face water. Figure 1 shows an application of ozone to DOC was determined by first filtering samples with aone of the batches of surface water. The target trans- prewashed 0.45-µm membrane filter followed byferred dose of 14 mg/L was achieved after approxi- analysis using a total organic carbon analyzer3 accordingmately 800 s. The mass transfer efficiency (MTE) is to method 5310C (Standard Methods 2005). UV-visdefined as the transferred ozone dose divided by the spectra were collected using a spectrophotometer.4 Sam-applied ozone dose. The instantaneous MTE had leveled ples were first filtered through a prewashed 0.45-µmoff to approximately 35% after 7 min. membrane filter before undergoing UV-vis scans. Each scan was conducted from wavelength 200 to 600 nm in Bench-scale jar testing. Bench-scale jar testing was 1-nm intervals.used to simulate coagulation–flocculation–sedimentation(CFS). A jar testing apparatus2 was used to conduct the Fluorescence excitation–emission spectroscopy wastest. Six jars were filled with 2 L of either surface water conducted to further characterize the dissolved organicor ozonated surface water. Each jar was dosed with matter. Before fluorescence analysis, samples were filteredapproximately 100 mg/L of polyaluminum chloride with a 0.45-µm membrane filter to remove particulates.(PACl) coagulant. The dose and jar testing sequence was Without further pretreatment, fluorescence excitation–chosen to match the conditions of a surface water treat- emission matrix (EEM) spectra were collected using ament plant that uses a solid-contact clarifier. This spectrofluorophotometer.5 The emission intensity read-sequence consisted of 11 s at 300 rpm to simulate a ings were captured in 1-nm wavelength intervals betweenrapid mix, 4 min and 14 s at 100 rpm to simulate mix- 280 and 600 nm for excitation wavelengths ranging froming at the inlet works, 8 min at 60 rpm to simulate the 200 to 400 nm in 5-nm intervals. The excitation andclarifier mixing zone, 7 min at 5 rpm to simulate the emission slits were set to a 10-nm band pass.clarifier flocculation zone, and 10 min at 0 rpm tosimulate settling. Water quality samples of the superna- The Rayleigh scattering effect was minimized bytant from each jar were collected independently, which subtracting the fluorescence spectra collected from asimulated a 0.6 m3/m2/h overflow rate. The jar testing blank sample of deionized water. Given that theand water quality testing were conducted in duplicate. organic content of the surface water was thought toThe supernatant from the jar test was then transferredto the feed tank of the bench-scale membrane apparatus. FIGURE 1 Transferred ozone dose curve Bench-scale UF testing. Bench-scale hollow-fiber UF Cumulative Ozone Dose—mg/L16Cumulative ozone dose 100membranes composed of a blend of polyethersulfone Instantaneous Mass Transfer Efficiency—%14Instantaneous mass transfer efficiency80and polyvinylpyrrolidone were used in the experiments. 12 60The membrane element was designed to be operated 10 100 200 300 400 500 600 700 40with an inside-out, dead-end flow path. Each module Time—s 20contained a total of 120 fibers, which made up a com- 8 0bined total active area of 0.08 m2. Each fiber had a 60.8-mm diameter and was 300 mm in length. The nom- 4inal pore size of the membrane was 0.010 µm (absolute 2pore size of 0.025 µm), and the molecular weight cutoff 0was 200,000 D. The filtration experiments were carriedout by pumping feedwater to the bench-scale UF module 0using a peristaltic pump. A permeate flux of 85 L/m2/hwas maintained. Flow was monitored using a digitalflowmeter connected to a data logger. Manual adjust-ments to the peristaltic pump were conducted to main-tain constant flux. Hydraulic backwashing at 255 L/m2/hBISCARDI & DURANCEAU  |  109:2 • FEBRUARY 2017  |  JOURNAL AWWA 17

TABLE 1 Description of FRI regions potentially contain more than 20 mg/L for some sam- ples, it was also important to account for the absorp- EEM Excitation Emission Description tion of light by the DOC of the sample (commonly Region Range Range referred to as the inner filter effect). A correction for nm nm the inner filter effect was applied to the blank-subtracted spectra following the method described by WesterhoffRegion I 200–250 280–330 Aromatic protein-like and colleagues (2001).Region II 200–250 330–380 Aromatic protein-likeRegion III 200–250 380–600 Fulvic acid-like Fluorescence regional integration (FRI) was used toRegion IV 250–340 280–380 Soluble microbial quantify and interpret the results of each EEM sample  by-product-like taken (Chen et al. 2003). FRI involves dividing an EEMRegion V 250–400 380–600 Humic acid-like into characteristic regions that represent different types of organic matter as shown in Table 1. A normalized,EEM—excitation–emission matrix, FRI—fluorescence regional integration integrated volume (i,n) was determined for the peaks in each region for quantitative comparison. The five regionsFIGURE 2 EEM as divided into characteristic regions are shown in Figure 2. by uorescence regional integration The apparent molecular weight distribution of the 400 NOM was determined using high-performance size- exclusion chromatography (HPSEC). Established HPSEC 350 methodology was followed and is detailed elsewhere (Zhou et al. 2000, Chin et al. 1994). Briefly, a high-performanceExcitation—nm 300 Region V liquid chromatography (HPLC) system6 consisting of a pump and autosampler was used. The mobile phase was Region IV a phosphate buffer that consisted of 2 mM dipotassium phosphate, 2 mM monopotassium phosphate, and 0.1 M 250 sodium chloride. The mobile phase was pumped at a 1-mL/min flow rate. The sample injection volume was Region Region Region III 150 µL. A size-exclusion column7 was used. Calibration I II was achieved with molecular weight standards prepared from HPLC-grade acetone, salicylic acid, and sodium 200 polystyrene sulfonate standards8 with molecular weights of 1.6, 5.2, 7.4, 16, and 34 kD (kilodalton). Before HPSEC 300 350 400 450 500 550 600 analysis, samples were filtered with a 0.45-µm membrane Emission—nm filter and adjusted to an ionic strength similar to the mobile phase using sodium chloride.Adapted from Chen et al. (2003)EEM—excitation–emission matrixTABLE 2 Comparison of water quality from both the preozonation experiment and the experiment without preozonation Without Preozonation With Preozonation Raw → Coagulation Raw → Coagulation → Raw → Raw → Ozone → Raw → Ozone → UF Ozone Coagulation Coagulation → UF Parameter Raw (Post-CFS) (UF Filtrate) (Post-CFS) 6.8 (UF Filtrate)pH 6.8 6.3 6.4 12.1 6.2DO—mg/L O2 8.0 8.3 8.4 1.8 10.3 6.2Turbidity—ntu 2.2 1.6 0.1 192 2.9 10.4Conductivity—µS/cm 186 209 209 79 208 0.1True color—PCU 180 15 15 18 209DOC—mg/L 19 7 7 0.56 8UV254—cm–1 0.85 0.19 0.19 3.1 9 9SUVA—L/mg/m 4.5 2.7 2.7 0.14 0.19 9Fe—mg/L 0.14 <0.10 <0.10 <0.10 1.9 0.17Mn—mg/L <0.10 <0.10 <0.10 <0.10 1.9 <0.10 <0.10 <0.10CFS—coagulation–flocculation–sedimentation, DO—dissolved oxygen, DOC—dissolved organic carbon, Fe—iron, Mn—manganese, O2—oxygen,PCU—platinum–cobalt color unit, SUVA—specific ultraviolet absorbance, UF—ultrafiltration, UV254—ultraviolet absorbance at 254 nm18 BISCARDI & DURANCEAU  |  109:2 • FEBRUARY 2017  |  JOURNAL AWWA

RESULTS AND DISCUSSION FIGURE 3 UV-vis absorbance scans for raw and Effect of preozonation on raw surface water. Table 2 ozonated surface watershows source water quality parameters and the effect of Raw waterozone on these parameters. Preozonation at 0.70 mg 2.0 Ozonated waterO3/mg DOC initially appeared to remove approximately5% of the DOC. The lack of significant DOC removal 1.8by ozonation was expected. Additionally, the pH andconductivity were not significantly affected. However, 1.6true color was reduced by approximately 56%. Thearomaticity of the remaining DOC was reduced as indi- Absorbance—cm–1 1.4cated by specific UV absorbance (SUVA), which droppedfrom 4.5 to 3.1 L/mg/m. UV-vis wavelength scans are 1.2presented in Figure 3. These scans indicated that ozonehad the most significant impact on UV absorbance at 1.0wavelengths between 200 and 250 nm. As expected,these results suggested that ozone preferentially 0.8destroyed or broke down the aromatic-rich, UV-absorbentfraction of the dissolved organic matter found in the raw 0.6Lake Manatee surface water. 0.4 Figure 4 shows the EEM data both before and afterozonation. The various peaks shown in part A of the 0.2figure suggested that the source water contained aromaticprotein-like, humic acid-like, and fulvic acid-like sub- 0.0 300 400 500 600stances. Figure 4, part B, indicated that, as expected, the 200 Wavelength—nmfluorescence of the organic substances had been dimin-ished following ozonation. UV-vis—ultraviolet-visible This difference was quantified using the EEM spectra predominant than regions I, II, and IV in the ozonatedFRI analysis (shown in Figure 5); analysis indicated that samples than in the raw surface water, although at sig-the greatest magnitude of reduction in integrated fluo- nificantly lower magnitudes.rescence intensity was from the fulvic acid-like region(region III). However, while region I (aromatic proteins) These results suggest that ozone preferentially trans-exhibited the least fluorescence, it experienced the great- formed the molecular structure of the organic substancesest percentage reduction (approximately 80%). Like- such that their fluorescence was reduced relative to thewise, region II (aromatic proteins) also experienced a raw surface water. Given that the DOC concentration57% reduction in fluorescence. As a result, regions III appeared to be reduced only by 5% in the ozonatedand V (humic-like substances) became relatively more water, the main effect of preozonation was the change in the characteristics of the organic matter fed to the coagulation step.FIGURE 4 EEM data for raw (A) and ozonated (B) surface water A Raw water Fluorescence B Ozonated water Fluorescence400 intensity—AU 400 intensity—AU350 1.0 350 1.0 0.8 0.8Excitation—nm Excitation—nm300 0.6 300 0.6 0.4 0.4250 250 0.2 0.2 200 0.0 200 0.0 300 350 400 450 500 550 600 300 350 400 450 500 550 600 Emission—nm Emission—nmAU—arbitrary units, EEM—excitation–emission matrix BISCARDI & DURANCEAU  |  109:2 • FEBRUARY 2017  |  JOURNAL AWWA 19

FIGURE 5 Results of the FRI analysis for raw surface water and preozonated surface water 16,000 Raw surface water Ozonated surface water 14,000 13,693FRI Volume, Фi,n—AU-nm2 12,000 7,858 9,658 10,000 5,704 5,324 8,000 6,000 4,000 1,755 2,284 3,124 2,000 343 1,445 0 Region I Region II Region III Region IV Region V (fulvic acid-like) (microbial by-product-like) (humic acid-like) (aromatic protein-like) (aromatic protein-like) EEM RegionAU—arbitrary units, EEM—excitation–emission matrix, FRI—fluorescence regional integration The average apparent molecular weight of the UV- revealed that the remaining UV-absorbent organics in theabsorbent organic matter was reduced as a result of preozonated water had a broader distribution of molecu-preozonation. The HPSEC chromatograph (Figure 6) lar weights compared with the unozonated raw water (indicated by the broader shape of the preozonated sur-FIGURE 6 HPSEC apparent molecular weight face water peak). Analysis of the HPSEC data showed distribution for raw water and ozonated that the weight-averaged apparent molecular weight of water the raw surface water was 952 D, compared with 693 D for the ozonated surface water (a 27% reduction). This Raw surface water result suggests that ozone had preferentially transformed Ozonated surface water aromatic compounds with molecular weights greater than 100 D and had less effect on low-molecular-weight UV- Retention Time—min absorbing organic matter. 2 4 6 8 10 12 14 16 18 20 22 24 26 28 Effect of preozonation on post-CFS water quality.UV254 Detector Response—AU 0.035 Although the direct impact of ozone on the surface water was not surprising, the results of the bench-scale 0.030 study revealed significant differences in the CFS process performance when preozonation was implemented. In 0.025 the experiment with preozonation, the CFS process improved with respect to color removal. The post-CFS 0.020 true color was 8 PCU (platinum–cobalt color unit) com- pared with 15 PCU when treated without preozonation. 106 105 104 103 102 101 100 10–1 10–2 10–3 However, when assessed as a percent removal, the CFS Apparent Molecular Weight—D performance for color removal was unchanged at approximately 90% true color removal in both cases. AAU—arbitrary units, HPSEC—high-performance size-exclusion similar trend applied to reduction of UVA at 254 nmchromatography, UV254—ultraviolet absorbance at 254 nm (UV254). Applying preozonation led to a lower post-CFS UV254; however, the percent removal was only slightly reduced. CFS reduced UV254 by approximately 80% relative to the raw surface water and achieved approxi- mately 70% UV254 reduction in the ozonated surface water. Fluorescent EEMs indicated that the post-CFS20 BISCARDI & DURANCEAU  |  109:2 • FEBRUARY 2017  |  JOURNAL AWWA

water had reduced fluorescence when preozonation was were reduced in the filtrate water produced during theapplied (Figure 7). However, turbidity and DOC were preozonation study. These results suggested that theboth found to be at higher levels in the post-CFS samples reduction of aromatic-rich and fluorescent compoundswhen preozonation was applied, indicating that ozone from preozonation was consistent in the post-CFS water;had impaired the performance of the CFS process for however, the reduction in turbidity and DOC attribut-removal of turbidity and overall organics. During treat- able to preozonation did not carry forward beyond thement of the preozonated water, smaller floc that settled preozonation step because of the adverse effect on themore slowly were visually observed, which was an addi- CFS process from the integration of preozonation.tional indicator of impaired coagulation performance.However, the application of preozonation led to reduced Effect of preozonation on membrane fouling. In order toaromaticity of the post-CFS water. The post-CFS sam- assess the changes in organic matter retention on theples from the preozonation experiment had a SUVA of membrane surface resulting from the integration ofapproximately 1.9 L/mg/m compared with 2.7 L/mg/m ozone and coagulation, differential EEMs were calcu-without preozonation. lated (Figure 9). The differential EEMs represent a sub- traction of the filtrate EEM from the post-CFS EEM and Effect of preozonation on filtrate water quality. These therefore are representative of the fluorescent com-trends mostly carried forward into the filtrate samples. pounds that were retained by the UF membrane.DOC was found to be at a higher concentration for Whereas Figures 7 and 8 show that the quantity offiltrate samples from the preozonation experiment. fluorescing organic material was less throughout theHowever, SUVA, true color, and fluorescence (Figure 8) process with preozonation, Figure 9 shows that theFIGURE 7 EEM data for post-CFS water without preozonation (A) and with preozonation (B)Excitation—nm A Post-CFS (without preozonation) Excitation—nm B Post-CFS (with preozonation) Fluorescence 400 Fluorescence 400 intensity—AU intensity—AU 350 1.0 1.0 0.8 350 0.8 300 0.6 300 0.6 0.4 0.4 250 250 0.2 0.2 200 0.0 200 0.0 300 350 400 450 500 550 600 300 350 400 450 500 550 600 Emission—nm Emission—nmAU—arbitrary units, CFS—coagulation–flocculation–sedimentation, EEM—excitation–emission matrixFIGURE 8 EEM data for UF ltrate water without preozonation (A) and with preozonation (B)Excitation—nm A Filtrate (without preozonation) Fluorescence Excitation—nm B Filtrate (with preozonation) Fluorescence 400 intensity—AU 400 intensity—AU 350 1.0 350 1.0 0.8 0.8 300 0.6 300 0.6 0.4 0.4 250 250 0.2 0.2 200 0.0 200 0.0 300 350 400 450 500 550 600 300 350 400 450 500 550 600 Emission—nm Emission—nmAU—arbitrary units, EEM—excitation–emission matrix, UF—ultrafiltration BISCARDI & DURANCEAU  |  109:2 • FEBRUARY 2017  |  JOURNAL AWWA 21

FIGURE 9 Differential EEMs (post-CFS EEM minus ltrate EEM) for experiment without preozonation (A) and with preozonation (B) A Differential EEM (without preozonation) B Differential EEM (with preozonation) 400 Fluorescence 400 Fluorescence intensity—AU intensity—AU 0.05 0.05 350 350 0.04 0.04Excitation—nm Excitation—nm3000.03300 0.03 0.02 0.02 250 250 0.01 0.01 200 0.00 200 0.00 300 350 400 450 500 550 600 300 350 400 450 500 550 600 Emission—nm Emission—nmAU—arbitrary units, CFS—coagulation–flocculation–sedimentation, EEM—excitation–emission matrixfluorescing substances were no longer being retained by Likewise, the differential UV-vis absorbance scansthe membrane. In other words, the process that incorpo- presented in Figure 10 were calculated by subtractingrated preozonation included DOC that was less fluorescent the filtrate UV-vis absorbance scan data from the post-and less aromatic and contained fluorescing compounds coagulation absorbance scan data. These plots illustratethat were less retained by the membrane. However, there the magnitude of the additional organics retained by thewas approximately 28% more DOC in the filtrate water membrane during the control experiment without pre-when using preozonation as compared with the control. ozonation. Although the DOC analysis was not preciseAlthough, this increase in filtrate DOC was apparently enough to quantify the removal of DOC by the UFattributable to the impaired CFS process and not because membrane, the differential EEMs (Figure 9) and the dif-additional organics were passing through the membrane. ferential absorbance scans (Figure 10) were sensitive enough to detect a change in organics retained on theFIGURE 10 Differential absorbance scans UF membrane. Given that aromatic fluorescent com- (post-CFS UVA minus ltrate UVA) for pounds are known to cause chemically irreversible foul- the experiment without preozonation ing, preozonation integrated with coagulation and UF and with preozonation may yield a more chemically reversible type of fouling. These results seem to indicate that although the DOCPost-CFS–Filtrate UVA—cm–1 0.11 Without preozonation was elevated in the water treated with preozonation 0.10 With preozonation followed by coagulation, the remaining fluorescent 0.09 600 organic matter known to cause irreversible fouling— 0.08 300 400 500 such as humic substances—were not retained by the 0.07 Wavelength—nm membrane. These EEM results suggest that foulants that 0.06 were less chemically irreversible were in the feedwater 0.05 during the experiment with preozonation. These results 0.04 agreed with recent pilot testing data that showed that 0.03 ozone was specifically effective at reducing chemically 0.02 irreversible fouling as opposed to hydraulically irrevers- 0.01 ible fouling (Biscardi et al. 2016). Future research should 0.00 expand this work to additional source waters to better –0.01 understand the relationship between ozone, natural organic matter, and membrane fouling. 200 CONCLUSIONSCFS—coagulation–flocculation–sedimentation, UVA—ultraviolet The goal of this work was to investigate the impact ofabsorbance integrated preozonation and coagulation on organic foulants. This study yielded the following major findings:22 BISCARDI & DURANCEAU  |  109:2 • FEBRUARY 2017  |  JOURNAL AWWA

•• Preozonation at 0.70 mg O3/mg DOC initially methods to reduce membrane fouling during drinking removed only 5% of the DOC; however, a significant water treatment. Steven J. Duranceau (to whom fraction of the humic acid-like, fulvic acid-like, and correspondence may be addressed) is an associate protein-like substances known to cause irreversible professor in the Department of Civil, Environmental, and membrane fouling were transformed to yield non- Construction Engineering, University of Central Florida, fluorescing molecules. 4000 Central Florida Blvd., POB 162450, Orlando, FL 32816-2450 USA; [email protected]. •• Preozonation at 0.70 mg O3/mg DOC, integrated with PACl coagulation, enhanced the overall removal https://dx.doi.org/10.5942/jawwa.2017.109.0013 of turbidity, color, UV254-absorbing constituents, and fluorescent constituents in surface water. However, PEER REVIEW DOC in the UF feed was increased by 28% because Date of submission: 04/16/2016 of impaired removal of DOC by CFS. Date of acceptance: 10/11/2016 •• Differential EEMs and absorbance scans con- REFERENCES firmed that the remaining aromatic fluorescent fraction of the organic matter was no longer Barry, M.C.; Hristovski, K.; & Westerhoff, P., 2014. Membrane Fouling by retained on the membrane when preozonation was Vesicles and Prevention Through Ozonation. Environmental integrated with CFS. Science & Technology, 48:13:7349. •• Future research should investigate the changes in the Biscardi, P.G. & Duranceau, S.J., 2016. Ultrafiltration Fouling long-term, chemically irreversible fouling rate of Reduction With the Pilot-Scale Application of Ozone Preceding surface water pretreated with ozone and coagulation Coagulation, Flocculation, and Sedimentation for Surface Water to further assess this integrated treatment configura- Treatment. Desalination and Water Treatment, 57:27433. tion with additional source waters. http://dx.doi.org/10.1080/19443994.2016.1180266.ACKNOWLEDGMENT Becker, W. & O’Melia, C.R., 1996. Optimizing Ozonation for Turbidity and Funding for this project was provided by Harn R/O Organics (TOC) Removal by Coagulation and Filtration. AWWA, Denver.Systems (Venice, Fla.) and the Alameda County Water Bose, P. & Reckhow, D.A., 2007. The Effect of Ozonation on NaturalDistrict (Alameda County, Calif.) under University of Organic Matter Removal by Alum Coagulation. Water Research,Central Florida (UCF) project agreement 16208088. The 41:7:1516.authors are grateful for the contributions of UCF stu-dents Maria Arenas, Martin Coleman, Cassidy Conover, Camel, V. & Bermond, A., 1998. The Use of Ozone and AssociatedAri Hadar, Carlyn Higgins, Hadi Toure, Ben Yoakum, Oxidation Processes in Drinking Water Treatment. Waterand David Yonge. The authors also gratefully acknowl- Research, 32:11:3208.edge the laboratory support of Maria Real-Robert. Thiswork would not have been possible without the source Chen, W.; Westerhoff, P.; Leenheer, J.A.; & Booksh, K., 2003.water provided by Bruce MacLeod and Katherine Fluorescence Excitation−Emission Matrix Regional Integration toGilmore of the Lake Manatee Water Treatment Plant, Quantify Spectra for Dissolved Organic Matter. EnvironmentalManatee County, Fla. Science & Technology, 37:24:5701.ENDNOTES Chin, Y.-P.; Aiken, G.; & O’Loughlin, E., 1994. Molecular Weight, Polydispersity, and Spectroscopic Properties of Aquatic Humic 1Guardian Manufacturing, Eustis, Fla. Substances. Environmental Science & Technology, 28:11:1853. 2Phipps & Bird, Richmond, Va. 3Teledyne Tekmar, Mason, Ohio Edwards, M. & Benjamin, M., 1992. Transformation of NOM by Ozone 4DR5000, Hach, Loveland, Colo. and Its Effect on Iron and Aluminum Solubility. Journal AWWA, 5RF-6000, Shimadzu, Kyoto, Japan 84:6:56. 6Series 200 HPLC, PerkinElmer, Waltham, Mass. 7Protein-Pak 125, Waters Corp., Milford, Mass. Farahbakhsh, K.; Svrcek, C.; Guest, R.K.; & Smith, D.W., 2004. A Review 8Scientific Polymer Products, Ontario, N.Y. of the Impact of Chemical Pretreatment on Low-Pressure Water Treatment Membranes. Journal of Environmental Engineering andABOUT THE AUTHORS Science, 3:4:237. Paul G. Biscardi is an assistant engineer with Hazen and Sawyer in Fujioka, T. & Nghiem, L.D., 2015. Fouling Control of a Ceramic Tampa, Fla. At the time that this Microfiltration Membrane for Direct Sewer Mining by article was written, he was a graduate Backwashing With Ozonated Water. Separation and Purification research assistant at the University of Technology, 142:268. https:/doi.org/10.1016/j.seppur.2014.12.049. Central Florida in Orlando, from which he received his BS, MS, and Furukawa, D., 2008. A Global Perspective of Low Pressure Membranes. PhD degrees in environmental NWRI-2008-03, Final Project Report. National Water Research Institute, Fountain Valley, Calif.engineering. For his dissertation research, he studied Gao, W.; Liang, H.; Ma, J.; Han, M.; Chen, Z.-L.; Han, Z.-S.; & Li, G.-B., 2011. Membrane Fouling Control in Ultrafiltration Technology for Drinking Water Production: A Review. Desalination, 272:1–3:1. Jacangelo, J.G.; Aieta, E.M.; Carns, K.E.; Cummings, E.W.; & Mallevialle, J., 1989. Assessing Hollow-Fiber Ultrafiltration for Participate Removal. Journal AWWA, 81:11:68.BISCARDI & DURANCEAU  |  109:2 • FEBRUARY 2017  |  JOURNAL AWWA 23

Jekel, M.R., 1998. Effects and Mechanisms Involved in Preoxidation and Peiris, R.H.; Hallé, C.; Budman, H.; Moresoli, C.; Peldszus, S.; Huck, P.M.; Particle Separation Processes. Water Science and Technology, & Legge, R.L., 2010. Identifying Fouling Events in a Membrane- 37:10:1. Based Drinking Water Treatment Process Using Principal Component Analysis of Fluorescence Excitation–EmissionJekel, M.R., 1986. Interactions of Humic Acids and Aluminum Salts in Matrices. Water Research, 44:1:185. the Flocculation Process. Water Research, 20:12:1535. Sam, S.; Yukselen, M.A.; Zorba, M.; & Gregory, J., 2010. The Effect ofJermann, D.; Pronk, W.; Meylan, S.; & Boller, M. 2007. Interplay of Ozone on the Reversibility of Floc Breakage: Suspensions With Different NOM Fouling Mechanisms During Ultrafiltration for High Humic Acid Content. Ozone: Science & Engineering, 32:6:435. Drinking Water Production. Water Research, 41:8. http://dx.doi. org/10.1016/j.watres.2006.12.030. Schlichter, B.; Mavrov, V.; & Chmiel, H., 2003. Study of a Hybrid Process Combining Ozonation and Membrane Filtration—Filtration ofJones, K.L. & O’Melia, C.R., 2001. Ultrafiltration of Protein and Humic Model Solutions. Desalination, 156:1–3:257. Substances: Effect of Solution Chemistry on Fouling and Flux Decline. Journal of Membrane Science, 193:2:163. Schneider, O.D. & Tobiason, J.E., 2000. Preozonation Effects on Coagulation. Journal AWWA, 92:10:74.Jucker, C. & Clark, M.M., 1994. Adsorption of Aquatic Humic Substances on Hydrophobic Ultrafiltration Membranes. Journal of Sethi, S.; Hugaboom, D.; Milton, D.; Nyfennegger, J.S.; Stone, E.; Simpson, Membrane Science, 97:37. http://dx.doi.org/10.1016/0376- M.; MacLeod, B.; & Gilmore, K., 2015. Planning a Plant Retrofit to 7388(94)00146-P. Result in Construction of Florida’s Largest Low-Pressure Membrane Facility. AWWA Annual Conference & Exposition, Anaheim, Calif.Kimura, K.; Tanaka, K.; & Watanabe, Y., 2014. Microfiltration of Different Surface Waters With/Without Coagulation: Clear Correlations Standard Methods, 2005 (21st ed.). Standard Methods for the Between Membrane Fouling and Hydrophilic Biopolymers. Water Examination of Water and Wastewater. APHA, AWWA, and WEF, Research, 49:434. https:/doi.org/10.1016/j.watres.2013.10.030. Washington.Lee, S.; Jang, N.; & Watanabe, Y., 2004. Effect of Residual Ozone on Szymanska, K.; Zouboulis, A.I.; & Zamboulis, D., 2014. Hybrid Ozonation– Membrane Fouling Reduction in Ozone Resisting Microfiltration Microfiltration System for the Treatment of Surface Water Using (MF) Membrane System. Water Science and Technology, 50:12:287. Ceramic Membrane. Journal of Membrane Science, 468:0:163.Moslemi, M.; Davies, S.H.; & Masten, S.J., 2014. Hybrid Ozonation– Van Geluwe, S.; Braeken, L.; & Van der Bruggen, B., 2011. Ozone Ultrafiltration: The Formation of Bromate in Waters Containing Oxidation for the Alleviation of Membrane Fouling by Natural Natural Organic Matter. Separation and Purification Technology, Organic Matter: A Review. Water Research, 45:12:3551. 125:0:202. von Gunten, U., 2003. Ozonation of Drinking Water: Part II. DisinfectionMozia, S.; Tomaszewska, M.; & Morawski, A.W., 2006. Application of an and By-product Formation in Presence of Bromide, Iodide or Ozonation–Adsorption–Ultrafiltration System for Surface Water Chlorine. Water Research, 37:7:1469. Treatment. Desalination, 190:1–3:308. Westerhoff, P.; Chen, W.; & Esparza, M., 2001. Fluorescence Analysis ofNetcher, A.C. & Duranceau, S.J., 2016. Modeling the Improvement of a Standard Fulvic Acid and Tertiary Treated Wastewater. Journal Ultrafiltration Membrane Mass Transfer When Using Biofiltration of Environmental Quality, 30:6:2037. Pretreatment in Surface Water Applications. Water Research, 90:258. http://dx.doi.org/10.1016/j.watres.2015.12.038. Zhou, Q.; Cabaniss, S.E.; & Maurice, P.A., 2000. Considerations in the Use of High-Pressure Size Exclusion Chromatography (HPSEC) forOrta de Velásquez, M.T.; Monje-Ramírez, I.; & Muñoz Paredes, J.F., 2013. Determining Molecular Weights of Aquatic Humic Substances. Effect of Ozone in UF-Membrane Flux and Dissolved Organic Water Research, 34:14:3505. Matter of Secondary Effluent. Ozone: Science & Engineering, 35:3:208. Zularisam, A.W.; Ismail, A.F.; & Salim, R., 2006. Behaviours of Natural Organic Matter in Membrane Filtration for Surface WaterPark, Y.G., 2002. Effect of Ozonation for Reducing Membrane-Fouling in Treatment—A Review. Desalination, 194:1:211. http://dx.doi. the UF Membrane. Desalination, 147:1–3:43. org/10.1016/j.desal.2005.10.030.24 BISCARDI & DURANCEAU  |  109:2 • FEBRUARY 2017  |  JOURNAL AWWA

Peer Reviewed Expanded SummaryEstimating the National Costs of RegulatingPerchlorate in Drinking WaterCAROLINE G. RUSSELL AND KEVIN M. MORLEYhttps://dx.doi.org/10.5942/jawwa.2017.109.0009 A prior assessment estimating the national compliance The perchlorate compliance strategy used for a givencosts associated with perchlorate by Russell et al. (2009) source was assigned on the basis of a survey of Californiawas updated in anticipation of a proposed National Pri- and Massachusetts PWSs required to implement strategiesmary Drinking Water Rule. The national cost estimate to comply with state MCLs. Treatment using single-passwas updated to account for utility compliance actions ion-exchange was assumed for 43% of perchlorate-affectedtaken in response to the maximum contaminant levels sources, blending was assumed for 43%, and source aban-(MCLs) for perchlorate in California (6 µg/L) and donment was assumed for the remaining 14%.Massachusetts (2 µg/L), which were promulgated sincethe 2009 study was completed. Including the California Capital and operation and maintenance (O&M) costor Massachusetts water systems that have already imple- curves were developed for a range of flows for eachmented strategies for compliance with state regulations potential compliance strategy (i.e., single-pass ion-would inappropriately skew the costs (and benefits) of a exchange, blending, or source abandonment) based onfederal regulation. Second, this study incorporated costs survey results, past studies, and engineering opinions offor source abandonment and/or blending, recognizing probable construction costs derived from actual projectthat some water systems may pursue alternative compli- costs for similarly sized systems. For blending, low-endance strategies in lieu of treatment. The 2009 study and high-end capital costs were developed to reflect sur-assumed that single-pass ion-exchange treatment would vey results for systems able to blend with existing sup-be implemented at all contaminated sources. plies, which incurred minimal capital costs, versus other systems where blending was not an option and therefore Perchlorate data included in the US Environmental incurred capital costs for a new well.Protection Agency’s final Unregulated ContaminantsMonitoring Rule 1 (UCMR 1) database were used to After assigning capital and O&M costs for each con-estimate the percent and number of affected public taminated source, the costs were aggregated to estimatewater systems (PWSs) for a given potential MCL. The the total national cost burden for water systems expectedmedian and 90th percentile perchlorate concentrations to comply with a given MCL. Amortized capital costs andwere calculated for each UCMR 1 entry point sample net present value O&M costs were calculated assuminglocation with at least one UCMR 1 sample above the 20 years of operation and for a 3% interest rate.detection limit. The updated cost assessment demonstrates that theFIGURE 1 Estimated compliance cost burdena potential impacts on individual systems are significant and are particularly burdensome for small water systemsSystem Size—population served Very small (Figure 1). Estimated costs for very small systems (popu- (<500) lation ≤ 500) are approximately $3/1,000 gal, represent- Small ing an increase in annual household water bills of more than $500 for a family of four. Some PWSs may avoid (501–3,300) incurring high treatment costs by abandoning impacted Medium sources. However, most of these systems will eventually incur costs to replace the abandoned supply. This study (3,301–10,000) demonstrates that opportunity costs for lost water (source Large abandonment) are comparable to treatment costs. (10,001–100,000) REFERENCE Very large (>100,000) Russell, C.G.; Roberson, J.A.; Chowdhury, Z.; & McGuire, M.J., 2009. National Cost Implications of a Perchlorate Regulation. Journal 0.00 1.00 2.00 3.00 4.00 AWWA, 101:3:54. Costs—$/1,000 gal Corresponding author: Caroline G. Russell is aMCL—maximum contaminant level, O&M—operations and principal technologist at Carollo Engineers, 8911maintenance Capital of Texas Hwy. North, Ste. 2200, Austin, TX 78759 USA; [email protected] capital and O&M costs to individual systems if federalperchlorate MCL is 4 µg/L RUSSELL & MORLEY  |  109:2 • FEBRUARY 2017  |  JOURNAL AWWA 25

Peer Reviewed Expanded SummaryData-Driven HR: Creating a Turnover ResponsePlan Based on the Kaplan-Meier EstimatorSTANLEY LAM AND AARON CHANhttps://dx.doi.org/10.5942/jawwa.2017.109.0011Employees—%The right data can help drive decisions regarding new The KM curve can be drawn for any organization usingstaff programs or budgetary impacts. While engineering data that most HR departments already possess: employeemanagers have numerous design programs and statistical rosters. Houston PUD’s curve (Figure 1) was drawn usingtools they can use to assist their decision-making, where only the hire and termination dates for employees who hadcan human resources (HR) professionals turn? left the utility over the past 10 years. Those points at which the curve drops off the steepest indicate when the most This article introduces the Kaplan-Meier estimator, a employees left the organization. The highlighted areas illus-statistical tool that can illustrate patterns of turnover trate two areas of particular concern for Houston’s utility:in an organization. Understanding these patterns is employees who have recently joined the organization andimportant because reducing employee turnover is a employees nearing their 20-year work anniversaries.universal challenge facing HR professionals at utilities.Best practices suggest calculating the utility’s turnover This knowledge helped the authors develop a planrate, which measures turnover across the entire organi- that targeted these specific segments of the workforcezation. However, the turnover rate does not provide with tailored turnover programs. For the first group,much actionable insight. HR professionals might also newer employees, the workforce plan included creatingwant to know who is leaving their organization or more-defined career paths and increasing learningwhether there are specific subsets of staff on which they opportunities. For the second group, those employeesshould focus their retention efforts. nearing their 20-year work anniversary, the workforce plan was intended to increase employee engagement by To shed light on these topics, utilities can use the providing cross-training and mentorship opportunities.Kaplan-Meier estimator, or KM curve. This statistical toolpaints a detailed picture of the patterns of turnover within Using the KM curve allowed HR professionals atan organization. It was initially developed in the field of Houston PUD to develop a data-driven response to turn-oncology and was first adapted for the utilities industry over by targeting specific areas within their organizations.by the authors in their work at the City of Houston Other utility managers can use the KM curve to identifyPublic Utilities Division (Houston PUD) in Texas. the portions of their workforce that have contributed to turnover in recent years (i.e., those who may also be at FIGURE 1 KM curve for Houston PUD (2007–2016) higher risk of further turnover). Armed with detailed knowledge about employee turnover, a utility team can 100 deploy targeted programs to address their organization’s specific areas of concern or to support a request for addi- 75 tional funding or resources. The KM curve is a simple yet powerful graphical tool that can help HR professionals 50 and their managers respond to the patterns of turnover specific to their own organizations. 25 Corresponding author: Stanley Lam is the consulting 0 analyst at City of Houston–Public Utilities Division, 0 10 20 30 611 Walker St., Houston, TX 77002 USA; Tenure—years [email protected]. KM curve—Kaplan-Meier estimator, PUD—Public Utilities Division Write for the Journal Journal AWWA is currently seeking Highlighting is intended for emphasis. peer-reviewed and feature articles. Find submission guidelines at www.awwa.org/submit.26 LAM & CHAN | 109:2 • FEBRUARY 2017 | JOURNAL AWWA

Peer Reviewed Expanded SummaryTotal Intensity of Odor: A New Methodto Evaluate OdorsALLISON JACOBSEN-GARCIA, MELISSA DALE, ROY DESROCHERS, AND STUART KRASNERhttps://dx.doi.org/10.5942/jawwa.2017.109.0010 Unpleasant odors in drinking water can have a neg- unexpected off-odor with intensity of 6 (“weak toative effect on consumer confidence. In fact, consum- moderate”) or greater on the FPA scale of 0 to 12 isers often perceive that if their water smells bad, then likely to elicit consumer complaints. The secondaryit must be bad. Regulatory compliance in the United maximum contaminant level (SMCL) for TON is 3,States recommends that utilities test for odor aesthet- meaning that samples with TON ≤3 are in complianceics as a secondary standard using the threshold odor and considered free of significant odors. Samples withnumber (TON), a dilution-to-threshold test. However, TON >3 are out of compliance, indicating the presencethe drinking water community has recognized the of odors likely to cause complaints. On the basis ofshortcomings of the TON method, namely for being a MWDSC’s history with FPA and consumer complaints,poor measure of consumer acceptance, and as a result the proposed TIO SMCL is an intensity of 4 (“weak”)have sought better options to evaluate aesthetic char- on a scale up to 8, indicating samples with TIO ≤4 areacteristics of drinking water. Many utilities have in compliance and deemed free of significant odors.turned to flavor profile analysis (FPA), a descriptive Samples with TIO >4 are out of compliance and wouldsensory technique from which consumer preferences likely result in complaints.can be reliably predicted. However, the nature of themethodology is complex, requiring extensive panel Comparative testing revealed that TON falls shorttraining and time commitment. Additionally, FPA by as a viable indicator of consumer perception of odor,design is not intended to be a pass/fail measurement but TIO accurately predicts consumer acceptance/and thus is not the appropriate alternative test to TON rejection. The majority of the spiked municipal watersfor compliance purposes. (92%) reported low FPA intensities (FPA <6) and would not be expected to cause consumer complaints. Total intensity of odor (TIO) is an alternative However, TON did not gauge this correctly, with 48%method that was developed to reliably predict con- of the samples suggesting the presence of off-odor(s)sumer acceptance/rejection of drinking water. On the likely to cause complaints. TIO significantly outper-basis of the principles of FPA, TIO derives a numerical formed TON in predicting consumer attitudes, withintensity of a sample’s perceived odor using prescribed 83% agreement with FPA. These low odor eventsreference standards. Unlike TON, TIO samples are demonstrate the most common failure of TON: odor-evaluated in full, eliminating potential dilution anom- ants in water at intensities that are acceptable to con-alies related to individual chemical dose responses. A sumers but that need four or more dilutions to reachpass/fail intensity number can be established, analo- a nondetect odor.gous to TON, for use in compliance applications.Furthermore, TIO does not require specialty glassware Direct comparisons of methodologies not only haveor complex sample preparation (TON does) and does demonstrated that TIO is a valid method for compliancenot require as extensive a training and time commit- testing, but that it is indeed a superior test to the TONment as FPA. method for the aesthetic evaluation of odor in drinking water. The information gathered through this project Treated municipal drinking waters were spiked with supports the adoption of TIO as an approved analyticalcommon drinking water odorants and analyzed in method through Standard Methods and/or ASTMparallel for TON, TIO, and FPA. The objective was International. Once approved, regulators can amend theto mimic low odor events and compare the effective- secondary standard to include TIO as a valid method forness of TON and TIO at predicting consumer accep- odor evaluation in drinking water.tance. FPA was used as the benchmark to evaluateconsumer acceptance. It is well documented that FPA Corresponding author: Allison Jacobsen-Garcia is aintensity is a good predictor of consumer acceptance. chemist at the Metropolitan Water District of SouthernMetropolitan Water District of Southern California’s California, 700 Moreno Ave., La Verne, CA 91750(MWDSC’s) historical FPA experiences indicate an USA; [email protected]. JACOBSEN-GARCIA ET AL.  |  109:2 • FEBRUARY 2017  |  JOURNAL AWWA 27

Feature ArticleDOUGLAS BROWN, TIM RYNDERS, CHRIS STILLWELL, CHRIS DOUGLASS, AND SCOTT NIEBURBrackish Water Reverse Osmosis:A Proven Cost-Effective Renewable Water Supply PURCHASING AND Before 2006, the East Cherry Creek Valley (ECCV) Water and TREATING BRACKISH Sanitation District (Aurora, Colo.) relied exclusively on deep, GROUNDWATER TO SERVE nonrenewable groundwater supplies and was experiencing sig- AS AN ALTERNATIVE FOR nificant declines in pumping-water levels. ECCV realized that these DEVELOPING RENEWABLE supplies would continue to decrease as well as become more costlyWATER SUPPLIES HAS BEEN to pump and maintain, so it began investigating alternatives for developing A SUCCESS FOR A WATER renewable water supplies. The evaluation process considered not only the DISTRICT IN COLORADO. technical feasibility of using various surface water and tributary groundwater but also the water rights and the costs for converting agriculture water to municipal use, trans-mountain diversions, wastewater reuse, purchasing senior water rights, and storage of junior rights.28 BROWN ET AL.  |  109:2 • FEBRUARY 2017  |  JOURNAL AWWA

Because of the scarcity of high- Overhead view of the East Cherry Creek Valley Water and Sanitation District’s Northern Waterquality water supplies and the diffi- Supply Project Membrane Water Treatment Plant in Brighton, Colo.culty in acquiring these sources, thefinal solution was to purchase agri- Existing surface water supplies along Barr Lake’s surface water becomescultural water rights that were of Colorado’s Front Range were over- brackish in the alluvial well fieldbrackish quality because of their loca- appropriated, and getting permanent from the source water, which istion. This article describes the various water rights to high-quality surface influenced by urban return flows,factors evaluated to treat the brackish water supplies was expensive and agricultural irrigation, and seepagegroundwater to meet ECCV treated- complicated. Brackish water supplies from the up-gradient dam andwater goals and how this affects the along the South Platte River down- becomes mineralized as it flowslife-cycle costs of the treated water. stream of the Denver metropolitan through layers of soils and sedi-This project is of interest because it area were available and identified as ments and solubilizes calcium, mag-shows how the value of the treated a potential source. A similar source of nesium, and sodium salts. The sus-water in water-scarce areas can justify brackish alluvial groundwater had pended solids in the groundwaterthe use of advanced treatment tech- been previously developed for the are reduced by natural filtrationnologies to maximize yield and mini- Town of Brighton, and additional through this alluvial aquifer beforemize residuals disposal costs. alluvial water supplies were available being pumped to the facility. The from the alluvial aquifer downstream brackish groundwater has low sus-RENEWABLE BRACKISH of Barr Lake. Because this alluvial pended solids, as characterized by aGROUNDWATER SUPPLY It was apparent that the nontributary The ECCV Northern Water Supply groundwater would not be viable as the onlyProject Membrane Water Treatment long-term water supply.Plant (ECCV Northern MWTP) inBrighton, Colo., treats brackish water source has a significantly silt density index level <3, allowinggroundwater using a low-pressure higher concentration of total dis- the use of cartridge prefiltration fol-reverse osmosis (LPRO) process. The solved solids (TDS) and hardness lowed by primary LPRO membranealluvial brackish groundwater is a compared with the existing ground- desalination. The alluvial ground­renewable water supply that is water, RO was used before blending water has high hardness (>350 mg/Lrecharged with surface water from it with the existing deep nontributary as calcium carbonate [CaCO3]) andBarr Lake, a shallow irrigation supply groundwater supplies. a TDS ~1,000 mg/L. The existingreservoir. The ECCV NorthernMWTP is shown in the photographon this page. The ECCV project was first dis-cussed in 2002 as an approach toreduce dependence on deep nontribu-tary groundwater. This nontributarywater aquifer was ECCV’s first watersupply and was estimated to last 100years when initially developed in thelate 1960s. By 2000, ECCV’s servicearea had increased to 30,000 custom-ers, and the water levels in existingwells were decreasing 10–20 ft/year,leading to increased pumping andmaintenance costs. It was also appar-ent that the nontributary groundwa-ter would not be viable as the onlylong-term water supply. Pumping water levels are now typ-ically 800–1,000 ft below groundsurface and drop several feet per year.Exclusive reliance on deep ground­water was not sustainable, and ECCVneeded to develop renewable watersupplies to augment the groundwater. BROWN ET AL.  |  109:2 • FEBRUARY 2017  |  JOURNAL AWWA 29

ECCV water supplies had <100 mg/L scaling at high recoveries, so RO was trains with an ultraviolet (UV) disin-hardness and 400 mg/L TDS. determined to be the best fit. Table 1 fection system for alluvial water presents the water qualities at various bypassing the RO for blending. TheEVALUATION OF TREATMENT points in the treatment train. alluvial well field was designed toALTERNATIVES take advantage of the natural filtra- Design of the RO system. After suc- tion of the source water through the Treatment selection was based on cessful pilot testing, design of the alluvium, which has more than 60the alluvial water quality, with the full-scale 10-mgd RO plant with days of travel time from the nearestobjective of reducing the high hard- potential expansion to 40 mgd surface water source.ness and TDS before blending with moved forward in 2007. The pro-the higher-quality deep groundwater duction capacity of the first phase A cost-effective brackish water ROsupplies. The relatively low TDS and would later be expanded by the system was created by blending themoderate hardness resulted in a addition of brine minimization. The low TDS RO permeate with a por-relatively low potential for mineral design included two parallel RO tion of groundwater bypassing the RO system to produce finished waterTABLE 1 ECCV Northern Water Supply Project Membrane Water with approximately 300–350 mg/L Quality Parameters TDS and 100–130 mg/L total hard- ness. Ancillary processes include sec- RO RO Finished Final ondary RO membrane desalinationParameter Feed Permeate Water Concentrate for decreasing the primary RO con- centrate volume, UV disinfection ofpH 7.4 6.3 6.8a 7.9 membrane system bypass flows forConductivity—µS/cm 1,500–1,800 130 500–600 18,000–20,000 permeate blending, and chemicalTDS—mg/L 900–1,100 80 300–350 11,000–13,000 addition for finished water condi-Calcium—mg/L tioning and disinfection. The finalMagnesium—mg/L 110 3 40 1,300 concentrate flow is managed onsitePotassium—mg/L 30 1 10 400 by high-pressure injection into aSodium—mg/L 3 0.5 1.5 40 10,000-ft-deep US EnvironmentalAlkalinity—mg/L as CaCO3 150 20 60 1800 Protection Agency class I nonhazard-Chloride—mg/L 230 40 100 2,300 ous disposal well. Twelve wells pro-Sulfate—mg/L 200 10 70 2,500 vide brackish alluvial groundwater asSilica—mg/L 250 4 90 3,000 the raw water supply. Raw water 8 1 3 70–80 pretreatment includes the addition of a scale inhibitor followed by filtra-CaCO3—calcium carbonate, RO—reverse osmosis, TDS—total dissolved solids tion through 5-µm pleated polypro-apH was raised to 8.0 with sodium hydroxide before distribution. pylene cartridge filters before convey- ance to the primary RO systems.FIGURE 1 Treatment process schematic Unchlorinated groundwater for Distribution the 10-mgd initial phase of the facil- ity is supplied from the 12 wells, allScale inhibitor Primary RO Finished water approximately 80 ft deep, and isaddition and desalination treatment and conveyed to the plant via 1 mi of prefiltration raw water pipeline. Water delivered blending to the facility is divided into two streams: one to the LPRO systemBrackish groundwater and the other to the UV train for disinfection before blending with Secondary RO Ultraviolet the RO permeate to reduce the cor- desalination disinfection rosion potential. The well water is fed directly to the RO system to Deep well injection Brackish groundwater minimize potential introduction of oxygen and bacteria from the use ofRO—reverse osmosis raw water storage tanks. A sche- matic of the treatment process is shown in Figure 1. Using multiple supply wells without a raw water reservoir to attenuate30 BROWN ET AL. | 109:2 • FEBRUARY 2017 | JOURNAL AWWA

ARSENICIS NOTA GAMEReady to make the right move?AdEdge is your one-stop resource for arsenic treatment systemsfor any size community. We now supply Bayoxide E33 adsorptivemedia and E33 media replacements for the largest and thesmallest systems. As always, AdEdge offers competitive pricing,experienced engineers, and unparalleled support.With AdEdge’s arsenic treatment systems’ global successhistory, why would you go anywhere else?When you’re ready to make the right move, call us at 866.823.3343. www.AdEdgetech.com

The system’s low-pressure reverse osmosis unit consists of two stages split between pressure brackish water membranes. Thevessels operating in parallel and two banks for cleaning. resulting rated capacity of each RO unit is 3.3 mgd of permeate produc-minor flow variations did, however, for cleaning. Demineralized water tion. The interstage turbo boostercomplicate the startup sequence for permeates through the RO mem- lowers the concentrate pressure to 10the wells and RO systems. The raw branes and is gathered into a collec- psi so that it can be discharged to thewell water is initially treated with a tion header. Reject water from stage 1 concentrate pond during startup andscale inhibitor designed to control is also gathered into a collection when the brine minimization RO isCaCO3 scaling up to a Langelier satu- header and piped to a second stage not available. During normal opera-ration index of 2.1, eliminating the of RO treatment. The two-stage tion, the stage 2 concentrate is sentneed to depress the pH with sulfuric LPRO unit is shown in the photo- directly to the brine minimizationacid. The raw water passes through graph on this page. system to eliminate potential expo-5-μm cartridge filter vessels before sure to air, which might initiate pre-distribution to individual RO feed An interstage turbo booster is used cipitation of supersaturated minerals.pumps dedicated to each LPRO unit. to recover energy from the high- Minimizing the concentrate pressureThe cartridge filters remove large pressure stage 2 concentrate and to from the LPRO system also makes itparticles, biological material, debris, balance production between the first possible to use a low-pressure peri-sand, and grit to protect the RO and second stages. Stage 1 concen- staltic pump to feed hydrochloricmembranes from particle fouling trate is piped to the pump side of the acid (HCl) to depress the concen-and damage. The 300-hp vertical turbo booster, where it receives a pres- trate pH before further concentra- sure boost of approximately 20 psi. tion before deep well injection for ultimate disposal.Design of the full-scale 10-mgd reverse osmosisplant with potential expansion to 40 mgd moved The LPRO units operate at a recov-forward in 2007. ery of 83–85%. At design flows, each RO train will produce 3.3 mgd ofturbine pumps provide the pressure This allows for higher recoveries in permeate with less than 30 mg/L TDSboost necessary for the LPRO system the second stage without the added and 5 mg/L of total hardness. Thisto reach 83–85% recovery operating cost of an electric booster pump and permeate is combined with well waterat an average flux of 15 gfd. controls for adding significant perme- disinfected with a medium-pressure ate back pressure to stage 1. UV system to provide a total treat- Each RO train consists of two ment train capacity of 5 mgd. Becausestages. Feed water is piped to stage 1, The second stage of the RO trains acid is not used for pH depression ofwhere it is split between 48 pressure extracts additional permeate from the raw well water, the bicarbonatevessels operating in parallel and the stage 1 concentrate through simi- alkalinity is not converted to the car-divided into two banks of 24 vessels lar low-pressure, high-rejection bon dioxide that will pass through the membranes into the RO perme- ate. This eliminates the need to decar- bonate the permeate before final pH adjustment. The final blended prod- uct water is treated with sodium hydroxide for pH adjustment and sodium hypochlorite to provide dis- infection residual before being pumped more than 30 mi to the ECCV distribution network. Brine minimization. HCl is added to the concentrate from the LPRO system to reduce the scaling poten- tial of CaCO3 during the brine minimization process. LPRO con- centrate is kept from contact with air to avoid precipitation of super- saturated salts. A two-stage medium-pressure RO skid is used for brine minimization. The pho- tograph on page 36 shows an32 BROWN ET AL. | 109:2 • FEBRUARY 2017 | JOURNAL AWWA

Your advanced meteringtechnology partner.With advanced metering technology, your business can bemore efficient, more profitable and more service orientedthan ever before.Automated Meter Reading (AMR) and Advanced MeteringInfrastructure (AMI) are more than industry buzzwords – theyprovide a steady flow of valuable data to achieve your businessgoals, and provide consumers critical understanding of theirown usage habits. The result is better business and improvedstewardship of our water resources.©2016 Ferguson Enterprises, Inc. FERGUSON.COM/WATERWORKS 1216 333828

image of the two-stage brine mini- on the face of the membranes, there project, and design additions weremization skid. would be a more significant decrease undertaken as scope changes under in permeate production over the three- the construction management at-risk A phosphate-based antiscalant or year operating period represented. contract. Logistic and other chal-scale inhibitor added to the feedwater lenges that needed to be addressedahead of the LPRO process is concen- The design of the brine minimiza- during the design of the brine mini-trated to 600% in the LPRO and ulti- tion and deep well injection systems mization and deep well injection sys-mately to 1,500% in the brine mini- began after construction of the initial tems included the following:mization RO unit. A total of 3 mg/L RO facility. Using a secondary RO • Installation of all requiredThe alluvial well field was designed to take equipment within an estab-advantage of the natural filtration of the source lished and limited footprintwater through the alluvium, which has more than60 days of travel time from the nearest surface • Installation of a hazardouswater source. chemical (HCl) storage and pumping system without alter-of antiscalant in the feedwater is con- system to reduce the volume of con- ing the facility’s classificationcentrated to 60 mg/L at the ultimate centrate being disposed in the deep or generating the need fordischarge. The high concentration of well also produced additional perme- modification to the existingantiscalants prevents the supersatu- ate that was blended with permeate ventilation designrated salts in the brine from precipitat- from the LPRO to increase overalling in the short time they reside in the recovery to 94–96%. • Management and disposal ofRO units, as illustrated by the rela- off-specification water generatedtively flat lines for permeate flow in MANAGEMENT CHALLENGES through normal system startup,each stage in Figure 2. If there were The ECCV Northern MWTP was shutdown, and operationsignificant mineral scaling occurring a construction management at-risk • Conveyance and management of high-pressure corrosive waterPermeate Flow—gpmFIGURE 2 Operators’ trend chart for monitoring performance of at a significant flow from the RO stages process building to the deep injection well, which was RO stages approximately 200 ft away 1 2 The challenges of installing the 3 brine minimization and deep well 4 injection systems within a facility that was under construction mim- 2,000.0 icked the challenges involved in 1,800.0 modifying an existing facility. To effi- 1,600.0 ciently and accurately establish a 1,400.0 layout within the available space, a 1,200.0 three-dimensional model of the orig- 1,000.0 inal facility’s design was developed first, then any new equipment was 800.0 modeled and added to the facility 600.0 layout. The three-dimensional model 400.0 allowed the design team and ECCV 200.0 to visualize and troubleshoot the facility layout as soon as the initial 0.0 design was developed. 9/14/2011 Injection of HCl into the LPRO 4/1/2012 concentrate for pH suppression 10/18/2012 resulted in several space and safety 5/6/2013 challenges. Because of the bicarbon- 11/22/2013 ate alkalinity levels, it was estimated 6/10/2014 that 300 mg/L HCl would be required 12/27/2014 to reduce the pH level to 6.7. At the design flow, this resulted in an injec- Date tion volume of 300 gpd of 30% HCl solution and required a large storageRO—reverse osmosis tank with double containment. Space34 BROWN ET AL. | 109:2 • FEBRUARY 2017 | JOURNAL AWWA

V E R S AT I L I T Y.Choose steel.  Decades of documented performance  Zero leak tolerance  High-performance, NSF 61-compliant coatings extend maintenance cycles  Multiple competitive bids on projects  Sustainable community asset  Lower cost of ownership than other materials: Calculate Total Cost of Ownership (TCO) with our free online tool!* LEARN ABOUT STEEL WATER TANKS FROM THE EXPERTS! Municipal and utility administrators, water tank owners and operators, our Field Erected Water Storage Tanks Seminar* is offered several times each year in major metropolitan area cities. Topics include:  Using and interpreting the standards  Locating, sizing, and selecting tanks  New tank specifications  Tank construction  Tank coating  Inspection and maintenance *Scan here to go to WWW.STEELTANK.COM

This two-stage brine minimization skid design includes two parallel reverse osmosis trains Brine disposal. The plant dischargewith an ultraviolet disinfection system. requirements at the ECCV facility are highly constrained. No concen-inside the equipment building was secondary containment and the under- trate may be sent to the local seweralready being used for the brine min- ground location eliminated the need system, nor may it be discharged toimization skids and high-pressure for ventilation and heating the vault local surface water ditches withoutbrine injection pumps, so options to except for heat tracing and insulation raw water blending, meaning all off-build a separate chemical storage area of the fume scrubber. An HCl carrier specification water would be treatedwere evaluated. pipe was routed below grade to the as if it were concentrate. To manage process building. Two peristaltic the off-specification water separately, The ultimate solution was to pumps provided the capacity and suc- a system to collect and store it wasinstall a 5,500-gal tank in an under- tion lift required to move chemicals implemented. All off-specificationground vault with peristaltic pumps. from the vault to the injection point. water generated by the system is nowThe precast concrete vault provided collected and stored onsite in a syn- thetic, flexible, lined storage pondFIGURE 3 Deep well injection overall pressure and ow providing 17 acre-ft of storage. 1,800 Pressure 2013 2014 2015 1,800 One of the innovative features of 1,600 Flow 1,600 the ECCV facility is its use of a deep 1,400 2012 1,400 class I nonhazardous injection well 1,200 1,200 for disposing the highly concentratedDeep Well Injection Pressure—psi 1,000 1,000 Deep Well Injection Flow—gpm waste brine. Testing of the injection 800 well showed that an inlet pressure of 800 600 3,100 psi would ultimately be 600 400 required for a 400-gpm brine flow 400 200 rate, which is typical for oil and gas 200 0 salt water disposal wells. The design flow rate for the injection well was 0 established in the 200–450 gpm range. Well testing indicated that6/20/2012 these flows would initially result in 1/6/2013 backpressures of 700–1,500 psi. 7/25/2013 2/10/2014 To generate the pressure required 8/29/2014 for injection, two horizontal multi- 3/17/2015 stage centrifugal pumps were 10/3/2015 installed in the membrane water treatment facility. In their current Date parallel configuration, the pumps can operate in a duty/standby mode. Each pump is approximately 30 ft long; the majority of that overall length is the 45 stages required to develop high pressures. Each pump is driven by a 600-hp variable-speed electric motor. Over time, the 10,000-ft-deep shale and sandstone formations where the brine is injected become saturated and the backpressure on the deep well injec- tion pumps increases. Fouling of the formations may occur as the high-TDS brine enters the well and undergoes various geo- chemical reactions under elevated pressure and temperature. In the event that the injection well backpres- sure increases over time, the discharge36 BROWN ET AL. | 109:2 • FEBRUARY 2017 | JOURNAL AWWA

THE Our Q46 Series monitoring systems are economical to purchase,economical to maintain, and provide long term accuracy and reliability for all your monitoring needs!Free Chlorine Dissolved Oxygen pH/ORPCombined Chlorine Suspended Solids ConductivityTotal Chlorine Residual Sulfite Dissolved OzoneFluoride Dissolved Sulfide Chlorine DioxideTurbidity Permanganate Hydrogen PeroxideDissolved Ammonia Peracetic Acid Auto-Clean Total Chlorine Toxic & Combustible Dissolved Oxygen Monitor Monitor Transmitter• Optical or membrane sensors available • A Gas-Phase approach to Chlorine Measurement • Interchangeable “Smart Sensors”• A blast of air keeps the Auto-Clean sensor clean • Prevents sensor fouling • Optional Sensor Daily Auto-Testing800-959-0299 www.analyticaltechnology.com

piping system has been designed to pumping and $6 million for deep hardness, but this results in increasedfacilitate reconfiguration of the injection well drilling, high-pressure costs for salt in ion exchange regen-pumps into a series arrangement that pumps, and the brine minimization eration, lime or sodium hydroxidecan boost the injection pressures up system. As of 2015, the combined for precipitating hardness, and theto 3,200 psi. Overall pressure and amortized capital and operating disposal of additional regenerationflow are shown in Figure 3. costs, including power, labor, brines or precipitated calcium sludge. chemicals, and administrative allo- All brine and off-specification cations, were estimated to be Normally, integrity testing of ROwater pumped by the high-pressure systems treating groundwater is not required because there are noThe ECCV state-of-the-art high-recovery requirements for reduction of GiardiaRO system is achieving an overall recovery and Cryptosporidium. The shallowof 94–96% with no offsite process alluvial wells that supply the ECCVwastewater discharge. RO system were tested for micro- scopic particles, but there were nodeep well injection system are $2.80/1,000 gal of product water indications that the wells were underdelivered to the deep injection well ($917/acre-ft). The ECCV raw the direct influence of surface water.via a below-grade high-pressure water supply well field, raw water ECCV wanted to demonstrate therepipeline. The total length of the delivery system, and the finished were additional barriers to potentialpipeline is approximately 200 ft. water pipeline and pumping systems pathogen contamination in theAll pump discharge piping, valves, were constructed independently of future, so it does routine integrityand appurtenances were selected the RO facility; those costs were testing of the RO membranes andand designed to withstand the approximately $70 million. The uses UV disinfection for the blend-build-out pressure of 3,100 psi. Oil 48-in.-diameter finished water ing flow. RO system integrity testingand gas technology was used to pipeline has a capacity of 47 mgd, is achieved by using the natural sul-develop this system. which will support expansion of the fate ion as a surrogate for Crypto- ECCV Northern MWTP for ECCV sporidium and Giardia. There is Data analysis performed biannu- and other utilities in southeast enough sulfate in the feedwater thatally continues to provide insight metropolitan Denver. the RO membrane can consistentlyinto long-term injection well behav- achieve a more-than-99% rejectionior. The data confirm that injection CONCLUSIONS of all sulfate, which demonstratespressure requirements are increas- The ECCV state-of-the-art high- 2 logs of rejection.ing yearly at a relatively predictablerate based on the normalized pres- recovery RO system is achieving an The reliable, cost-effective opera-sure trends. Also, the injection pres- overall recovery of 94–96% with no tion of a complex and intercon-sure requirements decrease after offsite process wastewater discharge. nected multistage RO system isperiods of rest or low flow but The typical brackish water RO plant achievable in a municipal treat-quickly return to previous high val- operates at 75–85% recovery, so ment system. The use of a high-ues when flow rate is increased. The approximately 15–25% of the raw pressure deep well injection systemrise in normalized pressure may be water supply is discharged as waste; allows for the cost-effective dis-indicative of well fouling or this represents a significant water loss posal of the RO concentrate at anincreased pressure mounding in the for areas with limited supplies. In inland site. The implementation ofinjection zone, but the injection addition, disposal of the waste the ECCV RO facility establishespressure fall-off after brine injec- streams can be a limiting criterion for the treatment of alluvial brackishtion is stopped indicates that most projects without access to an ocean water as a renewable source ofof the pressure increase is likely the discharge. The ability of the ECCV high-quality potable water. Theresult of pressure mounding in the Northern MWTP to achieve 96% findings have provided a wealth ofinjection zone. recovery represents a 70% reduction information regarding the feasibil- in brine volume compared with oper- ity of using RO for brine minimiza- Costs. The initial capital cost for the ation at 80% recovery. The costs of tion, and for using deep injectionECCV Membrane Water Treatment brine disposal are also reduced pro- wells as the primary long-termFacility was $33 million (in 2012), portionally. RO facilities that achieve method for drinking-water RO-which included approximately more than a 90% recovery typically concentrate disposal. The lessons$27 million for the RO facility soften the feedwater to remove learned through four years of con-infrastructure and finished water tinuous operation can be used by engineers and operations staff to optimize designs for new systems or when planning for the future38 BROWN ET AL. | 109:2 • FEBRUARY 2017 | JOURNAL AWWA

maintenance and operation of Brown has been active in the Tim Rynders is a project managerexisting systems. research and design of high-recovery and Chris Stillwell is a project RO processes and brine engineer, both with CDM Smith,ABOUT THE AUTHORS minimization technologies to allow Denver. Chris Douglass is a project Douglas Brown is for increased use of inland brackish manager and Scott Niebur is the a project manager water supplies in water-short areas. operations manager, both with East with CDM Smith, He is a regular presenter at Cherry Creek Valley Water and 555 17th St., Ste. membrane conferences and is Sanitation District, Aurora, Colo. 1100, Denver, CO involved with the AWWA 80202 USA; Membranes Standards Committee. https://dx.doi.org/10.5942/jawwa.2017.109.0020 browndr@ AWWA RESOURCEScdmsmith.com. He began designingbrackish water reverse osmosis (RO) • Pressure-Retarded Osmosis as Energy Recovery for Reverse Osmosissystems in 1990 to reduce nitrates Desalination: Module-Scale Modeling and Specific Energy. O’Toole, G. &and total dissolved solids in Achili, A., 2016. 2016 AWWA/AMTA Membrane Technology Conferencegroundwater for potable water use. Proceedings. Catalog No. MTC_0084024.Since then, he has designed morethan 10 large municipal RO systems • Programmatic Delivery Key to New Water Supply: San Antonio’s Brackishand has seen the technology evolve Groundwater Desalination Program. Davis, B.; Harrah, E.; & Timmermann,from a specialized treatment process D., 2015. Journal AWWA, 107:3:59. Product No. JAW_0081661.requiring high pressures, acid feed,and proprietary membrane design to • EL-10: Filtration Basics. Elearning self-paced. Catalog No. EL10.the low-pressure designs usinghighly effective mineral scale These resources have been supplied by Journal AWWA staff. Forinhibitors that are common today. information on these and other AWWA resources, visit www.awwa.org. BROWN ET AL. | 109:2 • FEBRUARY 2017 | JOURNAL AWWA 39

Feature ArticleTARRAH HENRIE, SARAH PLUMMER, AND J. ALAN ROBERSONOccurrence and State Approachesfor Addressing Cyanotoxins inUS Drinking Water C yanotoxins have long been recognized as potential sources ofCYANOTOXINS ARE contamination for drinking water supplies. As far back as 1930,BECOMING A MORE there have been suspected links between algal blooms in munic-IMPORTANT ISSUE FOR ipal water supply source waters and gastroenteritis outbreaksWATER SYSTEMS SINCE THE (Dillenberg & Dehnel 1960). However, cyanotoxins are not yetUS ENVIRONMENTAL regulated in the United States under the Safe Drinking Water Act (SDWA).PROTECTION AGENCY Although cyanobacteria and their toxins have been identified as microbialRELEASED HEALTH contaminant candidates on the four Contaminant Candidate Lists (CCLs)ADVISORIES (HAS) AND that the US Environmental Protection Agency (USEPA) has issued under theRECOMMENDATIONS FOR SDWA to date, cyanotoxins have not yet been included in any of the past threeSYSTEMS’ RESPONSES IN Unregulated Contaminant Monitoring Rules (UCMRs), primarily due to the2015. THIS ARTICLE lack of a robust and reliable analytical method at the time.SUMMARIZES HOW STATEPRIMACY AGENCIES HAVE In August 2014, the City of Toledo, Ohio, issued a do not drink/do not boilRESPONDED TO THE HAs. (DND/DNB) notice to all customers served by the city’s water system because of a detection of microcystin in the drinking water that exceeded the lifetime chronic exposure threshold concentration of 1 μg/L established by the Ohio Environmental Protection Agency at that time. This weekend-long DND/DNB notice affected nearly 500,000 residents and businesses served by the City of Toledo and was the largest DND/DNB order caused by cyanotoxin contamination in US history.40 HENRIE ET AL. | 109:2 • FEBRUARY 2017 | JOURNAL AWWA

This event was quickly followed microcystins and 3.0 μg/L for through the Water Industry Technicalby legislative and regulatory efforts cylindrospermopsin. Action Fund (WITAF), initiated ato manage the risk of algal toxins in study to investigate the observed lev-drinking water. On Jan. 8, 2015, the Concurrent with the HAs, USEPA els and associated actions by stateDrinking Water Protection Act released recommendations to water primacy agencies during the 2015(H.R. 212) was introduced into the systems for monitoring, treatment, cyanotoxin-producing bloom season.US House of Representatives to and public communications (USEPA The purpose of the study was toamend the SDWA by requiring 2015d). These recommendationsUSEPA to develop a strategic planfor assessing and managing risks Cyanotoxins have long been recognizedassociated with algal toxins in as potential sources of contaminationdrinking water provided by public for drinking water supplies.water systems. This bill was signedinto law on Aug. 7, 2015 (P.L. 114-45). used a tiered approach, with collect and analyze data to supportAlthough this bill did not specify or increasing actions recommended as development of state guidance andrequire regulatory limits for cyano- concentrations increased. to set the stage for future data col-toxins, it emphasized the need for a lection efforts. The report providedplan to manage cyanotoxins at the To meet its requirement in P.L. 114- background information describingfederal level. 45, USEPA released an Algal Toxin the guidance currently in place for Risk Assessment and Management cyanotoxins in drinking water, the In summer 2015, before ratifica- Strategic Plan for Drinking Water in data sources used for collecting cya-tion of H.R. 212, USEPA issued late 2015 (USEPA 2015e). This stra- notoxin occurrence data, availablehealth advisories (HAs) for two spe- tegic plan describes a number of cur- data, and recommendations forcific cyanotoxins: microcystins and rent and future initiatives that USEPA future data collection and analysiscylindrospermopsin. HAs are in- is undertaking to better understand efforts (AWWA 2016).tended to provide information for and appropriately manage the riskspublic health officials on pollutants from cyanotoxin-producing algal FEDERAL AND STATE APPROACH TOthat are not currently regulated blooms. Perhaps most important CYANOTOXINS IN DRINKING WATERunder the SDWA but are capable of from a regulatory perspective is theaffecting drinking water quality. repeated inclusion of cyanotoxins in Because cyanotoxins are not regu-HAs do not establish regulatory the final fourth CCL (81 FR 81099), lated at the federal level, state pri-limits but instead identify the con- and the publication of two liquid macy agencies can issue their owncentration of a contaminant in chromatography–tandem mass spec- guidance or regulations for cyano-drinking water below which adverse trometry methods for cyanotoxin toxins in drinking water. To date,health effects are not anticipated to analysis (USEPA 2015f, 2015g). three states have proactively issuedoccur over specific exposure dura- Cyanotoxins were not included in guidance or regulation for cyano-tions. USEPA also developed a the first three UCMRs because cya- toxins in drinking water. These val-health-effects support document for notoxin analytical methods were ues are summarized in Table 1antitoxin-a but concluded that insufficient (USEPA 2015h). The (Minnesota Department of Healthavailable toxicity data were inade- development of the new analytical 2015, OEPA 2015, OHA 2012).quate for deriving a specific health- methods, the listing of cyanotoxinsbased value (USEPA 2015a). on the final CCL 4, the listing of total Ohio and Oregon have guidance microcystins, as well as six individual values for anatoxin-a, saxitoxin, and For both microcystins and cylin- microcystins (-LA, -LF, -LR, -LY, -RR, cylindrospermopsin in addition todrospermopsin, USEPA established -RY), nodularin, cylindrospermopsin, microcystins. As Table 1 shows,two 10-day HA levels based on age and anatoxin-a in the final UCMR 4 Ohio has established different guid-ranges; a lower HA value was estab- (81 FR 92666) all point to the prob- ance values for children and sensi-lished for bottle-fed infants and able development of maximum con- tive populations for microcystins,young children of preschool age, taminant levels for microcystins, consistent with the HAs issued bywhile a higher HA value was estab- cylindrospermopsin, anatoxin-a, USEPA in 2015. The Minnesota andlished for school-age children and and/or other cyanotoxins. Ohio values were updated in 2015adults (USEPA 2015b, 2015c). The (Minnesota Department of Healthmicrocystin and cylindrospermop- On the basis of these consider- 2015; OEPA 2015, 2014). Previ-sin HA levels for children under the ations, and recognizing the scarcity ously Ohio had a microcystin guide-age of six were set at 0.3 μg/L and of existing data on cyanotoxin con- line of 1 µg/L, and Minnesota had a1.6 μg/L respectively. For children centrations in source water, AWWA, guideline of 0.04 µg/L.over the age of six and adults, theHA level was set at 1.6 μg/L for HENRIE ET AL. | 109:2 • FEBRUARY 2017 | JOURNAL AWWA 41

States are taking a variety of of the states have subtle nuances on representatives cited legal concernsapproaches for responding to USEPA’s one or more topics. No algal toxin over enforcing a standard that is notHAs. As part of this project, state pri- expert was reached in 14 states. Five in regulation. Ohio is taking themacy agencies were contacted to states are reviewing or developing an most proactive steps. States are taking a variety of approaches OCCURRENCE for responding to USEPA’s health advisories. In general, there appear to bedetermine how they were approaching approach to addressing cyanotoxins increasing occurrences of cyanotoxin-the issue of cyanotoxins in drinking in drinking water. producing algal blooms worldwidewater. These data were collected (de Figueiredo et al. 2004). Whileindependently by the Association of Most states are not tracking this there are still many unknownsState Drinking Water Administrators issue closely, with many responding regarding cyanobacteria ecology, anfrom a survey of states. A summary that they do not plan to take action increase in cyanotoxin-producingis shown in Table 2; however, many on the HA because the USEPA advi- algal blooms in the past has been sories are not regulations. Some state linked to two major factors. The first is the eutrophication of fresh-TABLE 1 States’ guidance for cyanotoxins summary water sources caused by nutrients, primarily nitrogen and phosphorus Microcystins Cylindrospermopsin Anatoxin-a Saxitoxin (Yuan et al. 2014, Dolman et al.State μg/L μg/L μg/L μg/L 2012). The other is climate change, which is producing a warmingMinnesota 0.1a NG NG NG trend in the majority of water bod- 0.2 ies (O’Reilly et al. 2015). WarmingOhiob 0.3 0.7 20 0.2 trends may increase the number 3 and intensity of harmful algalOhioc 1.6 3.0 20 blooms that in turn have a signifi- cant effect on monitoring and man-Oregon 1 1 3 agement of bloom events (Paerl & Paul 2012; Backer & Moore 2010;NG—no guidance Delpla et al. 2009).aMicrocystin-LRbChildren less than six years old and sensitive populations RECENT OCCURRENCE DATAcChildren six years old or more and adults IN THE UNITED STATESTABLE 2 States’ approach to USEPA health advisories Currently available data do not capture the full extent of cyano- State Action on Monitoring Intend to Written Guidance toxin occurrence in the United Health Required? Collect Complete or in States because the majority of statesOH, RI Data? Development? do not require cyanotoxin monitor-MD Advisory? Yes ing in drinking water. In states withAL, CO, CT, IL, KS, No Yes Yes testing programs, the monitoring Yes No Yes Yes data are not in a consistent format MA, ME, NH, OR, Yes Yes Yes from state to state. Since there is no VT No USEPA requirement for testing,SC various methods are used, but thoseCA, WI Yes No Yes No methods often are not indicatedAR, IA, UT No No No Yes with the data.AK, AZ, DE, FL, HI, No No Yes No MN, MT, NC, NM, No No No No Data presented in this section were NV, OK, PA aggregated from state data thatGA, ID, IN, KY, LA, No algal toxin expert reached were available on the Internet or MI, MS, NE, ND, were directly obtained from states SD, TN, VA, WA, Currently reviewing or developing an approach to addressing in December 2015. No states pro- WV cyanotoxins in drinking water vided surrogate data, such as cellMO, NJ, NY, TX, WY counts. Table 3 provides details on the source of the data collected for this study. Knowing that some rec- reational water bodies are also42 HENRIE ET AL. | 109:2 • FEBRUARY 2017 | JOURNAL AWWA

TABLE 3 Summary of state dataState Recreational Source Treated Total Years Website Samples Drinking Drinking Samples Included Water Water Samples SamplesMinnesota 671 0 0 671 2006–2007, Data not online 2012Ohio 2,741 4,869 2,678 http://epa.ohio.gov/ddagw/HAB.aspxOregon 0 129 27 10,301 2010–2015 Data not onlineVermont 3 532 532 156 2011, 2015 www.drinkingwater.vt.gov/pcwswqbga.htmWashington 0 1,067 2015 www.nwtoxicalgae.org/ 6,593 0 6,593 2000, 2007–2015West Virginia 0 48 24 72 2015 www.wvdhhr.org/oehs/public_health/BGA_Sample_ Results.aspdrinking water sources, recreational test for saxitoxin but did test for the only data available. Although 45%cyanotoxin data were included. Each other three cyanotoxins. Vermont of recreational samples had cyano-state classified the data as recre- also tested for cylindrospermopsin. toxin detections, it should be notedational, source drinking water, or that many states sample only whentreated drinking water. In some In Figure 1 the total number of blooms have been visually detectedcases, the recreational data are from samples is shown in blue and include in a water body. Out of the 3,261the same water body as the source all of the cyanotoxins tested. These treated water samples collected,drinking water data; however, the data are further divided into recre- 2.7% of those were detections. All ofsource drinking water samples were ational water data (orange) and the treated water detections werecollected at an intake to a water drinking water data (green). Each from Ohio.treatment plant, whereas the recre- small cube represents 37 samples.ational samples were collected at Figure 2 shows the seasonal pat-nonsource water locations. Recreational data are included tern of microcystin occurrence from because, in most states, those are the The majority of drinking waterdata are from Ohio (7,547 out of FIGURE 1 Cyanotoxin occurrence data in drinking and recreational water8,839 samples). Additional drinkingwater data are from Oregon, 37 samplesVermont, and West Virginia. Thefreshwater recreational cyanotoxin 8,839 drinking water samples 37.6% 2.7%data are from Minnesota, Ohio, source DW treated DWVermont, and Washington. Although 5,578 source DW samples detections detectionsthe focus of this study was the 2015 3,261 treated DW samplesbloom season, older data wereincluded when available. 18,860 total samples For this project, anything reported 10,021 45.1%as below the detection limit was recreational recreationalincluded as zero. Data points reported detectionsas greater than a given value (122 samplesdata points) were included as thatgiven value. For example, data DW—drinking waterreported as greater than 5 µg/L wereincluded as 5 µg/L. Every state withmonitoring results tested for micro-cystin. Ohio and Washington alsotested for saxitoxin, anatoxin, andcylindrospermopsin. Oregon did not HENRIE ET AL. | 109:2 • FEBRUARY 2017 | JOURNAL AWWA 43

2006 through 2015. Sixty-one data shown as light-orange squares, and the highest concentrations occurring inpoints (1%) of recreational data are source drinking water concentrations the late summer and early fall.not shown because they are so high are shown as light-green diamonds. Figure 3 presents the concentrationthat they obscure the rest of the distribution of microcystins in bothdata. The highest reported level was Although detections of microcys- source drinking water and recre-80,000 µg/L. Recreational data are tin are possible year-round, the data ational water. Those samples of showed a seasonal trend with the source drinking water with the highest concentrations of microcys-FIGURE 2 Recreational and source drinking water microcystin data tins are still substantially lower than (2006–2015) the concentration of microcystins in the highest concentration recre- Recreational water ational water samples. Source drinking water The majority of source drinking 2,000 water samples have nondetect results, with 57.5% of samples 1,800 below detection limits. Many of the recreational water samples also had 1,600 nondetect results, with 44.3% of samples below detection limits. In 1,400 source drinking water, 51.4% of those water samples with detectionsMicrocystin—μg/L 1,200 are between 0.3–1.6 µg/L. The recre- ational water detections have higher 1,000 concentrations overall with 30.1% in the 1.6–20 µg/L range. 800 Ohio cyanotoxin data. Since the 600 majority of cyanotoxin occurrence data was gleaned from the Ohio 400 database, it is worth looking at Ohio data separately from the other data 200 sets. Analyses of the drinking water data set collected from Ohio are pre- 0 Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. Jan. sented in this section. Jan. 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2006 Date Ohio drinking water data. Most of the drinking water data comeFIGURE 3 Microcystin concentration histogram from Ohio, so it is informative to look at those data in more detail. 3,000 Source drinking water Figure 4 shows drinking water sam- Recreational water ples in light green. The detection data are inclusive of all cyanotoxins 2,500 that were tested (microcystin, cylin- drospermopsin, anatoxin-a, saxi-Number of Samples 2,000 toxin). Samples exceeding the micro- cystin HA level of 0.3 µg/L are 1,500 shown in yellow. 1,000 Although nearly 44% of source drinking water samples exceeded the 500 microcystin HA of 0.3 µg/L, just over 1% of treated drinking water sam- 0 <0.3 0.3–1.6 1.6–20 20–100 100–2,000 >2000 ples exceeded that same threshold. ND Microcystin—μg/L No significant correlation (R2 of 0.55) between source water concen-ND—not detected trations and treated water concentra- tions was observed. Anatoxin-a, cylindrospermopsin, and saxitoxin were not detected44 HENRIE ET AL. | 109:2 • FEBRUARY 2017 | JOURNAL AWWA

above the HA or Ohio limits in the detection of microcystin in treated sample. This incident resulted in thesource or treated drinking water. water, followed by two days of con- first DND order in Ohio due to cya-Table 4 summarizes the number of firmation samples ranging in levels notoxins (OEPA 2014).source drinking water samples col- from 1.43 to 3.56 µg/L. No addi-lected in Ohio and the range of tional samples were collected until The 2014 Toledo event includedresults for each cyanotoxin. Table 5 10 days after the final confirmation 15 treated water detections thatsummarizes the number and range of occurred between August 1 andresults for treated drinking watercyanotoxin detections. FIGURE 4 Microcystin occurrence data in Ohio drinking water (2010–2015) Of the 254 treated drinking water 16 45.10% 1.43%saxitoxin samples collected, there samples source DW treated DWwere 56 detections. The average detections detectionsdetection limit for saxitoxin was 6,405 drinking water samples0.022 µg/L, which is approximately 4,162 source DW samplesan order of magnitude lower than the 2,243 treated DW samplesaverage microcystin detection limit(0.23 µg/L). This difference in detec- 8,323 microcystin samples 43.90% 1.16%tion limit likely contributes to the 604 sample pointshigher number of detections. For per- samples with microcystin >0.3 µgspective, the highest saxitoxin detec- DW—drinking watertion was 0.064 µg/L, which is onlyslightly higher than the minimum TABLE 4 Cyanotoxin detections in Ohio source drinking watermicrocystin detection of 0.054 µg/L. Average/ Saxitoxin samples were collected Minimum Maximum Result Number of Samplesfrom 32 locations, and 65% of the Cyanotoxin μg/L μg/L μg/L Detections Collecteddetections in the source water werefrom upstream of a single plant Anatoxin-a    2 1 4(Barberton). In treated drinkingwater, 82% of saxitoxin detections Cylindrospermopsin     0.11 1 270were from samples collected at theBarberton water treatment plant Microcystin 0.054 20,000 25 1,877 4,162(WTP). At the Barberton WTP therewere 44 detections from July 31 Saxitoxin 0.022 0.88 0.22 164 433through Sept. 14, 2015. That planthad two more detections—one on TABLE 5 Cyanotoxin detections in Ohio treated drinking waterSeptember 21 and the other inDecember. The remaining treated Average/water detections were from three Minimum Maximum Result Number of Sampleslocations—two locations with single Cyanotoxin μg/L μg/L μg/L Detections Collecteddetections and one location (West Anatoxin-aFarmington WTP) with eight detec-    02tions over a nine-day period. Cylindrospermopsin     0 179 Regulatory implications in Ohio.A more detailed study of the Ohio Microcystin 0.054 11 1.33 32 2,243finished water detections is helpful inunderstanding the occurrence and Saxitoxin 0.022 0.064 0.037 56 254impact of cyanotoxins in drinkingwater production, and in understand-ing the implications of the current HAfor utilities. All of the microcystindetections over 0.3 µg/L at OhioWTPs occurred during nine incidents,which is defined as a treated waterdetection over 0.3 µg/L. In September 2013 at the CarrollTownship WTP, there was an initial HENRIE ET AL. | 109:2 • FEBRUARY 2017 | JOURNAL AWWA 45

August 19. Under the current regu- for microcystins. It is not clear control are other areas of interest.latory framework, Toledo would whether this is fully representative Sarah Plummer is a junior waterhave likely had to issue a second of the potential risks from cyano- process engineer at CoronaDND/DNB order in mid-August toxin contamination, or if ex- Environmental Consulting in2014 in addition to the notice that panded monitoring beyond micro- Newark. J. Alan Roberson was thewas in effect August 2–4. The 2014 cystins is more appropriate. director of policy at Corona Environmental Consulting at theIt would be useful to investigate why there time this article was written. are so many single-data-point detections of cyanotoxins reported in finished water. https://dx.doi.org/10.5942/jawwa.2017.109.0022Toledo event included 15 treated Cyanotoxins are clearly a regulatory REFERENCESwater detections that occurred priority for USEPA, given their inclu-between August 1–19. This incident sion on the final CCL 4, the final AWWA, 2016. Cyanotoxins in US Drinkingaccounts for 57% of the detections UCMR 4, and the 2015 HAs and Water: Occurrence, Case Studies andabove 0.3 µg/L in treated drinking action plan. While it is not clear State Approaches to Regulation. www.water among all Ohio samples. which specific cyanotoxins might be awwa.org/resources-tools/water- regulated and at what levels, state knowledge/cyanotoxins.aspx (accessed The other seven incidents of primacy agencies and water systems Jan. 11, 2016).treated water detections had initial will have to continue to workdetections, but the follow-up sam- together as the UCMR 4 monitoring Backer, L. & Moore, S., 2010. Harmful Algalpling did not confirm these initial data begin to be released in late Blooms: Future Threats in a Warmerdetections. Three of the detections 2018. Now is the time for every sys- World. Nova Science Publishers Inc.,were on the same day, July 24, 2015, tem to prepare appropriately. New York.at three locations. The samples wereanalyzed at the same laboratory. ACKNOWLEDGMENT de Figueiredo, D.R.; Azeiteiro, U.M.; Esteves,Source water concentrations ranged The authors would like to thank S.M.; Gonçalves, F.J.M.; & Pereira, M.J.,from nondetect to 37.2 µg/L. 2004. Microcystin-Producing Blooms—a AWWA, which supported this project Serious Global Public Health Issue.RECOMMENDED NEXT STEPS through WITAF, and research partners Ecotoxicology and Environmental Safety, This project was initiated to col- Damon Roth and David Cornwell at 59:2:151. https://doi.org/10.1016/j. Cornwell Engineering Group. ecoenv.2004.04.006.lect information from state primacyagencies on the levels of cyanotoxins ABOUT THE AUTHORS Delpla, I.; Jung, A.-V.; Baures, E.; Clement, M.;in recreational water and drinking Tarrah Henrie is a & Thomas, O., 2009. Impacts of Climatewater and the resulting impacts. water process Change on Surface Water Quality inAfter the compilation and analysis of scientist with Relation to Drinking Water Production.the data, two trends warrant further Corona Environment International, 35:8:1225.investigation. Environmental https://doi.org/10.1016/j. Consulting, 39899 envint.2009.07.001. First, in many instances, detec- Balentine Dr., Ste.tions of cyanotoxins in the finished Dillenberg, H.O. & Dehnel, M.K., 1960. Toxicwater are not replicated in follow- 200, Newark, CA 94560 USA; Waterbloom in Saskatchewan, 1959.up sampling. It is not clear from the [email protected]. She has Canadian Medical Association Journal,data whether these detections rep- over 13 years of experience solving 83:22:1151.resent false positives, actual detec- inorganic and organic drinkingtions that are not repeated as a water contaminant and regulatory Dolman, A.M.; Rücker, J.; Pick, F.R.; Fastner,result of changes in external condi- compliance challenges for utilities. J.; Rohrlack, T.; Mischke, U.; & Wiedner,tions or plant operation, or other For the past several years, Tarrah C., 2012. Cyanobacteria and Cyanotoxins:phenomena. It would be useful to has been involved with hexavalent The Influence of Nitrogen Versusinvestigate why there are so many chromium research, planning, Phosphorus. PloS One, 7:6:e38757.single-data-point detections of cya- treatment, and legislation. Nitrate, https://doi.org/10.1371/journal.notoxins reported in finished water. ammonia, nitrification, pone.0038757. 1,2,3-Trichloropropane, Second, by far the majority of cyanotoxins, and cross-connection Minnesota Department of Health, 2015.the analyses for cyanotoxins dur- Toxicological Summary for:ing the 2015 bloom season were Microcystin-LR. Minnesota Department of Health, St. Paul, Minn. www.health. state.mn.us/divs/eh/risk/guidance/gw/ microcystin.pdf (accessed Jan. 11, 2016). OEPA (Ohio Environmental Protection Agency), 2015. Public Water System Harmful Algal Bloom Response Strategy. OEPA, Columbus, Ohio. http://epa.ohio. gov/Portals/28/documents/HABs/PWS_ HAB_Response_Strategy.pdf (accessed Jan. 11, 2016).46 HENRIE ET AL. | 109:2 • FEBRUARY 2017 | JOURNAL AWWA

OEPA, 2014. Draft Public Water System Microcystin Toxins. 820R15100. USEPA, dirEntryId=306953. (accessed Sept. 2, Harmful Algal Bloom Response Strategy. Washington. 2016). OEPA, Columbus, Ohio. USEPA, 2015c. Health Effects Support USEPA, 2015g. Method 545: Determination ofOHA (Oregon Health Authority), 2012. Document for the Cyanobacterial Toxin Cylindrospermopsin and Anatoxin-a in Cyanotoxins in Drinking Water— Cylindrospermopsin. EPA-820R15103. Drinking Water by Liquid Frequently Asked Questions. OHA, Salem, USEPA, Washington. Chromatography Electrospray Ionization Ore. https://public.health.oregon.gov/ Tandem Mass Spectrometry (LC/ESI-MS/ HealthyEnvironments/DrinkingWater/ USEPA, 2015d. Recommendations for Public MS). EPA 815-R-15-009. USEPA, Monitoring/HealthEffects/Pages/ Water Systems to Manage Cyanotoxins in Washington. www.epa.gov/ cyanotoxins.aspx (accessed Jan. 11, 2016). Drinking Water. EPA 815-R-15-010. dwanalyticalmethods/method-545- USEPA, Washington. www.epa.gov/sites/ determination-cylindrospermopsin-and-O’Reilly, C.M.; Sharma, S.; Gray, D.K.; production/files/2015-06/documents/ anatoxin-drinking-water-liquid Hampton, S.E.; Read, J.S.; Rowley, R.J.; cyanotoxin-management-drinking-water. (accessed Sept. 2, 2016). Schneider, P., et al., 2015. Rapid and pdf (accessed Sept. 2, 2016). Highly Variable Warming of Lake Surface USEPA, 2015h. Revisions to the Unregulated Waters Around the Globe. Geophysical USEPA, 2015e. Algal Toxin Risk Assessment Contaminant Monitoring Rule (UCMR 4) for Research Letters, 42:24:10,773. https:// and Management Strategic Plan for Public Water Systems and Announcement doi.org/10.1002/2015GL066235. Drinking Water. 810R04003. USEPA, of a Public Meeting. Federal Register. Washington. www.epa.gov/sites/ www.federalregister.gov/articles/2015/12/Paerl, H.W. & Paul, V.J., 2012. Climate production/files/2015-11/documents/ 11/2015-30824/revisions-to-the- Change: Links to Global Expansion of algal-risk-assessment-strategic- unregulated-contaminant-monitoring-rule- Harmful Cyanobacteria. Water plan-2015.pdf (accessed Sept. 2, 2016). ucmr-4-for-public-water-systems-and#t-4 Research, 46:5:1349. https://doi. (accessed Sept. 2, 2016). org/10.1016/j.watres.2011.08.002. USEPA, 2015f. Method 544: Determination of Microcystins and Nodularin in Yuan, L.L.; Pollard, A.I.; Pather, S.; Oliver, J.L.;USEPA (US Environmental Protection Agency), Drinking Water by Solid Phase & D’Anglada, L., 2014. Managing 2015a. Health Effects Support Document Extraction and Liquid Microcystin: Identifying National‐Scale for the Cyanobacterial Toxin Anatoxin-A. Chromatography/Tandem Mass Thresholds for Total Nitrogen and EPA- 820R15104. USEPA, Washington. Spectrometry (LC/MS/MS). Chlorophyll a. Freshwater Biology, EPA/600/R-14/474. USEPA, 59:9:1970. https://doi.org/10.1111/USEPA, 2015b. Drinking Water Health Washington. https://cfpub.epa.gov/si/ fwb.12400. Advisory for the Cyanobacterial si_public_record_report.cfm? Everlube R-75 is now certified to NSF/ANSI 61Everlube R-75—a PTFE, Solid Film Lubricant—is the ideal coating for manufacturers that make products that come into contact with drinking water. This versatile coating offers these benefits: • Excellent wear life • Very good release properties • Very good thermal stability • Ideal for lighter load carrying applications • RoHS compliant Contact Information:Call 1-800-428-7802, visit www.everlubeproducts.com or type “Everlube R-75” in your browser for technical data sheet. HENRIE ET AL. | 109:2 • FEBRUARY 2017 | JOURNAL AWWA 47

AWWA HAS THE SMALL SYSTEMSRESOURCES YOU NEEDcrop marksM54: DEVELOPING RATES FOR AWWA SMALL SYSTEMS SMALL SYSTEMS RESOURCESMALL SYSTEMS FIELD GUIDE COMMUNITYHere is the essential information This field guide may be the The Small Systems Resourceyou need to determine your revenue handiest book ever written for Community links you torequirements, analyze your rates for water professionals. It provides information on certificationdifferent customer classes, develop information and data needed for and operator training, ratesa financial plan, and design a better operating small drinking water and planning, understanding USEPArate structure, even if you have wastewater treatment facilities. regulations, and much more,limited resources and data. including fact sheets on SMCLs AWWA Catalog No. 20746 and ammonia.AWWA Catalog No. 30054-2E AWWA member/nonmember $44$56 AWWA member$88 nonmember www.awwa.org/smallsystems