PROJECT NO. 4719

Biofiltration Guidance Manual for Drinking Water Facilities

Biofiltration Guidance Manual for Drinking Water Facilities

Prepared by: Jess Brown, Giridhar Upadhyaya, Jennifer Nyfennegger, Greg Pope, and Stetson Bassett Carollo Engineers, Inc.

Ashley Evans, Jason Carter, and Victoria Nystrom Arcadis U.S., Inc.

Samantha Black, Christina Alito, and Chance Lauderdale HDR, Inc.

Jennifer Hooper, Benjamin Finnegan, and Laurel Strom CDM Smith

Lynn Williams Stephens and Emily Palmer Brown and Caldwell

Eric Dickenson, Stephanie Riley, and Eric Wert Southern Nevada Water Authority

Lauren Weinrich and Peter Keenan American Water

2020

The Water Research Foundation (WRF) is a nonprofit (501c3) organization which provides a unified source for One Water research and a strong presence in relationships with partner organizations, government and regulatory agencies, and Congress. The foundation conducts research in all areas of drinking water, wastewater, stormwater, and water reuse. The Water Research Foundation’s research portfolio is valued at over $700 million. The Foundation plays an important role in the translation and dissemination of applied research, technology demonstration, and education, through creation of research‐based educational tools and technology exchange opportunities. WRF serves as a leader and rmodel fo collaboration across the water industry and its materials are used to inform policymakers and the public on the science, economic value, and environmental benefits of using and recovering resources found in water, as well as the feasibility of implementing new technologies. For more information, contact: The Water Research Foundation 1199 North Fairfax Street, Suite 900 6666 West Quincy Avenue Alexandria, VA 22314‐1445 Denver, Colorado 80235‐3098 www.waterrf.org P 571.384.2100 P 303.347.6100 [email protected] ©Copyright 2020 by The Water Research Foundation. All rights reserved. Permission to copy must be obtained from The Water Research Foundation. WRF ISBN: 978‐1‐60573‐516‐0 WRF Project Number: 4719 This report was prepared by the organization(s) named below as an account of work sponsored by The Water Research Foundation. Neither The Water Research Foundation, members of The Water Research Foundation, the organization(s) named below, nor any person acting on their behalf: (a) makes any warranty, express or implied, with respect to the use of any information, apparatus, method, or process disclosed in this report or that such use may not infringe on privately owned rights; or (b) assumes any liabilities with respect to the use of, or for damages resulting from the use of, any information, apparatus, method, or process disclosed in this report. Carollo Engineers, Inc., Arcadis U.S., Inc., HDR, Inc., CDM Smith, Brown and Caldwell, Southern Nevada Water Authority, American Water This document was reviewed by a panel of independent experts selected by The Water Research Foundation. Mention of trade names or commercial products or services does not constitute endorsement or recommendations for use. Similarly, omission of products or trade names indicates nothing concerning The Water Research Foundation's positions regarding product effectiveness or applicability.

ii The Water Research Foundation Acknowledgments

Research Team Principal Investigators: Jess Brown, Ph.D., P.E. Carollo Engineers, Inc.

Co‐Principal Investigators: Giridhar Upadhyaya, Ph.D., P.E. Carollo Engineers, Inc. Jennifer Nyfennegger, Ph.D., P.E. Carollo Engineers, Inc. Greg Pope, Ph.D., P.E. Carollo Engineers, Inc. Ashley N. Evans, P.E. Arcadis U.S., Inc. Jason T. Carter, P.E. Arcadis U.S., Inc. Samantha Black, Ph.D., P.E. HDR, Inc. Christina Alito, Ph.D., P.E. HDR, Inc. Chance Lauderdale, Ph.D., P.E. HDR, Inc. Jennifer Hooper, P.E., MS CDM Smith Lynn Williams Stephens, P.E. Brown and Caldwell Eric Dickenson, Ph.D. Southern Nevada Water Authority Eric Wert, Ph.D., P.E. Southern Nevada Water Authority Lauren Weinrich, Ph.D. American Water Project Team: Stetson Bassett Carollo Engineers, Inc. Patrick Carlson, P.E. Carollo Engineers, Inc.

Biofiltration Guidance Manual for Drinking Water Facilities iii Benjamin Finnegan, P.E., BCEE, MS CDM Smith Vincent Hart, P.E. Carollo Engineers, Inc. Peter Keenan, P.E. American Water Kelsey Kenel, EIT HDR, Inc. Victoria E. Nystrom, EIT Arcadis Emily Palmer, EIT Brown and Caldwell Stephanie Riley, Ph.D. Southern Nevada Water Authority Laurel Strom, MS CDM Smith WRF Project Subcommittee or Other Contributors Technical Advisory Committee Mark LeChevallier, Ph.D., Dr. Water Consulting, LLC Leigh Terry, Ph.D., University of Alabama Benay Akyon, Ph.D., Xylem Inc. Alan Roberson, P.E., Association of State Drinking Water Administrators Utility Advisory Subcommittee Gerard Yates, Mike Rau, Central Utah Water Conservancy District Chaise Holmgren, Dallas Water Utilities Ying Hong, Ph.D., P.E., Greater Cincinnati Water Works Randall Emory, P.E., Anthony Whitehead, Greenville Utilities Commission Denise Funk, Collin Hubbs, Gwinnett County (Denise was formerly with Gwinnett County) Wendy Krkosek, Ph.D., Halifax Water Sun Liang, Ph.D., P.E., Metropolitan Water District of Southern California Abhay Tadwalker, P.Eng., formerly with Toronto Water Kevin Linder, Aurora Water Regulatory Advisory Subcommittee Brian Noma, P.E., Minnesota Department of Health Kurt Souza, P.E., California Water Boards Academic Advisory Subcommittee Gary Amy, Ph.D., Clemson University Graham Gagnon, Ph.D., P.Eng., Dalhousie University Peter Huck, Ph.D., P. Eng., Peter Huck & Associates Mary Jo Kirisits, Ph.D., University of Texas‐Austin Lutgarde Raskin, Ph.D., University of Michigan Bruce Rittmann, Ph.D., NAE, FAEESP, Arizona State University Scott Summers, Ph.D., University of Colorado‐Boulder

iv The Water Research Foundation Manufacturer Advisory Subcommittee Vadim Malkov, Ph.D., Carlos Williams, Hach Co. Dave Tracey, P.Eng., LuminUltra Technologies, Ltd. Pam London, Melanie Lasch, Veolia North America (Pam was formerly with Veolia)

Project Advisory Committee Eva Nieminski, Ph.D., with the Utah Department of Environmental Quality during this project Christine Owen, Ph.D., Hazen & Sawyer Robert Jurenka, U.S. Bureau of Reclamation Zaid Chowdhury, Ph.D., P.E., Garver

Water Research Foundation Staff John Albert, MPA Chief Research Officer Grace Jang, Ph.D. Research Program Manager

Biofiltration Guidance Manual for Drinking Water Facilities v Abstract and Benefits

Abstract:

Unlike conventional granular media filters which only remove particles, biological filters serve the dual purpose of removing particulates and labile compounds through biodegradation or biotransformation mechanisms. Over the last 15 years, considerable research has been done to advance the science and engineering of surface water biofiltration, focusing on monitoring, tracking, and control strategies, enhancing and engineering biofiltration, leveraging upstream processes, and removing multiple contaminants simultaneously. This project combined that research with extensive biofilter design and operating experience across the water industry to develop a consolidated set of guidelines for the design, operation, maintenance, and monitoring of biologically active rapid‐rate gravity filters; guidelines intended to benefit existing biofiltration plants, filtration plants that intend to convert to biofiltration, and future greenfield biofiltration plants. In short, this project produced the definitive resource for biofiltration design and operation, which will help utilities leverage intentional biofiltration, mitigate unintended consequences, and improve overall biofilter performance.

Benefits:

In addition to providing practical and comprehensive guidance for operators, engineers, regulators, manufacturers, and researchers, the guidance manual also provides access to multiple biofiltration‐ related tools, including:

• Biofiltration Terminology • Frequently Asked Questions • Biofiltration Calculations • Operations Checklist • Troubleshooting Guide • Monitoring Tool Standard Operating Procedures • Biofiltration Tools Compendium • List of Biofilter Optimization Case Studies • Biofilter Optimization Decision Trees • List of Biofilter Conversion Case Studies • List of Select Full‐Scale Biofiltration Plants with Drivers and Key Parameters • Sample Biofilter Testing Plans

Keywords: Biofiltration, biofilter, greenfield biofiltration, biofilter conversion, biofilter optimization.

vi The Water Research Foundation Contents

Acknowledgments ...... iii Abstract and Benefits ...... vi Tables ...... x Figures ...... xi Acronyms and Abbreviations ...... xii Executive Summary ...... xv Chapter 1: Background ...... 1 1.1 History ...... 1 1.2 Treatment Objectives ...... 3 1.3 Removal Mechanisms ...... 3 1.4 Biofilm Characteristics and Dynamics ...... 5 1.5 Biofilter Performance Drivers ...... 7 Chapter 2: Monitoring Tools and Instrumentation ...... 9 2.1 Filter Integrity Monitoring ...... 9 2.1.1 Media Inspection ...... 9 2.1.2 Underdrain Inspection ...... 15 2.2 Hydraulic Monitoring ...... 16 2.2.1 Filtration Rate ...... 16 2.2.2 Headloss ...... 16 2.2.3 Filter Run Time and Unit Filter Run Volume ...... 18 2.2.4 Empty Bed Contact Time ...... 19 2.2.5 Backwash Pressure Monitoring ...... 20 2.3 Water Quality Monitoring ...... 20 2.3.1 General Water Quality (Turbidity, Temperature and pH) ...... 20 2.3.2 Biodegradable Organic Matter (BOM) ...... 21 2.3.3 Inorganic Compounds ...... 25 2.3.4 Nutrients ...... 25 2.4 Biological Monitoring ...... 27 2.4.1 Biofilm Formation Rate ...... 27 2.4.2 Dissolved Oxygen Consumption ...... 29 2.4.3 Adenosine Triphosphate (ATP) ...... 31 2.4.4 Extracellular Polymeric Substances (EPS) ...... 32 2.4.5 Microbial Community Analysis ...... 34 2.4.6 Other Biological Monitoring Tools ...... 35 2.5 Recommended Monitoring Tools ...... 36 2.6 Developing a Monitoring Strategy ...... 39 2.6.1 Routine Operation ...... 39 2.6.2 Startup and Troubleshooting ...... 40 2.6.3 Special Studies, Optimization, and Research ...... 41 2.7 Data Management ...... 41

Chapter 3: Optimizing Existing Biofiltration Plants ...... 43 3.1 Planning ...... 43 3.1.1 Suitability ...... 43 3.1.2 Testing ...... 44 3.2 Optimization Strategies ...... 45 3.2.1 Media Selection ...... 45 3.2.2 Backwash Protocol ...... 47 3.2.3 Nutrient Augmentation ...... 48 3.2.4 Pre‐Oxidation ...... 50 3.2.5 pH Adjustment ...... 53 3.2.6 Holistic Optimization ...... 54 Chapter 4: Converting Conventional Filters to Biofilters ...... 55 4.1 Planning ...... 55 4.1.1 Suitability ...... 55 4.1.2 Testing ...... 57 4.2 Biofilter Conversion Strategies ...... 57 4.2.1 Decreasing Chlorine Dose ...... 57 4.2.2 Addressing Pre‐Loaded Manganese ...... 58 4.2.3 Upgrading Backwash Capabilities ...... 58 4.2.4 Installing Preoxidation ...... 59 4.2.5 Modifying Filter Design ...... 59 Chapter 5: Greenfield Biofiltration ...... 61 5.1 Planning ...... 61 5.1.1 Suitability ...... 61 5.1.2 Testing ...... 61 5.2 Biofiltration Design ...... 61 5.2.1 Pre‐Treatment ...... 61 5.2.2 Filter Underdrains ...... 63 5.2.3 Trough Height Above Media ...... 66 5.2.4 Media Type and Configuration ...... 66 5.2.5 Empty Bed Contact Time and Filter Loading Rate ...... 67 5.2.6 Chemical Feeds ...... 67 5.2.7 Backwash System ...... 68 5.2.8 Chlorine and Oxidant Addition ...... 68 5.2.9 Residuals Handling ...... 69 5.2.10 Hydraulics ...... 69 Chapter 6: Operation and Maintenance ...... 71 6.1 Monitoring ...... 71 6.2 Start‐Up and Acclimation ...... 71 6.3 Steady‐State Operation ...... 71 6.4 Shutdown, Idling, and Restart ...... 72 Chapter 7: Biofiltration Testing...... 75 7.1 Defining Testing Objectives ...... 75 7.2 Benchmarking Water Quality and Treatment Characteristics ...... 76 7.2.1 Historical Water Quality and Performance ...... 76 7.2.2 Estimating Potential Biological Organic Carbon Removal ...... 79 7.2.3 Filter Design and Operation ...... 79

viii The Water Research Foundation 7.3 Selecting Testing Scale(s) ...... 80 7.4 Designing a Desktop Evaluation ...... 83 7.4.1 Literature Review ...... 83 7.4.2 Modeling ...... 84 7.5 Designing Bench, Pilot or Demonstration Tests ...... 84 7.5.1 Duration ...... 84 7.5.2 Design ...... 85 7.5.3 Testing Conditions ...... 88 7.5.4 Equipment and Instrumentation ...... 90 7.5.5 Basic Monitoring ...... 92 7.5.6 Data Management and Interpretation ...... 93 7.5.7 Staffing Options ...... 94 7.5.8 Quality Control ...... 95 7.5.9 Safety ...... 96 7.6 Overcoming Common Testing Challenges ...... 97 7.7 Understanding Expected Outcomes ...... 98 7.8 Resource Planning ...... 102

Appendix A ...... 105 Appendix B ...... 113 Appendix C ...... 119 Appendix D ...... 121 Appendix E ...... 123 Appendix F ...... 129 Appendix G ...... 137 Appendix H ...... 139 Appendix I ...... 141 Appendix J ...... 147 Appendix K ...... 149 Appendix L ...... 151 References ...... 163

Biofiltration Guidance Manual for Drinking Water Facilities ix Tables

1‐1 Contaminant Drivers and Total Removal Potential by Rapid‐Rate Biofiltration With and Without Pre‐Ozonation ...... 3 1‐2 Biofilter Performance Drivers and their Impact ...... 8 2‐1 Filter Media Integrity Monitoring Techniques ...... 11 2‐2 Hydraulic Parameter Summary ...... 16 2‐3 Factors that Impact Biofilter Headloss ...... 17 2‐4 Typical Filter Influent Ranges of General Water Quality Parameters ...... 21 2‐5 Typical Influent Concentrations and Removal Ranges of Different Components of NOM ...... 24 2‐6 Typical Influent Concentrations and Removal Ranges of Nutrients ...... 26 2‐7 Other Biological Monitoring Tools Not Recommended for Regular Monitoring ...... 36 2‐8 Summary of Monitoring Tools ...... 37 2‐9 Summary of Monitoring Tool Quality Assurance and Quality Control (QA/QC) and Relative Cost ..... 38 2‐10 Recommended Biofilter Monitoring During Routine Operations ...... 39 2‐11 Additional Recommended Biofilter Monitoring During Startup and Troubleshooting ...... 40 2‐12 Additional Recommended Biofilter Monitoring During Special Studies, Optimization, and Research ...... 41 3‐1 Optimization Strategies ...... 44 3‐2 Media Selection Characteristics ...... 46 3‐3 Parameters for Tracking Backwash Effectiveness ...... 47 3‐4 Summary of C:N:P Ratios Tested ...... 49 3‐5 Summary of Pre‐Oxidant Doses Tested ...... 51 3‐6 Summary of pH Increase Tested in Previous WRF Studies ...... 53 3‐7 Optimal pH Ranges for Ammonia, Nitrite, and Manganese Oxidizers ...... 53 5‐1 Chemicals Added in Biofiltration Applications ...... 62 5‐2 Common Biofilter Media Design ...... 66 5‐3 Example Backwash Protocol ...... 68 7‐1 Recommended Historical Water Quality and Operational Data ...... 77 7‐2 Description, Objectives and Limitations of Each Testing Scale ...... 80 7‐3 Testing Scale Selection Based Upon Common Testing Objectives ...... 83 7‐4 Comparison of Typical Testing Durations for Different Testing Scales ...... 84 7‐5 Comparison of Filter Design for Different Testing Scales ...... 85 7‐6 Comparison of Media Design for Different Testing Scales ...... 86 7‐7 Comparison of Backwash Design for Different Testing Scales ...... 87 7‐8 Comparison of Testing Conditions for Different Testing Scales ...... 88 7‐9 Challenge Testing Strategies ...... 89 7‐10 Comparison of Monitoring and Automation for Different Testing Scales ...... 91 7‐11 Testing Equipment Procurement Considerations ...... 92 7‐12 Recommended Basic Monitoring Plan for Biofiltration Testing ...... 93 7‐13 Comparison of Options for Staffing ...... 94 7‐14 Common Quality Control Measures ...... 95 7‐15 Common Safety Considerations ...... 96 7‐16 Overcoming Common Biofiltration Testing Challenges ...... 98 7‐17 Expected Outcomes of Each Testing Scale ...... 99 7‐18 Resource Planning for Each Scale of Testing ...... 102

x The Water Research Foundation Figures

1‐1 Timeline Detailing Progression in the Development, Implementation, and Research of Biofiltration ...... 2 1‐3 Effluent Concentration as a Function of Time Comparing Removal Trends Using an Inert Filter Media (e.g., Sand, Anthracite) and Fresh Adsorptive Media (e.g., GAC) from Start‐Up ...... 4 1‐3 Illustration of Contaminant Removal Mechanisms by Biological Filtration with a Porous Media ...... 5 1‐4 Typical Biological Characteristics of a Conventional Anthracite Filter with an Oxidant Residual and Anthracite/GAC Biological Filter ...... 6 2‐1 Example Alternative Methods Used for Collecting Filter Media Profile Samples, Including Use of a Check Valve on the Bottom of the Corer (a) and (b), Use of an Open Pipe with Handles on the Top (c), a Box Corer (d), and an Example of a Grain Thief Sampling Device (e) ...... 12 2‐2 Typical Solids Retention Curves in a Filter Bed Before and After a Backwash ...... 12 2‐3 Filter Media Replacement Decision Matrix ...... 14 2‐4 Pilot Biofilter Underdrain Caps Operated Under (a) Nutrient‐Limited and (b) Phosphorus‐ Enhanced Conditions ...... 15 2‐5 Microbial Biomass Macro and Micro Nutrient Composition, Based on E. coli...... 26 2‐6 Biofilm Attachment and Growth on Coupons in a Pipe Loop ...... 28 2‐7 Biofilm Formation Rate Pipe Loop Example (a) installation, (b) Coupon Harvesting and (c) ATP Analysis ...... 28 2‐8 Biofilm Formation Rate Pipe Loop Example Schematic ...... 29 2‐9 DO Consumption by Bacteria During Heterotrophic Cellular Respiration ...... 30 2‐10 DO Probe Installation Methods as a (a) Flow Through Cell or (b) Stilling Well ...... 31 2‐11 ATP Test Kit Protocol for Filter Media Samples ...... 32 2‐12 Scanning Electron Micrograph on a Sand (a) and BAC (b) Biofilter at Greater Cincinnati Water Works ...... 32 2‐13 EPS Formation and Interference with Particle Collection in Biofilter Media...... 33 3‐1 Demonstration of Holistic Optimization of Coagulant Dose...... 54 4‐1 Tasks to be Completed During Conversion Planning ...... 56 4‐2 A WTP in Texas Converted to Biofiltration in 2001 and Observed Rapid Decreases in UFRVs ...... 60 5‐1 Example of a Typical Gravel Underdrain System ...... 64 5‐2 Example of a Block Underdrain System ...... 64 5‐3 Example of a Nozzle Underdrain System ...... 65 5‐4 Cleaned (left) and Clogged (right) Underdrain Cap ...... 65 5‐5 Failed Porous Plate Underdrains: Stripped Anchors (left) and Blown Mastic (right) ...... 66 5‐6 Process Flow Diagram of Backwash System with Unchlorinated and Chlorinated Backwash Capability ...... 68 7‐1 Key Testing Plan Questions Addressed in This Chapter ...... 75 7‐2 Common Biofiltration Testing Objectives ...... 76 7‐3 Data Analysis Tools ...... 78 7‐4 Photos of Bench, Pilot, and Demonstration Testing Facilities ...... 82 7‐5 Next Steps After Evaluating Testing Outcomes ...... 100

Biofiltration Guidance Manual for Drinking Water Facilities xi Acronyms and Abbreviations

°C Degree Celsius AASG Average apparent specific gravity AD Apparent density ADP Adenosine diphosphate AMP Adenosine monophosphate AN Abrasion number AOC Assimilable organic carbon APHA American Public Health Association AS Acid solubility ATP Adenosine triphosphate AWWA American Water Works Association BAC Biological active carbon BCA Pierce bicinchoninic acid BDOC Biodegradable dissolved organic carbon BOM Biodegradable organic matter BSA Bovine serum albumin cm Centimeter CT Contact time DBPs Disinfection byproducts DBP‐FP DBP formation potential DNA Deoxyribonucleic acid DO Dissolved oxygen DOC Dissolved organic carbon EBCT Empty bed contact time EDC Endocrine disrupting compound EDTA Ethylenediaminetetraacetic acid EPA Environmental Protection Agency EPS Extracellular polymeric substances ES Effective size ft Feet/foot FRT Filter run time GAC Granular activated carbon gal/ft2 Gallons per square feet gpm/ft2 Gallons per minute per square feet HAAs Haloacetic acids HPCs Heterotrophic plate counts HPLC High performance liquid chromatography IBM Integrated Biofilm Model IC Ion chromatography

xii The Water Research Foundation ICP Inductively coupled plasma IESWTR Interim Enhanced Surface Water Treatment Rule IN Iodine number lbs/ft3 Pounds per cubed feet L Liter L/min Liters per minute LDO Luminescent dissolved oxygen LTESWTR Long‐Term Enhanced Surface Water Treatment Rule LT2ESWTR Long‐Term 2 Enhanced Surface Water Treatment Rule m Meter µg‐C/L Micrograms per liter µg‐C/L Micrograms of carbon per liter mg Milligrams mg/L Milligrams per liter MH Moh’s Hardness MIB Methylisoborneol mm Millimeter mM Millimolar MS Mesh size mV Millivolt NDMA N‐Nitrosodimethylamine NEWPP Northeast Water Purification Plant nm Nanometer NOM Natural organic matter NTU Nephelometric turbidity units ORP Oxidation‐reduction potential PLFA Phospholipid fatty acids ppb Parts per billion PPCP Pharmaceutical and personal care product PSW Partnership for Safe Water qPCR Quantitative polymerase chain reaction RLU Relative luminescence units SCADA Supervisory Control and Data Acquisition SDWA Safe Drinking Water Act SG Specific gravity SM Standard methods SOP Standard operating procedure SU Standard unit SUVA Specific ultraviolet absorbance SWTR Surface Water Treatment Rule T&O Taste and odor TCEP Tris‐2‐chloroethyl phosphate

Biofiltration Guidance Manual for Drinking Water Facilities xiii THMs Trihalomethanes TOC Total organic carbon TOrCs Trace organic compounds TSMSBM Transient‐State, Multiple‐Species Biofilm Model UC Uniformity coefficient UFRV Unit filter run volume U.S. United States UV Ultraviolet

UV254 Ultraviolet absorbance 254 nanometers WRF Water Research Foundation WTP Water treatment plant

xiv The Water Research Foundation Executive Summary

Biofiltration Guidance Manual Overview Granular media filtration is used in most surface water treatment plants (WTPs) throughout the U.S. as a key element of a multi‐barrier approach for removal of pathogens and other contaminants from drinking water supplies. The primary focus of granular media filtration is removal of fine particulate matter that is note otherwis removed in an upstream clarification or other pre‐treatment step. Filterable particulate matter can include bacteria, protozoa, and other microorganisms, as well as natural organic matter (NOM), naturally occurring silts and precipitated metals, and flocculated chemical coagulants. Historically, a common practice was to feed chlorine and/or other pre‐oxidants upstream of filtration, as it was found to improve particle removal effectiveness, and it also provided disinfection credits. However, chlorine can also react with NOM to form harmful disinfection byproducts (DBPs), so many water utilities have sought to postpone chlorine disinfection to later in the treatment process after the bulk ofM NO has been removed. Without a continuous feed of pre‐chlorine, granular media filters will inevitably become biologically active to some degree. In recent years, water utilities have begun to recognize that biologically active granular media filtration, which is also referred to as biofiltration, can offer several valuable benefits, including: • Removal of common cyanobacterial metabolites like 2‐methylisoborneol (MIB) and geosmin, which can lead to objectionable taste and odor (T&O) characteristics in drinking water. • Improving the biological stability of finished water. • Removal of assimilable organic carbon (AOC) and/or biodegradable dissolved organic carbon (BDOC), which can otherwise contribute to biofilm formation, bacteriological regrowth, loss of disinfectant residuals, and even increased corrosion within drinking water distribution systems. • Removal of DBP precursors and some biodegradable DBPs if preformed within the treatment plant or present in the source water. As a result, biofiltration is increasingly being used throughout the U.S. to enhance drinking WTP performance and improve finished water quality. However, biologically active granular media filters can create operational or performance challenges for water treatment facilities if physical or operational conditions are not optimized. The purpose of this Biofiltration Guidance Manual is to provide engineering and operations personnel with an understanding of critical design and operational parameters that should be considered to improve the effectiveness and maximize the reliability of biofiltration processes. Common questions about biofiltration in this section will orient the reader to the information within the individual chapters of this Biofiltration Guidance Manual. Answers to these and other pertinent questions are provided herein as well as in the Appendix B: Frequently Asked Questions section. What is the development, implementation, and research history of biofiltration? Biological filtration could arguably be described as incidental and there was relatively slow implementation until water quality regulations became more stringent and the benefits of biofiltration were recognized. In 2012, the need for information and guidance of biofilter implementation and

Biofiltration Guidance Manual for Drinking Water Facilities xv operation led to the Water Research Foundation’s (WRF’s) defined focus area Biofiltration: Defining Benefits and Developing Utility Guidance in 2012. The Biofiltration Guidance Manual begins in Chapter 1: Background by providing background information on the history of biofiltration and the mechanics of granular media filtration and microbiological processes. An overview of the various treatment objectives and removal mechanisms are described in order to answer the questions related to the removal potential that biofiltration provides for improving water quality such as DBP precursors, T&O compounds, inorganics, and other micropollutants. The first chapter also addresses performance drivers related to the way that preoxidation and biomass development impact water quality passing through biofilters compared to conventional filters. Biofiltration involves a complex interaction of physics, chemistry, and microbiology that is currently understood mostly on a macroscopic basis, but can still be monitored and controlled to achieve predictable and reliable performance. Because biofiltration relies on naturally occurring microbial species, biofiltration processes are influenced to varying degrees by source water quality and biota, as well as the physical and chemical conditions that exist within a WTP. Furthermore, the dynamics of a heterogeneous microbial community affect and are affected by the physical and chemical process mechanisms that occur within granular media filters. For example, shear forces within granular media filters cause detachment of microbial biofilms, which is an essential function to allow for effective mass transport of trace organic molecules into biofilms and minimizes plugging of the hydraulic pathways through the media and underdrain. The presence of chlorine in washwater can also be an important tool to combat biofouling of filter underdrains but will also have an impact on the make‐up and viability of the microbial community in the biofilter. The beginning of Chapter 2: Monitoring Tools & Instrumentation orients the reader with the guidance needed for inspecting biofilter media beds and underdrains as well and hydraulic performance parameters for assessing conditions, ensuring performance, and reducing the potential for underdrain failure. The Biofiltration Guidance Manual also provides a valuable overview of how microbial growth and activity are affected by the various physical and chemical conditions that exist within the biofilter and wider WTP. For example, the availability of macronutrients like nitrogen and phosphorus are critical to the growth of microbial populations in general, including within biofilters. The Biofiltration Guidance Manual describes how water treatment processes can change the availability of such macronutrients, as well as how adding supplemental macronutrients can be beneficial for establishing and sustaining a healthy microbial community within a biofilter. What monitoring tools are used to assess the performance of the biofilters? What is similar or different to conventional filter monitoring? Process monitoring and control methods that should be used routinely to monitor the “health” of biofilters on both an ongoing and periodic basis are described in Chapter 2: Monitoring Tools & Instrumentation. A detailed review of available monitoring and control techniques and technologies is provided, including those that are recommended for routine use, as well as to aid troubleshooting or more detailed process performance assessments. Techniques for inspecting biofilter media and guidance on the frequency of inspection and replacement are discussed. A decision matrix that incorporates aspects related to these assessments using media analysis and water quality targets is provided. There is a section dedicated to process monitoring that will be valuable to engineers to validate that plants are designed and constructed with sufficient monitoring and control devices in place to ensure operability and reliable performance. Similarly, monitoring guidance will be valuable to operations and maintenance staff because it explains the purpose and importance of using and

xvi The Water Research Foundation maintaining various process monitoring devices and methods to achieve consistent high‐quality process performance. In addition, useful operational tools and techniques that can be used either continually or periodically to ensure efficient and high‐quality filter performance on a reliable and continuous basis are presented. Hydraulic parameters are described with typical values that can be tracked and trended for proactive monitoring and avoidance of adverse operational issues. Guidance and decision matrices are useful resources that can be put in place in the treatment plants filter monitoring program. Chapter 7: Biofiltration Testing provides detailed guidance for a well‐designed testing plan that begins with defining the testing objectives, including experimental design to meet those objectives, determining how much data is necessary to meet those objectives and when the effort is completed, and planning for the necessary resources. What optimization strategies should be defined and implemented to attain reliable, sustained, and robust treatment with biofiltration? Chapter 3: Optimizing Existing Biofiltration Plants provides resources for projects targeted for improving biofiltration and associated planning efforts. Optimization planning is designed to deliver improvements in overall treatment efficiency and water quality. The approach covers clearly defined objectives and associated strategies with a suggested duration to achieve the testing objectives. Examples and strategies for questions such as “How do I manage and identify synergistic effects, for example pH adjustment and coagulation” are explored. Valuable information is provided in this Biofiltration Guidance Manual with respect to the advantages, disadvantages, and effectiveness of sand, anthracite, and granular activated carbon (GAC), which are the three most common types of media currently being used in granular media filters and biofilters. Each type is capable of supporting a robust microbial community, but the selection of media type should consider source and finished water quality objectives, as well as other practical factors like bed depth and condition of existing filter media when adding or converting to biofiltration. Optimization strategies including nutrient augmentation, pre‐ oxidant addition, and pH adjustment are also discussed. The addition of nutrients can promote biological activity, but careful selection of nutrient doses should be tested prior to system‐wide implementation, as benefits can be site‐specific. Pre‐oxidant addition can be used to push biology deeper into the filter bed and remove excess biomass. Optimal pH levels, like nutrient augmentation, can promote biological activity. The Biofiltration Guidance Manual includes a detailed discussion of numerous physical features and process elements that should be considered in the design of new biofiltration systems or conversion of existing granular media filters to biofilters (Chapter 4: Converting Conventional Filters to Biofilters and Chapter 5: Greenfield Biofiltration). When a utility decides to convert conventional filters to biofilters, they have recognized that biofiltration will offer various benefits related to improved T&O, biological stability of finished water, and DBP precursors. These and other benefits are valuable in providing excellent water quality to the customer. Given these and other advantages, planning is critical to determining an effective conversion strategy from careful considerations. The planning section for the benefits, challenges, opportunities and concerns when considering conversion are found in Chapter 4: Converting Conventional Filters to Biofilters. The Conversion Assessment Tool is available as a Microsoft Excel‐based program to assist with focusing on what information is most useful for determining the suitability of the plant to convert to biofiltration. This can be found in Appendix G: Tools Compendium along with other useful tools including Excel‐based computer programs, numerical models, standards, databases and manuals available to utilities that are currently operating or considering biofiltration. Such information will be extremely useful to help engineers and operators incorporate features that are important for maximizing the effectiveness of a biofiltration facility while minimizing risks and

Biofiltration Guidance Manual for Drinking Water Facilities xvii operational difficulties. For example, substantial information is provided with regards to filter underdrain design, because underdrain failures have been attributed to microbial fouling in a few notable biofiltration installations. The Biofiltration Guidance Manual emphasizes the importance of monitoring clean‐bed headloss and washwater‐system pressures on a regular basis for early signs of problems (Chapter 2: Monitoring Tools and Instrumentation and Chapter 5: Greenfield Biofiltration). Operational tools and techniques are also discussed to help operators mitigate the impacts of biofouling that can otherwise lead to possible declines in capacity, deterioration of performance, or expensive physical damage to critical plant infrastructure. One of the more challenging aspects of biofiltration as compared to conventional granular media filtration with pre‐chlorination is sustaining a high degree of manganese removal through the filtration process. Manganese is a naturally occurring mineral element that may be present in source waters at varying levels or may be added in treatment facilities if potassium or sodium permanganate is being used as a pre‐oxidant. When chlorine is applied upstream of granular media anthracite or sand filters, manganese removal is typically highly effective and reliable. However, manganese removal can become more difficult when using GAC as a filter media or eliminating pre‐chlorine to foster a biofiltration regime. The Biofiltration Guidance Manual offers a good explanation of how biofilters can be monitored and operated to ensure continuous and reliable removal of manganese and other particulate matter within the biofiltration process throughout Chapters 2 through 7. Other valuable components of the Biofiltration Guidance Manual are the data and information provided from numerous case study facilities that have had success and difficulties with transitioning from conventional granular media filtration or starting up a new biofiltration process (Chapter 4: Converting Conventional Filters to Biofilters and Chapter 5: Greenfield Biofiltrations). Such information can help utilities avoid errors or omissions in the design and operation of biofiltration facilities. In many ways, learning about the challenges and difficulties that other utilities have experienced is often the most valuable information that can be gleaned from case studies, because it can help identify pitfalls that may not be obvious when simply reviewing design criteria and performance data from successfully operating facilities. The Biofiltration Guidance Manual also includes 12 appendices comprised of supplemental and stand‐ alone reference material, including: • Appendix A – Biofiltration Terminology with commonly used language terms for describing various scientific and engineering aspects related to this topic. • Appendix B – Frequently Asked Questions and short answers to inquiries from utility personnel during various points when considering or implementing biofiltration through planning, evaluation, design, and operation phases. • Appendix C – Biofiltration Calculations provided in this section summarize equations for data needed in calculating media uniformity coefficient (UC) and size, filter hydraulics and design, water quality by specific ultraviolet absorbance (SUVA), and biological data from adenosine triphosphate (ATP) monitoring for biofilm formation rates and biofilteractivity on the media using ATP. • Appendix D – Biofilter Operations Checklist for monitoring using online monitoring or grab sampling across the pertinent water quality, biological, and operational characteristics with specific parameters, locations to be monitored, and for what purpose based on daily/weekly, monthly, annually and long‐term timelines. • Appendix E – Biofilter Troubleshooting Guide summarizes the indicators, causes, and solutions to typical biofilter challenges.

xviii The Water Research Foundation • Appendix F – Monitoring Tool Standard Operating Procedures (SOPs) summarize techniques for ATP analysis, determining biofilm formation rate using a pipe loop setup and ATP analysis, dissolved oxygen consumption across the filter for determining bioactivity, analysis of media for extracellular polymeric substances (EPS), and measuring BDOC. • Appendix G – Tools Compendium includes eight references and access links to Excel‐based computer programs, numerical models, standards, databases and manuals available to utilities that are currently operating or considering biofiltration. • Appendix H – Biofilter Optimization Studies summarizes biofiltration optimization studies and the various optimization strategies that were investigated. • Appendix I – Biofilter Optimization Decision Trees for use in improving overall treatment efficiency and water quality from oxidant addition as well as assist in avoiding unintended consequences. • Appendix J – Biofilter Conversion Case Studies summarizes various full‐scale utilities that converted and the water quality or operational drivers for conversion of each of the plants. • Appendix K – Full‐Scale Biofiltration Plants expands the previous list with additional examples of plants and the pre‐treatment and media types used at those utilities. • Appendix L – Sample Testing Plans provide examples of the approach for collecting information related to key considerations detailed in Chapter 8: Biofiltration Testing. Biofiltration is a reliable and effective unit treatment process that offers numerous benefits over traditional granular media filtration. However, biofilters pose unique challenges that must be understood to assure effective long‐term performance benefits. Operators and engineers should evaluate source water quality and finished water treatment objectives to decide if the benefits of biofiltration are desirable. The Biofiltration Guidance Manual was developed to provide engineers and operators with the knowledge and understanding necessary to determine if biofiltration would be reliable and effective for meeting treatment objectives at new or existing WTPs.

Biofiltration Guidance Manual for Drinking Water Facilities xix

xx The Water Research Foundation

Background

Biological filtration (biofiltration) is defined by the American Water Works Association (AWWA) Biological Drinking Water Treatment Committee as “the operational practice of managing, maintaining, and promoting biological activity on granular media in a filter to enhance the removal of organic and inorganic constituents before treated water is introduced into the distribution system” (Brown et al. 2016). In the presence of low or no oxidant residuals, indigenous organisms attach on the filter media surface and form a biofilm. Once established, the quantity of biomass maintained within the biofilm matrix tends to remain consistent over time, and the biomass is resilient to varying temperatures, backwashing, substrate loading rates, and in some cases, low concentrations (e.g., < 0.1 mg [milligram]/L [liter] free chlorine) of filter influent oxidant residuals (Evans et al. 2013a, Hooper et al. 2019, Lauderdale et al. 2018, Pharand et al. 2014). Engineered Biofiltration refers to design and operational strategies to enhance biofilter performance through enhancing microbial health. Some of the strategies include media optimization, pH adjustment, nutrient addition, and backwash optimization. Studies have shown that engineered biofiltration can result in increased biological activity (due to biodegradable organic matter [BOM] availability), improved water quality, and improved hydraulic performance (e.g., reduced headloss and increasedn filter ru times [FRTs]) (de Vera et al. 2019, Lauderdale et al. 2018, 2014, 2012, 2011, Metz et al. 2016). Conversely, passive biofiltration may also be observed, in which conventional media filters develop biomass over time, albeit not by design or intentional practice, and demonstrate biological removal of contaminants (Evans et al. 2010). A recent AWWA survey covering 45 states, one U.S. territory, and nine Canadian Provinces concluded that biofiltration is widely accepted and does not have any regulated design or monitoring requirements beyond those in place for conventional filtration (Nieminski and Perry 2015). 1.1 History Media filtration, often incidental biofiltration, of municipal drinking water was first applied using slow‐ sand filtration (0.02 to 0.08 gpm/ft2 [gallons per minute per square feet]) around 1800. Following cholera outbreaks in Europe during the latter half of the century, and the realization that filtration prevented waterborne diseases, water filtration quickly spread throughout Europe and North America (Crittenden et al. 2012). The U.S. began rapid‐rate filtration (2 to 6 gpm/ft2), which was favored as it allowed for a reduced footprint (relative to slow‐sand filtration) and improved water quality and production in the early 1900s (Crittenden et al. 2012). Rapid‐rate filters outnumbered slow‐sand filter installations by the 1930s. It wasn’t until the recognition of DBPs and biological instability in distribution lines that biofiltration was more purposefully implemented (Chaudhary et al. 2003). Biofiltration was first implemented in Europe in the 1970s to remove NOM and slowly reached North America in the 1980s (Uhl 2008). Implementation has been relatively slow due to public perception and resistance to intentional biological growth in drinking water treatment systems, counter to traditional practice (Evans et al. 2010). However, this mentality has shifted as increasingly stringent regulations for turbidity, disinfection, and DBPs have been implemented [e.g., Surface Water Treatment Rule (SWTR), Interim Enhanced SWTR, Long‐Term 1 Enhanced SWTR, and Stage 1 and Stage 2 Disinfectants and Disinfection Byproducts Rules (DBPRs)] and additional benefits of biofiltration have been recognized (e.g., improved biological stability, reduced chlorine demand, removal of DBPs and T&O compounds, chemical reduction/green approach, etc.) (Evans et al. 2013a, Zhu et al. 2010). Today, there are hundreds of biofiltration facilities across North America. To better inform and guide utilities on the implementation

Biofiltration Guidance Manual for Drinking Water Facilities 1 and operation of biofilters, the WRF launched the research focus area “Biofiltration: Defining Benefits and Developing Utility Guidance” in 2012. The objectives were to 1) “provide guidance documents for implementing, enhancing, monitoring, and optimizing biofiltration,” and 2) “communicate the attributes of biofiltration and how it can enhance drinking water treatment effectiveness” (Water Research Foundation 2019). As part of the focus area, over $4 million in funding was provided between 2012 and 2018 to research biofiltration. A summary of the historical progression of biofiltration practice and WRF studies are detailed in Figure 1‐1, with published reports available online via WRF. This Biofiltration Guidance Manual draws from the findings of studies supported by WRF, AWWA, and other organizations to create a comprehensive reference that guides the design, operation, monitoring, conversion, and optimization of rapid rate biofilters for drinking water utilities.

Figure 1‐1. Timeline Detailing Progression in the Development, Implementation, and Research of Biofiltration. Source: Cecen and Aktas 2011, Crittenden et al. 2012, Culp and Clark 1983, WRF 2019.

2 The Water Research Foundation 1.2 Treatment Objectives Table 1‐1 summarizes the removal capacity (with and without pre‐ozonation) and drivers for biofiltration for multiple classes of contaminants. Similar to conventional filtration, biofiltration is applied for the removal of pathogens and particles. Additional treatment benefits beyond those achieved by conventional filters include the reduction of T&O compounds [e.g., MIB and geosmin], NOM, and regulated DBPs (e.g., trihalomethanes [THMs] and haloacetic acids [HAAs]) (Bouwer and Crowe 1988, Dickenson et al. 2018, Emelko et al. 2006, Evans et al. 2013a, Hooper et al. 2019, Lauderdale et al. 2018, McKie et al. 2015, Persson et al. 2007). Biofilters are also implemented for removal of metals (e.g., iron and manganese) and ammonia, and have been shown to improve the biological stability [e.g., AOC and biodegradable organic carbon (BDOC) removal] of drinking water in distribution systems (Brown et al. 2017, Kohl and Dixon 2012, Lauderdale et al. 2018, LeChevallier et al. 2015, Simpson 2008, Wert et al. 2008). Although not currently regulated, and therefore not a current significant driver, biofilters can reduce some anthropogenic compounds like endocrine disrupting compounds (EDCs) and pharmaceutical and personal care products (PPCPs) (Dickenson et al. 2018, Snyder et al. 2007, Zearley and Summers 2012). Table 1‐1. Contaminant Drivers and Total Removal Potential by Rapid‐Rate Biofiltration With and Without Pre‐Ozonation. Total Removal Potential Contaminant Typical Contaminant Biofiltration with Class Parameter Driver Biofiltration Only Pre‐Ozonation Particles, Turbidity Regulatory Compliance High (with appropriate High (with appropriate Pathogens (SDWA, SWTR, particle conditioning) particle conditioning) LT2ESWTR)

Oxidation AOC, BDOC, Carboxylic Biostability, Reduced Moderate to High High* Byproducts acids, Aldehydes, Chlorine Demand Ketones

DBP Precursors THMs, HAAs, NDMA, Regulatory Compliance Low to Moderate Moderate (NOM) TOC (Stage 1&2 D/DBP Rule)

Algal MIB, Geosmin, Aesthetics, Health Moderate to High High Metabolites Cyanotoxins Effects

Anthropogenic EDCs, PPCPs, High‐Quality Water None to Moderate None to High Compounds Pesticides

Inorganic Iron, Manganese, Secondary Drinking None to High Moderate to High Compounds Ammonia Water Standards, Reduced Chlorine Demand

* Removal may decrease with lower temperatures 1.3 Removal Mechanisms Three primary contaminant removal mechanisms may be observed in biofiltration: 1) physical separation (i.e., particle removal), 2) adsorption, and 3) biodegradation/biotransformation (Figures 1‐2 and 1‐3). Biofilters act as conventional rapid‐rate filters by physically removing particulates from the water through a contact process when particle charge has been stabilized.C If GA media is used, rather than non‐adsorptive (i.e., inert) media like anthracite and sand, some contaminants can also be removed by adsorption to the media, as shown in Figure 1‐2. However, this adsorption capacity becomes exhausted over time, and the breakthrough of adsorbable compounds occurs. Constituents with a lower sorption affinity, often including those that are less biodegradable, may also desorb from

Biofiltration Guidance Manual for Drinking Water Facilities 3 the GAC. Upon biofilter acclimation, compounds can be removed by biodegradation and biotransformation. Acclimation refers to the period from startup until steady‐state removal of the target contaminant(s) is achieved when the biofilm is establishing on the media and adapting to the water quality. This period can take weeks to years., (i.e some PPCPs), depending on the target contaminant and origin of the filter media (e.g., virgin or used) (Brown et al. 2016, Dickenson et al. 2018, Zearley and Summers 2012). During this phase, microorganisms attach to the filter media and proliferate, increasing the biomass on the filter media. Biodegradation/biotransformation then becomes the primary removal mechanism, regardless of media type (e.g., anthracite or GAC), depicted in Figure 1‐2. Contaminants are degraded by bacteria within the biofilm that is attached to the media surface and sometimes macropores [pore radius greater than 25 nm (Sontheimer et al. 1988)].

Figure 1‐2. Effluent Concentration as a Function of Time Comparing Removal Trends Using an Inert Filter Media (e.g., Sand, Anthracite) and Fresh Adsorptive Media (e.g., GAC) from Start‐Up. Adapted from Brown et al. 2018.

Typically, biofilters are not operated for adsorption, but rather operated for numerous years relying on biodegradation, biosorption, and biotransformation processes for contaminant removal. When a biodegradable organic contaminant adsorbs to the biofilm or media, microorganisms degrade or transform it via two metabolic pathways: direct catabolism (i.e., primary substrate utilization) or co‐ metabolism (i.e., secondary substrate utilization) (Benner et al. 2013). Primary substrates are compounds that are present at concentrations high enough to provide energy for primary cellular processes without the use of another substrate. In drinking water biofilters, the primary substrate utilized is BOM (Benner et al. 2013). In secondary utilization, secondary substrates (compounds present at concentrations too low to directly support primary cellular functions) are biodegraded by bacteria using enzymes generated by primary substrate metabolism (Rittmann and McCarty 2001). Micropollutants (e.g., EDCs, PPCPs, pesticides) (Greenstein et al. 2018, Zearley and Summers 2012) and algal metabolites (e.g., MIB and geosmin) (Elhadi et al. 2006) are often secondary substrates. However, the extent or rate at which micropollutants are degraded varies, as some contaminants are more

4 The Water Research Foundation susceptible to biotransformation (e.g., MIB, geosmin, microcystin, acetaminophen, ibuprofen, triclosan) than others (e.g., atrazine, carbamazepine, meprobamate, sucralose) (Dickenson et al. 2018, Greenstein et al. 2018, Zearley and Summers 2012).

Figure 1‐3. Illustration of Contaminant Removal Mechanisms by Biological Filtration with a Porous Media. 1.4 Biofilm Characteristics and Dynamics To fully understand biofilter performance, particularly during the startup of new or newly converted biofilters, it is important to understand biofilm characteristics and lifecycle. Typical biofilm contains 10% to 90% EPS and the remaining fraction microorganisms, by weight (Flemming and Wingender 2010). EPS supports the biofilm and is excreted by microorganisms, with a mature biofilm consisting of 90% proteins and 10% carbohydrates (high molecular weight polysaccharides) and deoxyribonucleic acid (DNA) (Hooper et al. 2019, Keithley and Kirisits 2018). However, at time of production, EPS contains nearly 50% carbohydrates (easily degraded) and 50% proteins, with microorganisms consuming the carbohydrate and protein fractions as needed to sustain growth (Wang et al. 2007). Conversely, under periods of high stress (e.g., nutrient‐limited conditions or exposure to a strong oxidant), they may produce excess EPS and filamentous biofilm morphology that can impair hydraulic performance (e.g., increased headloss due to media or underdrain restriction caused by EPS fouling) until conditions in the biofilter improve (Keithley and Kirisits 2019, Lauderdale et al. 2011). Regular backwashing (typically 24 to 96 hours, but as high as 200 hours) (Brown et al. 2016), often combined with collapse‐pulse air scouring (Brown et al. 2016, Emelko et al. 2006, Lauderdale et al. 2018) or sometimes low doses of hydrogen peroxide (Lauderdale et al. 2012), is used to control biomass and mitigate biomass release during filter runs. Analysis of heterotrophic plate counts (HPCs) and cellular ATP can be used as an indicator for biomass detachment during filter runs (Dickenson et al. 2018, Evans et al. 2013a). Selection and abundance of microorganisms within the biofilter is site‐specific and is dependent on the water source, influent water quality (i.e., nutrient availability), media type, and operational practice (e.g., pre‐oxidation, backwashing) (de Vera et al. 2018, Lauderdale et al. 2011, Moll et al. 1998, Pinto et al. 2012). Every natural system supports a unique microbiota containing indigenous microorganisms that effectively seed the biofilter and form a biofilm on the filter media. Bacteria present in the filter influent are available to continuously seed the biofilters and form biofilm on the filter media. The abundance and diversity of microbial populations also vary with filter depth, as some microorganisms require more substrate, nutrients, or dissolved oxygen (DO) than others (Moll et al. 1998). Despite seasonal changes and occasional fluctuations in influent water quality, the biofilter biomass (Evans et al. 2013a, Pharand

Biofiltration Guidance Manual for Drinking Water Facilities 5 et al. 2014, Hooper et al. 2019) and microbiome in a fully acclimated biofilter (i.e., a biofilter with established mature and active biomass that results in sustained, steady‐state biodegradation) are typically stable (Pinto et al. 2012). Figure 1‐4 depicts the biological characteristics of a typical conventional filter with pre‐chlorine disinfection and anthracite media compared to a biological filter consisting of anthracite or GAC media. When a disinfectant residual (i.e., free chlorine) is maintained during filtration, it will limit bioactivity and support minimal biomass. Conversely, when operating the filter as a biological filter (minimal or no oxidant residual is maintained in the filter), microorganisms present in the filter influent may attach to the filter media and grow as a biofilm. The biofilter influent source water quality (e.g., oxidant presence/absence) will dictate the degree of microbial diversity and bioactivity of the biofilter. In general, not all bacteria present in the influent will proliferate in the biofilter (e.g., the purple bacteria pass through in Figure 1‐4). Other bacteria could proliferate in the biofilter and not detach frequently (Pinto et al. 2012) (i.e., the blue/green bacteria are “strict” colonizers in Figure 1‐4). However, most bacteria present within the biofilter are considered “leaky” colonizers (e.g., red/orange bacteria in Figure 1‐4). These bacteria have the greatest potential to be present post biofiltration as they are derived from a stable reservoir of bacterial biomass on the filter that can detach. After biofiltration, post disinfection is employed to inactivate remaining bacteria. Due to the reduced organic matter in the biofilter effluent, as compared to conventional filtration, oxidant demand (e.g., chlorine) of the biofilter effluent typically decreases (Lauderdale et al. 2018).

Figure 1‐4. Typical Biological Characteristics of a Conventional Anthracite Filter with an Oxidant Residual and Anthracite/GAC Biological Filter. Adapted from de Vera et al. 2018, Lauderdale et al. 2011.

6 The Water Research Foundation 1.5 Biofilter Performance Drivers Several design and operating parameters (e.g., empty bed contact time [EBCT], temperature, pre‐ oxidation, medium type) affect contaminant removal across biofilters, most notably when targeting simultaneous removal of multiple contaminants (e.g., manganese, DBP precursors, PPCPs, EDCs). General observations and trends from previous studies are briefly summarized here (Table 1‐2), with more detail provided in subsequent chapters. Influent water quality affects the performance of biofilters and should be closely monitored to ensure optimal conditions for biofilter performance. Biofiltration is most effective at warmer temperatures (exceeding 15°C) (Dickenson et al. 2018) as reaction kinetics double with every 10°C increase (Evans et al. 2013). Significant increases in the biofilm formation rate have been observed at facilities with temperatures greater than 15°C, with declined biofilm formation at locations below 5°C (Hooper et al. 2019). A pH between 6 to 9 is recommended to promote biological activity – most drinking water sources are 6.5 to 8.5 (Evans et al. 2013). The typical EBCT of biofilters is five to 15 minutes (Brown et al. 2016), with little improvement in contaminant removal beyond 10 minutes at temperatures exceeding 15°C (Dickenson et al. 2018). However, an EBCT greater than 10 minutes is beneficial for trace organic compounds (TOrCs) removal when the temperature is below 15°C (Dickenson et al. 2018). Ozonation is often coupled with biofiltration and improves the overall removal of contaminants (Table 1‐1), including NOM, trace contaminants, dissolved manganese, MIB, and geosmin (Dickenson et al. 2018, Hozalski et al. 1999, Westerhoff et al. 2005). Pre‐oxidation promotes biodegradationM of NO through increased formation of AOC; however, too high of an oxidant residual in the biofilter influent, especially when using sand or anthracite media (i.e., no quenching capacity of the oxidant residual), impairs bioactivity and performance (Evans et al. 2013). Pre‐chlorination and pre‐chloramination are not recommended without a quenching step before biofiltration, as residuals greater than 0.1 mg/L free chlorine adversely affect the biofilm, especially in biofilters with non‐GAC media (Hooper et al. 2019). Pre‐chlorination with filters using GAC has also reduced the removal of TOrCs, whereas effects on manganese, HAAs, and ammonia removal were lesst significan (Dickenson et al. 2018). Media selection should be carefully considered in relation to influent water quality and desired biofilter effluent quality (including particle removal). Following biofilm acclimation, inert (e.g., sand or anthracite) and adsorptive media (e.g., GAC) can demonstrate different contaminant removal efficiencies, particularly adsorbable micropollutants. In general, both types of media have demonstrated the ability to achieve simultaneous removal of DOC, AOC, MIB and geosmin, manganese, ammonia, and DBP precursors (Dickenson et al. 2018, Greenstein et al. 2018). However, GAC usually demonstrates better removal of TOrCs than inert media (Zhang et al. 2017). The adsorption capacity of GAC, while exhausted from a TOC/DOC perspective, may not become completely exhausted to some recalcitrant TOrCs for several years. Studies have reported adsorption of carbamazepine and tris‐2‐chloroethyl phosphate (TCEP) for one year of operation, and adsorption of fluoxetine and triclocarban for over 10 years (Dickenson et al. 2018, Stanford et al. 2017). GAC media has also been shown to support higher biomass due to a combination of high surface area for attachment and irregular surface structure providing protection from high shear rates during backwashing (Evans et al. 2013, Huck et al. 2000, Lauderdale et al. 2018), although this does not always translate to higher bioactivity or substrate utilization (Evans et al. 2013, Pharand et al. 2014). Further, studies have indicated that biofilters containing inert media can have similar BOM removal to spent GAC media at warm temperatures (Emelko et al. 2006, Huck et al. 2000).

Biofiltration Guidance Manual for Drinking Water Facilities 7 Table 1‐2. Biofilter Performance Drivers and Their Impact. Performance Driver Impact Temperature Biofilm formation rate, biological activity, contaminant removal

pH Biological activity and contaminant speciation/removal

EBCT Contaminant removal

Pre‐oxidation Biodegradability, contaminant removal, and hydraulic performance

Media Contaminant removal, biological activity

8 The Water Research Foundation

Monitoring Tools and Instrumentation

An organized and streamlined monitoring strategy is important for assessing the performance and operability of biofilters. Monitoring must account for physical characteristics to confirm media and underdrains are within original and operational specifications, as well as assess the efficacy of biological processes and potential unintended consequences such as filter headloss and underdrain fouling. Monitoring for chemical and water quality parameters are equally important, which can be key indicators of the biological health of the filters and inform decisions on operations upstream and downstream. This chapter describes the monitoring tools that are recommended for plants operating biofilters to inform operators on the health of their filters and when to take action to prevent a lapse in water quality. SOPs for monitoring are also included in Appendix F. 2.1 Filter Integrity Monitoring Careful monitoring of filter media and underdrain performance is crucial to biofilter operation and water efficiency. At a minimum, biofiltration requires similar monitoring as in conventional media filtration. A summary of conventional filter monitoring requirements is provided, but these procedures and protocols are not discussed in detail as several excellent resources are available on this topic. For more information, the reader may consult: • Integrated Design and Operation of Water Treatment Facilities by Susumu Kawamura (2000). • Filter Evaluation Procedures for Granular Media by Daniel Nix and John Taylor (2018). • Filter Maintenance and Operations Guidance Manual (WRF 2511) by Gary Logsdon et al. (2002). • Filter Troubleshooting and Design Handbook by AWWA (Beverly 2005). This section discusses monitoring parameters and techniques for filter media beds and underdrains specifically applicable to biofiltration (i.e., in addition to conventional filtration techniques). 2.1.1 Media Inspection The following section describes several methods, procedures, and recommendations related to biofilter media inspection. The section also discusses recommended inspection frequencies and provides a decision tree to assist water utility personnel in determining the right time to replace their media. 2.1.1.1 Media Testing Similar to conventional filtration, biofilters have the primary objective of particulate removal. Filters treating surface water supplies must meet the requirements for turbidity that with the Surface Water Treatment Rule. To provide effective filtration performance, filter media needs to meet minimum industry requirements. In North America, ANSI/AWWA B100 – Granular Filter Materials is the industry standard for filter media. Biofilters incorporating GAC also need to meet ANSI/AWWA B604 – Granular Activated Carbon. Below are definitions of key filter media parameters, adapted from the latest versions of AWWA B100 and B604. • Effective Size (ES): The sieve size (or mesh size [MS]) opening that will pass 10% by dry weight of a representative sample of the filter material. This means that 10% by dry weight is finer than the ES. The ES, also known as D10, is measured in millimeters (mm) and is specified with a range of tolerance, typically +/‐ 0.05 mm. • Mesh Size (MS): For filtration applications, mesh size refers to the U.S. standard sieve size used for classification of filter media. A filter media may be specified in accordance with the percent passing values of oversize (upper bound) and undersize (lower bound) mesh sizes as an alternative or in

Biofiltration Guidance Manual for Drinking Water Facilities 9 addition to the ES. The mesh size specification is more commonly used to describe GAC media rather than anthracite or sand. Commercial GAC products are typically classified by mesh size. For example, an 8x20 mesh GAC is defined by specific percent passing limits on No. 8 and No. 20 mesh sieves. • Uniformity Coefficient (UC): A ratio calculated as the size of the opening that will pass 60% by dry weight of a representative sample of the filter material (D60), divided by the opening that will just pass 10% by dry weight of the same sample (D10). Typical biofilter media specifications call for a UC between 1.3 and 1.5; however, some proprietary GAC media products used for filtration have UC values up to 2.1. Higher UC values (UC = 1.5 ‐ 2.1) are more commonly used in GAC media whose primary purpose is adsorptive contaminant removal. • Specific Gravity (SG): Calculated as the ratio of the density of filter media to the density of water. Several methods for determining the specific gravity of medium are outlined in AWWA B100. According to AWWA B100, the SG values for all methods are reported on lab test results. The most commonly used and specified values are saturated surface dry SG for silica sand and average apparent SG (AASG) for anthracite. Typical filter media‐specific gravities range from 2.6 to 2.65 for silica sand and from 1.6 to 1.7 for anthracite mined in North America. The SG of anthracite mined in Europe or elsewhere may range as low as 1.4. • Apparent Density (AD): GAC densities are typically not related to water as due to the skeletal nature of the particle the SG is highly affected by the pore volumes. Instead, the apparent density is typically reported, measured in lbs/ft3 (pounds per cubic feet), g/cc, or g/mL. The apparent density may also be reported by GAC suppliers as “apparent density, backwashed and drained”, which represents the mass of GAC in a packed bed volume after backwashing and draining. • Moh’s Hardness (MH): A measure of the hardness of a media and shear strength, typically specified as greater than 2.7 for anthracite and greater than 6.5 for sand. • Acid Solubility (AS): A test in which a media sample is immersed in acid. The sample weight before and after is measured to determine how much acid‐soluble material was present. Per AWWA B100, acid solubility is a measure of the degree of chemical fouling (i.e. coagulant solids, calcium carbonate precipitation, metals, etc.) and the media’s inherent resistance to acid. Acid solubility of filter media is typically specified lessn tha 5%. • Abrasion Number (AN): A parameter reported for GAC to indicate the relative resistance of the carbon particle to abrasion and breakage. For bituminous GAC, typically employed in surface WTPs, an abrasion number greater than 75 is typically specified. If any degradation of the abrasion number is observed, the user should consider the replacement of the media. • Iodine Number (IN): A measurement used as an indicator of adsorptive capacity of GAC, expressed in units of mg/g. Virgin GAC is typically specified with an IN of 900 mg/g. As the adsorptive capacity of the GAC is exhausted, the IN will decrease. Analysis of the above parameters is performed upon purchase of the filter media. A particle size distribution analysis is performed to demonstrate thate th media meets specified requirements for ES and UC. However, over time, the media bed characteristics may change due to many operational factors including loss of fine particles during backwash, abrasion during backwash, and deposition and adsorption of filtered material onto the media (e.g., calcium carbonate). Additionally, abrasion of filter media over time can decrease the particle size and result in a higher concentration of fines at the top of the filter bed and result in higher clean‐bed headloss. As smaller media particles are removed, and larger heavier granules remain, filtration performance may be impacted. Biofilter media core sampling and characterization should be performed annually to evaluate any changes in effective size, mesh size, and uniformity coefficient. Figure 2‐1 shows examples of methods used for collecting filter media profile samples.

10 The Water Research Foundation Additional testing can be performed to determine if unwanted materials have deposited on the surface of the filter media, including acid solubility tests and chemical composition analysis. Investigation of the media angularity can also be performed by media analysis laboratories using microscope imaging. Several tests can be performed to evaluate filter performance and conformance with the original specified media characteristics over the life of filter operation. These are listed in Table 2‐1. A resource to consult on the mechanics of these tests is Filter Evaluation Procedures for Granular Media by Daniel Nix and John Taylor (2018). Table 2‐1. Filter Media Integrity Monitoring Techniques. Test/Monitoring Frequency for Technique Description Purpose and Value of Test Biological Filters Observations during backwash and operation to check for: • Uneven distribution of water and air (or surface wash water) • Evidence of plugged nozzles (if surface wash is used) • Excessive deposition of materials on the surface of the filter Identifies obvious problems that • Excessive biological growth on the Visual Inspection require further attention and more Weekly filter basin structure and in‐depth inspection and analysis components • Noticeable changes to visible biological growth (i.e. color change, etc.) • Media loss during backwash, mudballs, media mounding and/or cratering • Backwash pressure trends Sampling of the entire depth of the Used to determine conformance with Core Sampling/ media bed for particle size distribution specified requirements and ascertain if Annually Media Analysis and other parameters. media replacement is required. Measurement of the total media Useful in determining if capping of bed depth. Methods such as punch rod Media Depth with additional media is required to measurement, as described by Nix and Annually Measurement meet specified depth or if media Taylor (2018), measure the whole replacement is required. media bed depth. Measurement of the amount of bed expansion during backwash. Methods Determines if backwash process Backwash for determination include using a pipe Semi‐Annually or provides effective bed expansion, Expansion Analysis organ apparatus, Secchi disk, or by Annually typically 30‐50% or more. measuring the depth of penetration of a pipe into the fluidized bed. Used to determine the quantity of solids retained throughout the media A test which determines the relative Floc/Solids bed. Profiling before and after concentration of solids throughout the Annually Retention Profiling backwash also is valuable to assess media depth. and optimize the performance of the backwash process (see Figure 2‐2). Excavation of the filter bed to Determines depth of each filter media Once every 5 to Filter Excavation determine conditions throughout the and gravel layer, used to inspect filter 10 years bed. underdrains

Biofiltration Guidance Manual for Drinking Water Facilities 11

Figure 2‐1. Example Alternative Methods Used for Collecting Filter Media Profile Samples, Including Use of a Check Valve on The Bottom of the Corer (a) and (b), Use of an Open Pipe with Handles on the Top (c), a Box Corer (d), and an Example of a Grain Thief Sampling). Device (e Sources: Panels a‐d, Evans et al. 2013a, Panel e, Hooper et al. 2019. Figure 2‐2 shows solids retention via the turbidity profile before and after backwashing. The right line (blue) represents a normal solids retention curve at the end of a filter run (before backwashing). The solids retention is greatest at the top of the media bed and tapers off throughout the bed.e Th left (yellow) line in the figure represents a floc retention profile after backwash. A properly backwashed filter will have 30‐60 nephelometric turbidity units (NTU) of retained solids per 100 grams of media throughout the depth of the media bed (Kawamura 2000).

Figure 2‐2. Typical Solids Retention Curves in a Filter Bed Before and After a Backwash. Reprinted from the Journal of NEWWA, Vol. 113 (No. 1) by permission. Copyright ©1999 the New England Water Works Association.

12 The Water Research Foundation 2.1.1.2 Inspection Frequency Table 2‐1 in the previous section lists the recommended frequency for several filter media inspection techniques. Annual filter media sampling and analysis is recommended to determine if the media has deviated from the specified characteristics. If several successive years of testing have demonstrated that the media remains compliant, operators may consider reducing the frequency of sampling. Alternatively, if GAC is used in the biological filter and the plant is targeting adsorptive removal of trace constituents, operators may consider more frequent water quality sampling to verify remaining adsorptive capacity. At least once per year, plant operators should perform traditional filter surveillance techniques such as floc/solids retention profiling (before and after backwash), media depth measurements, backwash bed expansion, and filter component condition assessment. Regular filter surveillance provides an important opportunity that can inform decision making and guide filter operation modifications. 2.1.1.3 Media Replacement Frequency Determining when to change out biofilter media is a common question. Some biological filtration utilities have operated with the same media for 25 to 30 years even with GAC media. However, as some utilities top off filter media to account for losses and over time this eventually results in a full changeout. Example facilities include Arlington Water’s John F. Kubala and Pierce‐Burch WTPs in Arlington, Texas, the Central Lake County Joint Action Water Agency in Illinois, and the Lanier and Shoal Creek Filter Plants at Gwinnett County, Georgia. However, a targeted approach can be developed to make informed decisions on media replacement frequency. Figure 2‐3 shows a media replacement decision tree. Here a media sample would be collected from three locations: the entire profile, the top 20% of the profile and bottom 80% of the profile. Many utilities use the IN and/or the abrasion number to track when replacement is required. Typical triggers for replacement would include IN of 500 to 550 mg/g (ANSI/AWWA B604‐12) and/or abrasion number less than 75. Acid solubility testing could also be included, with media replacement recommended when values are above 5% (AWWA B100). However, replacement of GAC based on IN is not required if the media is not being targeted for adsorption. In this case, replacement of the GAC based on physical changes is recommended, as would be the case for replacement considerations for anthracite and sand. Other utilities have used TOC breakthrough, or an increase in effluent TOC levels above a pre‐established baseline, as a trigger for GAC media replacement. Biological filters that are also targeting adsorptive removal with GAC may have more stringent triggers for media replacement, which will be driven by the adsorption capacity of the GAC for a given target contaminant.

Biofiltration Guidance Manual for Drinking Water Facilities 13

Figure 2‐3. Filter Media Replacement Decision Matrix.

14 The Water Research Foundation 2.1.2 Underdrain Inspection Biologically active filters require additional monitoring to prevent damage to filter equipment. Excess biomass accumulation may lead to long‐term fouling of media and filter underdrain equipment, significantly limiting production rates/water efficiency and, in some cases, has led to catastrophic underdrain failures (Lauderdale et al. 2011). Underdrains can foul due to biofouling or mineral scaling if they are not effectively cleaned during backwash cycles. Measurement of backwash pressure or backwash underdrain headloss, ideally at the filter, can be used to indicate if underdrain fouling is occurring (see Section 2.2.5). Figure 2‐4 shows an underdrain support cap with biofouling. Media support caps using sintered high‐ density polyethyleneplastic beads are not recommended for biofiltration due to the potential for irreversible fouling. Biological material can enter the media support cap and proliferate in the tortuous paths formed by the sintered beads. This biomass is very difficult to remove once established. Many biological filters operate with these types of caps and operators of those plants are cautioned to be diligent with backwash pressure or backwash headloss monitoring to verify that fouling does not develop. For these filters backwash water containing low doses of oxidants such as peroxide or chlorine can be used for control. If an upward trend in fouling is observed, the utility should consider replacement with a more suitable media support design. Slotted media support designs are better suited for biological filtration. Gravel media support is also suitable in cases where direct media retention is not provided by the underdrain.

Figure 2‐4. Pilot Biofilter Underdrain Caps Operated Under (a) Nutrient‐Limited and (b) Phosphorus‐Enhanced Conditions. Source: Lauderdale et al. 2014

Filter media, underdrain, and underdrain plenum inspections are very effective tools as a condition assessment. However, due to the cost and time involved they are typically employed on five‐ to 10‐year frequencies. A filter excavation is the preferred tool to inspect the underdrains and is recommended when increased backwash pressure indicates fouling. Filter underdrain plenum inspections are only possible with monolithic false floor nozzle style underdrains, Wheeler‐style underdrains, plastic block flume, or other suspended floor style underdrains. Many suspended slab underdrains include manways in the false floor whereby the plenum can be accessed when the media is removed org durin a filter excavation. If a manway into the plenum is accessible from the filter gallery, a plenum inspection can be performed without having to do a filter excavation and thus can be performed more frequently.

Biofiltration Guidance Manual for Drinking Water Facilities 15 2.2 Hydraulic Monitoring Careful hydraulic monitoring of biofilters can facilitate the early detection of potential upset events, provide opportunities for process optimization, and allow for a comparison of various treatment methods. Hydraulic parameters important for biofilter operation are shown in Table 2‐2. Each of the parameters has various monitoring frequencies, locations, and typical operating values (Hooper et al. 2019). Additionally, hydraulic parameters can influence each other; therefore, it is recommended that related parameters be tracked and trended together (Nyfennegger et al. 2016).

Table 2‐2. Hydraulic Parameter Summary. Parameter Typical Operating Values Varies with filter design. Look for increases over multiple filter runs. Headloss Accumulation Rate Should be normalized with loading rate. Clean‐Bed: Varies with filter design (typically 1 to 4 feet [ft]) and loading rate. Clean‐bed and Terminal Headloss Terminal: Varies with filter design (typically 8 to 10 ft, although older filter designs may have lower ranges such as 4 to 6 ft) Differential Pressure Profiling Varies with filter design. Look for changes over time. Filter Run Time (FRT) 24 to 96 hours 8,000 to 16,000 gpm/ft2. Look for decreases over time and Unit Filter Run Volume (UFRV) understand what controls filter backwashing (turbidity, headloss or time). Backwash Pressure Varies with filter design. Look for increases over time. Filtration Rate: Typically, 0.5 to 10 gpm/ft2 Filtration Rate and Empty Bed Contact EBCT: 5 to 15 minutes (while this range is ideal, in practice EBCT can Time (EBCT) range from 2 ‐ >20 minutes) Adapted from Hooper et al. 2019

2.2.1 Filtration Rate Biological filtration can be performed at higher filtration rates, assuming a reasonable contaminant surface loading rate and enough EBCT is provided (see Section 2.2.4). High rate biofiltration with loading rates up to 6.8 gpm/ft2 is being implemented at large (2,300 square ft) filters at the 320 mgd Northeast Water Purification Plant (NEWPP) in Houston, Texas. Higher filtration rates, up to 10 gpm/ft2, have been implemented successfully at other plants (AWWA 1998, Brown et al. 2016, Evans et al. 2013a). However, consideration should be given to the higher shear rates that take place at higher filtration rates. One study indicated that filtration rates up to 7 gpm/ft2 did not have a negative impact on filter biomass. However, higher rates were not investigated, so care should be taken to ensure that biomass release does not occur when higher filtration rates are implemented (Schulz 2014, Servais et al. 1994). For constant‐head/constant level filters, the degree of change in filtration rate over time can be an indication of filter biofouling. The filtration rate can be calculated as follows:

𝐹𝑖𝑙𝑡𝑟𝑎𝑡𝑖𝑜𝑛 𝑅𝑎𝑡𝑒 𝑓𝑙𝑜𝑤 𝑟𝑎𝑡𝑒 (Equation 2‐1) 2.2.2 Headloss Headloss is a measure of the change in pressure across the filter media bed. Pressure can be detected using online pressure transducers, level monitoring instrumentation, or pressure gauges (or a combination of these). Headloss is lowest at the beginning of a filter run, termed the clean‐bed headloss, and increases overe tim until backwashing is triggered. Backwashing may be triggered due to excessive headloss, FRT, or filter effluent turbidity/particle counts. The filter headloss at the end of a run is the terminal headloss. Changes in the pattern of clean‐bed headloss and headloss accumulation rate are important hydraulic parameters for assessing whether backwash cleaning procedures are sufficient

16 The Water Research Foundation or if biofouling is constraining available pore space for water flow. Other factors that may impact clean‐ bed headloss and headloss accumulation rate are shown in Table 2‐3. Coarser, deeper media is typically employed in new biological filter designs as the larger pore space creates less headloss, and thus, the design is more robust with respect to the impacts of biofouling on filter bed headloss. 2.2.2.1 Headloss Accumulation Rate For constant‐flow filters, headloss increases over a filter run as particles become trapped and as bacteria grow within a biofilter (Nyfennegger et al. 2016). The headloss accumulation rate describes how quickly this process occurs. If the progression of headloss accumulation is relatively linear over time, Equation 2‐2 can be used:

𝐻𝑒𝑎𝑑𝑙𝑜𝑠𝑠 𝐴𝑐𝑐𝑢𝑚𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑅𝑎𝑡𝑒 (Equation 2‐2)

However, changes in headloss accumulation can also follow an exponential increase over time. For those cases, follow Equation 2‐3 to calculate the headloss accumulation rate constant (Hooper et al. 2019):

𝐻𝑒𝑎𝑑𝑙𝑜𝑠𝑠 𝐴𝑐𝑐𝑢𝑚𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑅𝑎𝑡𝑒 𝐶𝑜𝑛𝑠𝑡𝑎𝑛𝑡 (Equation 2‐3)

The headloss accumulation rate over several sequential filter runs should remain relatively constant (if the filter loading rate is the same). The headloss accumulation rate will change during a single filter run when headloss is increasing exponentially. This could be a sign of changes in influent solids or loading rate, influent filter aid polymer dose, or the formation of mudballs or cracks within the biofilter (Brown et al. 2016). Table 2‐3 presents several factors that can impact biofilter headloss.

Table 2‐3. Factors that Impact Biofilter Headloss. Factor Impact on Filter Headloss Conditioning of particles upstream with coagulants, flocculants or filter aids may induce filter clogging through floc carryover. Settled water typically has fewer particles and Pre‐Treatment/ thus less friction and clogging, reducing headloss. Solids may also be generated from Particle Removal biological processes, such as manganese dioxide precipitation from biological oxidation of dissolved manganese. Media Grain Size Smaller filter media has greater friction loss and higher headloss. Filtration Rate Higher filtration rates have higher headloss. Water Temperature Colder temperatures have a higher dynamic viscosity, leading to higher headloss. Dissolved gases increase the oxygen or CO2 that is released, potentially leading to air‐ Dissolved Gases and binding and subsequent headloss. Filter headloss that exceeds the static difference Negative Available between the water level and the top of the media will result in negative pressures in Pressure the bed which will pull gas from solution and result in air binding and turbidity breakthrough. With more biological growth and production of biological polymers such as EPS, more Biological Growth headloss is possible. Biologically generated solid precipitates, such as manganese dioxide, may cause slight impacts to headloss. Adapted from Hooper et al. 2019.

Biofiltration Guidance Manual for Drinking Water Facilities 17 2.2.2.2 Clean‐Bed and Terminal Headloss To evaluate headloss, most utilities use online pressure and/or level monitoring instrumentation, which allows for continuous headloss measurements. If continuous monitoring is not possible, it is recommended that headloss be monitored by the operator at least three times per day (Brown et al. 2016). While the clean‐bed headloss is relatively low under most operating conditions, terminal headloss can range from 4.0 to greater than 10 ft (Lauderdale et al. 2011, TEEX 2013). Clean‐bed headloss can be used to detect the onset of biofouling within the biofilter media and is often an indication of insufficient backwashing (Brown et al. 2016). Additionally, an increase in clean‐bed headloss can result in higher water elevations within the filters (assuming the filter is not level‐ controlled), making this parameter important to monitor over the life of the biofilter (Lauderdale et al. 2011). Typically, biofilter operations do not observe higher head loss than conventional filters. However, excess headloss may occur, particularly during warmer temperatures, if overproduction of EPS or other substantial changes in biofilm morphology occurs. Terminal headloss signals when a biofilter should be placed into backwash for the removal of accumulated particles, making this an important monitoring parameter. In constant head filters, the terminal headloss setpoint should not exceed the physical distance from the top of the media to the operating water level. Air binding will occur when the positive driving head above the media is less than the losses through the bed, generating negative pressure in the filter. Air binding will create a decrease in flow at the end of a filter run (typically noted by the filter effluent valve being 100% open), which will create issues with constant filter process operations. 2.2.2.3 Differential Pressure Profiling Biofilter differential pressure can be monitored using manometer tubes or pressure gauges/transducers placed at various locations through the profile of the filter bed, as well as below the underdrains. The pressure head at each monitoring point is read by either recording the height of water for the manometer tubes, or the difference in pressure reading on the pressure gauges/transducers. Most full‐scale plants do not have sample taps at different locations within the filter bed and this analysis is more commonly evaluated in pilot plants, but this option can be retrofitted or incorporated into new biofilter plant designs. Where installed, full‐scale installations typically have differential pressure profiling on one or two filters for use during full‐scale filtration testing and assessment of hydraulic conditions within the bed. Increases in differential pressure can also indicate where within the filter bed headloss is accumulating, due to the filtration of coagulated particles or attachment and growth of bacteria on media, e.g., if it is occurring at the top of the filter (making filter binding more likely), throughout the filter, at the media interface of dual‐media filters, or across the underdrain (Eighmy et al. 1997). In the event air binding is occurring (typically denoted by air release from the filter when the filter is taken off‐line at the end of a filter run), differential pressure profiling in the bed will allow an operator to pinpoint exactly where the negative pressures occur (and make adjustment to terminal headloss). Alternatively, some full‐scale filters include differential pressure instrumentation for monitoring headloss during backwash. This instrument can be piped up with several feed lines on the high‐pressure side which can be toggled between different points in the bed to obtain pressure profile readings. One study used differential pressure to evaluate the impact of oxidants in the filter backwash and found that adding hydrogen peroxide to the backwash water reduced underdrain fouling (Lauderdale et al. 2018). 2.2.3 Filter Run Time and Unit Filter Run Volume FRT is the time between backwashing. FRTs of less than 24 hours could be due to filter design, high settled water turbidity, filter aid polymer dosage, direct filtration operation, or a loading rate that is too high. Unintentionally short FRTs could also indicate hydraulic challenges experienced by the biofilter that need to be addressed, making FRT an important monitoring parameter with respect to addressing hydraulic issues (Brown et al. 2016, Nyfennegger et al. 2016). It should be noted that FTRs below

18 The Water Research Foundation 24 hours can result in plant flowrate limitations due to the potential inability to backwash filters quickly enough or recycle the large volumes of backwash waste washwater. Plants are encouraged to quantify their backwash/backwash waste washwater system limitations to understand the potential impacts of these systems when experiencing shortened FRTs. TheRV UF is calculated by dividing the total volume produced in one filter run by the surface area of the filter, and usually ranges from 8,000 to 16,000 gal/ft2. UFRV can also be calculated by multiplying the filtration rate (in gpm/ft2) by the run time (in minutes) if the filtration rate was constant throughout the run. UFRVs can be used to compare multiple filter runs that have different flow rates or run times (Brown et al. 2016). For simple monitoring of UFRV, the plant Supervisory Control and Data Acquisition (SCADA) system can be programmed to calculate a total volume produced in a filter run which can be divided by the filter area, allowing for a real‐time tracking. If calculated manually, UFRV should be evaluated weekly. Monitoring of UFRVs over time can also be used to monitor whether changes to filter hydraulics are being impacted. 2.2.4 Empty Bed Contact Time The EBCT is a measure of the duration water is in contact with the filter media (and microbial community) and is a function of the hydraulic loading rate and the filter bed depth. EBCT is calculated by dividing the volume of the media bed (V) by the filter flow rate (Q, same flow rate used to calculate the filtration rate) (Equation 2‐4).

𝐸𝐵𝐶𝑇 𝑚𝑖𝑛𝑢𝑡𝑒𝑠 Equation 2‐4

Media bed depth is determined based on filter pilot column performance, required depth for GAC adsorption and/or targeting a minimum total bed depth (L) to media diameter (d) ratio of 1,000 or higher. The L/d guidance applicable to filtration for turbidity removal also applies to biofilters (AWWA 1998). It is well documented that EBCT is one of a few critical variables controlling the degree of contaminant biodegradation (Krasner 1993, Nugroho 2010). In many cases (e.g., biofilter conversions or optimizations), EBCT is not an actual design parameter due to the fact that the existing filter box structure limits bed depth changes and thus changes in EBCT. However, when designing a greenfield biofiltration plant, pilot testing is typically used to identify an optimal EBCT (i.e., achieves treatment goals at minimized cost), which can then be used to design the new filter box structure. Contaminant surface loading rate (i.e., the mass loading of a given contaminant per unit time per unit medium surface area) is a more fundamental biofilter design parameter and is described in detail in AWWA Chapter 17). Similar to EBCT, there is limited opportunity to adjust this parameter for existing biofilters. As EBCT increases, greater constituent removal is often observed. In practice, EBCTs can range from two minutes to greater than 20 minutes but are typically between five and 15 minutes (Brown et al. 2016, Evans et al. 2013b). New biofilter plant designs should target a minimum EBCT of five minutes at maximum plant capacity. Some studies have documented that EBCTs up to 30 minutes are required for some trace chemical constituents (Dickenson et al. 2018), but diminishing returns for NOM removal are typically observed in the first five to 10 minutes (Hozalski et al. 1995). EBCT can be modified during normal operation when the hydraulic loading rate is modified by adding or removing filters from service (Evans et al. 2013b). The depth of media used in the calculation of EBCT is the entire bed depth, except in cases where GAC is used, as there is additional adsorptive contaminant removal. In those cases, EBCT is typically only reported for the GAC layer only.

Biofiltration Guidance Manual for Drinking Water Facilities 19 2.2.5 Backwash Pressure Monitoring Backwash pressure can be used to indicate if underdrain fouling is occurring (Brown et al. 2016, Nyfennegger et al. 2016). Backwash pressure should be monitored near the backwash inlet of each filter or at a common pressure gauge located on a common backwash header. If a common gauge is used, the backpressure at the underdrain will need to be adjusted for the additional losses in the header. To track pressure increases over time, the pressure shall be recorded for each filter at start‐up of a new filter installation. This shall be used as the baseline value for comparison with future pressure monitoring data. It is recommended that the backwash pressure be read at the same backwash rate, during the high flow backwash sequence, to allow for an accurate comparison between filters or the same filter over multiple filter runs (Brown et al. 2016). The backwash flow rate can only be monitored to track underdrain fouling if the backwash is provided from a constant head source (Brown et al. 2016, Nyfennegger et al. 2016). The critical increase in backpressure should be evaluated on a case‐by‐case basis depending on the type of underdrain. However, if the pressure gauge used to monitor filter backwash backpressure is not an online instrument, routine measurements should be taken by operators during backwash cycles. An on‐line pressure transducer gauge on the main header is an effective way to track backpressure on any filter over time. If backpressure increases more than 2 psi or 10% of the baseline value, or above the underdrain manufacturers’ specifications, whichever is less, fouling should be investigated and mitigated. Other methods to monitor backpressure include simple pitot tubes on the backwash pipe at the filter which allow an operator to read the hydraulic grade line directly to track backwash pressure over time. Backwash pressure monitoring is becoming an industry standard on both retrofitted and new biological filter plants. A properly monitored filter is at a lower risk for developing damaging clogging conditions in the filter underdrain; however, as an added factor of safety, many new and retrofitted filter designs also include passive pressure relief systems to prevent damage to the underdrain in the event clogging develops over time. These systems can include pressure relief valves or static vent pipes in the backwash system downstream of the backwash control valve. The venting systems should be sized to evacuate enough water to prevent the underdrain rfrom ove ‐pressurizing beyond its design limits under a clogged condition. 2.3 Water Quality Monitoring 2.3.1 General Water Quality (Turbidity, Temperature and pH) General water quality parameters that must be monitored on biofilters include temperature, pH, and turbidity. Table 2‐4 presents typical temperature, pH and turbidity ranges. Continuous monitoring is often accomplished through use of SCADA systems, but these parameters should also be verified manually, via grab samples. Water temperature has a strong impact on biological removal rates, with an approximate doubling for every 10°C increase in water temperature. However, the impacts of temperature on biofilter performance are less significant if GAC is used (Emelko et al. 2006, Huck et al. 2000, Lauderdale et al. 2018). Although temperature cannot be easily controlled in a biofilter, temperature should be evaluated alongside other constituents including NOM removal, biological process parameters (including biofilm formation rate and DO consumption), as temperature is known to have an impact on these parameters (Evans et al. 2013a, Huck et al. 2000, Moll et al. 1998, Nyfennegger et al. 2016). During periods of colder temperatures, more filters could be placed online to increase EBCT and allow for additional degradation of contaminants. Periods of colder temperatures also often coincide with lower water demand, which can decrease plant flow and, consequently, increase EBCT. Increases in turbidity can correlate with increased concentrations of some constituents, including bacteria, Giardia, and Cryptosporidium (Crittenden et al. 2012). Although biofilter influent turbidity typically ranges from 0.01 to 1.5 NTU (Evans et al. 2013b), some studies have reported higher values (Nyfennegger et al. 2016). Filter effluent turbidity is regulated under the SWTR and is the primary

20 The Water Research Foundation treatment objective for filters. The Interim Enhanced SWTR (IESWTR) (1998) states that for systems serving at least 10,000 people with surface water sources, or groundwater sources directly under the influence of surface water, combined filter effluent turbidity must be ≤ 0.3 NTU in at least 95% of samples collected each month, ande non may exceed 1.0 NTU at any time (sampled every four hours). Many utilities operate filters to meet tighter requirements (<0.1 NTU 95% of time) per the Partnership for Safe Water (PSW) goals to provide an improved Cryptosporidium removal barrier (Linder and Martin 2015). The IESWTR states that individual filters must be monitored continuously, and turbidity measurements must be recorded every 15 minutes. The Long‐Term Enhanced Surface Water Treatment Rule (LTESWTR) requires that individual filters require follow‐up action if the filter turbidity exceeds 1 NTU for two consecutive readings or more (EPA 2006). Additionally, filter effluent turbidity is often used as a trigger for backwashing (along with time and headloss). Highly colored water can interfere with turbidity measurements due to the absorption of light (turbidity measures the scattering of light). Grab samples should be collected to validate online turbidimeter readings (SM [standard method] 2130; samples should be analyzed immediately after sample collection). The optimal pH for biofilters is between 6 and 9 SU; operating outside of this range may impair most biological reactions (Evans et al. 2013b, Lauderdale et al. 2014). The pH directly influences chemical dose requirements throughout the treatment plant. For example, many coagulants are optimal at slightly acidic pH (less than six standard units, SU) and a higher pH (in excess of 9 SU) is targeted for distribution system corrosion control at many facilities. pH can be measured at the filter influent continuously using a probe or by measuring grab samples using a portable or bench‐top pH meter following SM 4500‐H+ (APHA et al. 2012). If measured using a portable/bench‐top pH meter, the sample should be measured immediately, if possible, within 15 minutes after collection (Evans et al. 2013b). If the sample is unsaturated or oversaturated with carbon dioxide, the pH will change over timee as th sample equilibrates with the carbon dioxide in the atmosphere. Although other water quality parameters do not interfere with pH, measurements can be impacted by electrical interferences (e.g., electrical noise), since the electrode has a high impedance (Evans et al. 2013b). Table 2‐4. Typical Filter Influent Ranges of General Water Quality Parameters. Analyte Typical Filter Influent Ranges Temperature 5°C to 30°C

pH 6.0 to 9.0 SU

Influent turbidity varies with source water and season. Turbidity typically ranges from Turbidity 0.01 to 1.5 NTU in biofilter influent (Evans et al. 2013b).

2.3.2 Biodegradable Organic Matter (BOM) Biodegradable organic matter (BOM) is the fraction of NOM which serves as the primary substrate for heterotrophic microorganisms (those organisms which utilize organic carbon as a carbon source). If BOM is not sufficiently removed through treatment, it may act as a precursor for DBPs and decrease treated water biostability, fueling regrowth of bacteria in the distribution system. Effective removal of BOM through biofiltration is important for mitigating DBP formation and enhancing biostability. While organic compounds are an important parameter for assessing performance of biofilters, not all are practical for use (Evans et al. 2013a). AOC and BDOC assays can be used to characterize the biodegradable fraction of carbon present. While these tools may be useful for research studies and evaluating longer‐term trends, they are not practical for shorter‐term operational decisions due to their labor‐intensive analysis requirements, cost, lack of sensitivity and lengthy turnaround time (Evans et al. 2013a). Table 2‐5 summarizes typical NOM analyses for biofiltration monitoring.

Biofiltration Guidance Manual for Drinking Water Facilities 21 2.3.2.1 Bulk Carbon Indicators Total organic carbon (TOC) is the sum of particulate and dissolved organic carbon (DOC) in a water sample. DOC is the fraction of TOC that passes through a 0.45 µm filter and represents approximately 80 to 95% of the TOC by mass (Crittenden et al. 2012) ind settle water. Glass fiber filters should be used to analyze for DOC to avoid TOC contamination. DOC removal can be used for assessing bioactivity as the removal of primary substrate (carbon) is a direct indicator of removal from biological processes (e.g., biodegradation, assimilation into biomass, and sorption into biomass). Since particulate organics may be removed via physical filtration, TOC removal is not an appropriate tool for assessing biological removal processes. DOC is also often cited as the primary reason for use of biofiltration, as it pertains to control of DBP formation, biostability and disinfectant residual stability in distribution systems (Brown et al. 2016, Evans et al. 2013a, LeChevallier et al. 2015).

Ultraviolet (UV) absorbance at 254 nm (nanometers; UV254) is used to quantify the UV‐absorbing NOM, including humic and fulvic acids with aromatic rings, in water samples using Standard Method 5910B or SM 9510 with a spectrophotometer (APHA et al. 2012). UV254 is often used as a surrogate measure for NOM due to its quick, relatively simple analysis, and its ability to be measured in‐house (Crittenden et al. 2012, Nyfennegger et al. 2016). SUVA is calculated by dividing UV254 in a 1‐centimeter (cm) cell pathlength by the DOC concentration (mg/L) and multiplying this value by 100 (EPA method 415.3) (Equations 2‐5 and 2‐6):

𝑆𝑈𝑉𝐴 100 Equation 2‐5

where

SUVA = Specific UV absorbance DOC = dissolved organic carbon cm = centimeters m = meters 𝑈𝑉𝐴 𝑐𝑚 Equation 2‐6

where

UVA = UV absorbance of the sample A = UV absorbance at 254 nm d = The quartz cell path length in cm.

Typically, higher values of SUVA indicate a higher fraction of NOM with complex structures such as aromatic rings (Evans et al. 2013a). SUVAd is use for evaluating coagulation efficacy, where values less than 2 L/mg‐m are indicative of lower DOC removal potential with coagulation due to the presence of hydrophilic, non‐humic matter, whereas SUVA greater than 2 L/mg‐m tends to be easier to coagulate due to the presence of hydrophobic humic matter. Higher SUVA values lead to higher chlorine demands and increased TTHMs (EPA 2012). Correlations can often be made between UV254, TOC, and/or DOC although correlations are site‐specific, and can be impacted by source water quality including NOM characteristics, pre‐treatment (e.g., coagulation/sedimentation and pre‐oxidation) and biological activity (Evans et al. 2013a). UV254 has also been shown to correlate with DBP formation, though the degree of formation is also dependent on chlorine dose, temperature, pH, and time (Archer and Singer 2006, Kingsbury 2010). DBPs are a result of a reaction between NOM and an oxidant, and many are regulated (or being considered for regulation) by the Environmental Protection Agency (EPA) as they are suspected carcinogens (EPA 2012). Some DBPs that are not regulated by the EPA are regulated in certain states (e.g., N‐Nitrosodimethylamine [NDMA]) (EPA 2012). DBP formation potential (DBP‐FP) measurements

22 The Water Research Foundation may be conducted by using a standard residence time, chlorine residual, pH, and temperature or simulate the distribution system at a utility’s maximum distribution system hydraulic residence time, chlorine residual, and temperature. This testing is often used to inform whether treatment is sufficient to reduce the formation of DBPs. DBP concentrations are measured using SM 5710, SM 6232B, SM 6232C, or SM 6200 (APHA et al. 2012, Evans et al. 2013a, 2013b). It is important that laboratories take caution to use organic‐free bottle‐ware and use advanced cleaning procedures when reusing glassware. DBP‐FP can be measured at the filter influent and effluent to assess removal efficacy or just the filter effluent as an indication of how effectively biofiltration removes organic compounds that can form DBPs. Typical removals across biofilters are shown in Table 2‐5. 2.3.2.2 Biodegradable Dissolved Organic Carbon (BDOC) BDOC is the fraction of DOC that is degradable by microorganisms and AOC is the fraction of DOC that is consumed by bacteria to produce biomass. BDOC is calculated as the change in DOC after exposure to biomass over a given duration. Several methods have been developed, including biomass attached to sand (Joret and Levi 1986, Joret et al. 1991, Volk et al. 1994), batch reactors with suspended biomass (Lim et al. 2008, Servais et al. 1987), a recirculating column with biomass attached to media (Mogren et al. 1990, Terry et al. 2019), single‐pass column with biomass attached to media (Frías et al. 1992) and a shaken batch reactor (Allegeier et al. 1996). The sand method is commonly used and has a detection limit of approximately 0.2 mg/L (Escobar and Randall 2001). The turnaround time for BDOC by an outside lab is approximately three weeks, but there are few options available besides university labs. Two common methods for AOC analysis are SM 9217 (APHA et al. 2012) and the bioluminescent method (Haddix et al. 2004, Weinrich et al. 2009), both of which have detection limits of 10 ppb acetate‐C and holding times ranging from two to seven days post‐pasteurization. Samples for AOC analysis is typically sent to an outside commercial laboratory. Both methods measure the growth of two strains of oligotrophic bacteria: Pseudomonas P17 and Spirillum NOX. The standard method measures their growth through plate counting on culture media and the bioluminescent method measures growth through luminescence emitted from bacteria degrading carbon. The turnaround time for SM 9217 is lengthy, typically three to four weeks. Compared to SM 9217, the bioluminescent method has a shorter turnaround time of approximately one week. Carboxylic acids and aldehydes are small chain organic compounds generated from upstream oxidation processes such as ozonation (Evans et al. 2013a, Peldszus et al. 1996a, Peldszus et al. 1996b). Systems that do not use pre‐oxidation prior to filtration typically have low carboxylic acid and aldehyde concentrations. Carboxylic acids are typically measured by a commercial laboratory using a modified version of EPA 300.1 (Peldszus et al. 1996a, 1996b). The carboxylic acids including acetate, formate and oxalate are measured at low concentrations, e.g., detection limits of 5 to 10 parts per billion (ppb). The sum of these acids, expressed as micrograms of carbon per liter (µg‐C/L), have been shown to correlate with AOC at higher concentrations (e.g., greater than 400 µg‐C/L) (Evans et al. 2013a). Analysis of carboxylic acids is advantageous over AOC because carboxylic acids are directly measured via ion chromatography (IC), whereas AOC uses a bioassay which relies on the growth of two organisms (Pseudomonas fluorescens P17 and Spirillum NOX) in a water sample. Carboxylic acids analysis has a higher reproducibility and is less expensive thanC. AO Aldehydes can be measured using EPA Method 8315A. This method uses high performance liquid chromatography (HPLC) and UV/vis to quantify 13 carbonyl compounds found in aqueous samples. Detection limits range from 5.9 to 110.2 µg/L, depending on the compound (EPA 1996). In surface water treatment where ozonation is applied, aldehydes have been shown to correlate with high TOC levels and background organic matter (Glaze et al. 1989a).

Biofiltration Guidance Manual for Drinking Water Facilities 23 Table 2‐5. Typical Influent Concentrations and Removal Ranges of Different Components of NOM. Analyte Typical Biofilter Influent Concentrations and Removal Ranges Total organic carbon Raw water TOC typically ranges from 1 to 20 mg/L in surface waters, and 0.1 to 2 mg/L in (TOC) groundwater (Crittenden et al. 2012, Nyfennegger et al. 2016). Ranges of biofilter influent TOC typically range from 0.3 to 4 mg/L. Removals at 17 full scale facilities ranged from 0 to SM 5310 1.6 mg/L (Evans et al. 2013a). DOC is typically comprised of 80 to 95% of the TOC (Crittenden et al. 2012). Average raw Dissolved organic water influent DOC has been reported to range from 0.5 to 3.7 mg/L in surface waters but carbon (DOC) can exceed 15 mg/L (Evans et al. 2013a). Typical DOC removal in biofiltration processes is 10 to 20% but varies depending on several factors such as media type, temperature, EBCT, SM 5310 and degree of pre‐oxidation. Concentrations ranging from 0 to 1.4 mg/L have been reported removed using biofiltration at 17 full scale facilities (Evans et al. 2013b). Ultraviolet Influent level varies with source water and season. Biofilter influent UV254 ranged from 0.01 absorbance at 254 nm to 0.08 absorbance units per centimeter (abs/cm) at 17 full scale biofilters (Evans et al. (UV ) and specific 254 2013b). Effluent UV254 varies with filter design. In some cases, UV254 correlates with DOC and ultraviolet absorbance thus UV254 could be used as a surrogate for DOC, if correlated through testing. UV254 has also NOM (SUVA) been shown to correlate with DBP‐FP for some water types (Archer and Singer 2006,

Bulk SM 5910 Kingsbury 2010), which could be investigated if DBPs are of concern. DBP‐FP is a standard method to characterize the “worst case” potential for DBPs to form after disinfection via chlorination. Simulated distribution system (SDS) testing is an alternative method (SM 5710D) to determine DBP formation at representative site‐specific doses, pH and residencee times. Th formation of DBPs in the distribution system are Disinfection impacted by chlorination strategies (dose, CT) and may increase with higher or more frequent Byproduct Formation use of chlorine during treatment. Biofiltration has been shown to remove carbonaceous and Potential (DBP‐FP) nitrogenous DBP precursors. If the formation of DBPs is a treatment objective for SM 5710B biofiltration, DBP‐FP can be either be measured at the filter influent and effluent, or just the filter effluent. Studies have indicated removal of total trihalomethane formation potential (TTHM‐FP) across a biofilter ranged from 13 to 84%, and haloacetic acid formation potential (HAA5‐FP) removal ranged from 6% to 75% (Arnold et al. 2018, Chaiket et al. 2002, Evans et al. 2013b, Farré et al. 2011, Hooper et al. 2019, Selbes et al. 2016). Carboxylic acids are measured as acetate, formate, and oxalate. Pyruvate is typically below or near detection. The sum of carboxylic acids expressed as carbon equivalents (mg‐C/L) has been shown to correlate with pre‐oxidant typed an dose and AOC (Evans et al. 2013a, Carboxylic acids and Lauderdale et al. 2018). Biofilter influent concentrations have been reported to range from aldehydes 2 to 183 µg‐C/L with removal rates up to 171 µg‐C/L (Evans et al. 2013a). No SM available for Formaldehyde, acetaldehyde, glyoxal, and methyl‐glyoxal are aldehydes that typically appear carboxylic acids and most in ozonated drinking water (Cervi 1996), which has been shown to correlate with ozone SM 8315A for dose, temperature, pH, DOC, and NOM (Andrews and Huck 1994, Glaze et al. 1989a, 1989b; aldehydes Najm and Krasner 1995). Weinberg et al. (1993) reported influent concentrations ranging from 10 to 147 µg/L across 13 sites with removal of up to approximately 115 µg/L in GAC

BOM biofilters.

Assimilable organic AOC has been shown to range from 10 to 1,440 µg‐C/L in biofilter influent and removal up to carbon (AOC) 220 µg‐C/L and an average AOC/TOC of 4.6% and an average BDOC/TOC of 20% have been SM 9217 achieved using biofiltration (Evans et al. 2013b, Terry and Summers 2018). Biodegradable dissolved organic BDOC has been shown to range from 0.2 to 1.1 mg/L in biofilter effluent, with removal of up carbon (BDOC) to 0.5 mg/L. No SM available

24 The Water Research Foundation 2.3.3 Inorganic Compounds Iron and manganese are frequently monitored nutrients in drinking water due to their nuisance effects on downstream household plumbing fixtures and aesthetic impacts on taste and color. Their presence can also promote the growth of bacteria and biofilms further enhancing water quality issues downstream of the treatment plant. Removal of both of these constituents primarily occurs through oxidation to an insoluble state ahead of (during pre‐treatment) or during filtration. If particulate, the filter media is then able to capture the particulate iron and/or manganese oxides. Commonly, plants with targeted removal rates of iron and manganese will dose a chemical oxidant such as sodium permanganate, chlorine, or ozone upstream of the filters (Kohl and Medlar 2006, Lauderdale et al. 2016). However, biological removal of iron and manganese can occur if the filters have active iron‐ and manganese‐oxidizing bacteria present on the filter media (Kohl and Dixon 2012) and the iron and manganese are in the dissolved form. Monitoring for influent and effluent particulate and dissolved iron and manganese is important to assess performance if these constituents are of concern. Oxidation‐ reduction potential (ORP) monitoring and pH are important water quality parameters as they control the speciation as soluble or particulate forms (Brown et al. 2016, Tremblay et al. 1998). Iron can be evaluated using spectrometric (SM 3111B and C), inductively coupled plasma (ICP) methods (SM 3120), or more simply using a colorimetric procedure (SM 3500‐Fe) with a detection limit of 0.01 mg/L. Similarly, manganese concentrations can be evaluated using spectrometric methods (SM 3111B and C) or more simply using a colorimetric method (SM 3500‐Mn) with a detection limit ranging from 0.042 to 0.21 mg/L (APHA et al. 2012). While iron and manganese concentrations are highly dependent on source water quality, influent iron may range between 0.013 to 1.55 mg/L and influent manganese between 0.002 to 1.28 mg/L (Brown et al. 2016). While recent studies have shown that manganese removal in full‐scale biofiltration can range 25% to greater than 80%, iron removal through biofiltration has not been as thoroughly documented. 2.3.4 Nutrients Macronutrients such as carbon, nitrogen, and phosphate are essential to microbial respiration and growth. A molar ratio of 100:10:1 of bioavailable carbon (C) to nitrogen (N) to ortho‐phosphate (P) at the biofilter influent is often cited as necessary to maintain and promote biomass production (Figure 2‐5;(Lauderdale et al. 2014, 2011, LeChevallier et al. 1991). The carbon concentration for this ratio is based on DOC consumed from the biofilter influent to effluent. Nitrogen can exist as ammonium, nitrate, nitrite and organic nitrogen, with ammonium being the most energetically favorable for assimilation. Ammonium and organic nitrogen are of interest as they can be an electron donor (energy source) for bacteria. Complete oxidation of ammonia to nitrate through nitrification is common in aerobic biofilters and removal of nitrate may occur through denitrification processes in anaerobic biofilters (Li et al. 2010). The micronutrients shown in Figure 2‐5 are also important for biological growth; they may act as cofactors for enzymes or perform specific cellular functions (Vaccari et al. 2006). Evans et al. (2016) found that supplementation with micronutrients at the ratios shown in Figure 2‐6 enhanced BDOC removal in batch experiments, as metals can be limiting cofactors to enzymes important for specific cellular functions or degradation of specific contaminants. However, the implications for full‐scale systems warrants additional research and validation.

Biofiltration Guidance Manual for Drinking Water Facilities 25 100.0% 50.0% 20.0% 14.0% 8.0% 10.0% Weight 3.0% Dry

1.8% 1.0% 1.0% Cell

1.0% 0.5% 0.5% by

0.2%

Percent 0.1%

Figure 2‐5. Microbial Biomass Macro and Micro Nutrient Composition, Based on E. coli. Source: Vaccari et al. 2006. Typical filter influent and removal ranges for nutrients are shown in Table 2‐6. Phosphate should be monitored as orthophosphate, which is the bioavailable reactive fraction. Coagulation and flocculation can bind phosphorus in flocs and cause low concentrations at the biofilter influent. A deficiency in the ratio of phosphorus to carbon has been shown in some studies to result in excess EPS production, which may impact filter hydraulics (Keithley and Kirisits 2019, Lauderdale et al. 2018, Lauderdale et al. 2011b). If nutrients are limiting, they can be added in the form of inorganic nitrogen (e.g., ammonium) and/or phosphate (e.g., orthophosphate, phosphoric acid) (Lauderdale et al. 2018). Changes in nutrient concentrations between the filter influent and effluent is dependent on the concentration of other macronutrients, and nutrient cycling may occur within biofilms which may complicate mass balance calculations. Monitoring nutrients is not recommended during normal routine operations. However, nutrient data collected monthly may be useful for troubleshooting or optimization studies to improve filter hydraulics.

Table 2‐6. Typical Influent Concentrations and Removal Ranges of Nutrients. Analyte Typical Influent Concentrations and Biofilter Removal Ranges Orthophosphate is evaluated using EPA Method 365.1 or EPA Method 300 with a detection limit of 0.05 to 20 mg/L, depending on the method used. Phosphate has a holding time of two days Ortho‐ using EPA method 300.1 and 28 days for EPA method 365.1. Influent total phosphate has been phosphate shown to range from <0.05 to 5.8 mg/L at 14 full‐scale facilities (Evans et al. 2013b). Ortho‐ phosphate has been shown to range from <0.003 mg/L to 0.16 mg/L (Nyfennegger et al. 2016). Ammonium is evaluated using EPA Method 350 or SM 4500‐NH3D with a detection limit of approximately 0.03 mg‐N/L (APHA et al. 2012). The holding time is 28 days. Nitrification of Ammonium influent ammonia ranging from <0.05 to 1.5 mg‐N/L resulted in 73 to 98% conversion to nitrate (Lauderdale et al. 2018). Nitrate can be evaluated using EPA Method 300 or SM 4500‐NO3 E with a detection limit of approximately 0.05 mg‐N/L (APHA et al. 2012). Nitrite is measured using EPA Method 300 or SM 4500‐NO2 B with a detection limit of 0.005 mg‐N/L (APHA et al. 2012). Nitrate and nitrite Nitrate and should be analyzed within 24 hours. Nitrate typically ranges from 0.06 to 0.67 mg‐N/L and nitrite nitrite is typically low, at or near detection (Nyfennegger et al. 2016). Biofilter influent nitrate has been reported to range from 0.06 to 0.67 mg‐N/L, and nitrite concentrations are typically at or near detection (Nyfennegger et al. 2016).

26 The Water Research Foundation 2.4 Biological Monitoring Biological process monitoring is important for decoupling the biological component from chemical and physical processes occurring within the filter. Understanding the relative contribution of biological degradation can provide a reference point for performance and aid in determining whether further enhancement of bioactivity is warranted or possible. The biological process monitoring toolsd presente in this chapter are readily available for immediate use and have been shown to provide reliable data, be cost‐effective, and operator‐friendly (Hooper et al. 2019). These tools include biofilm formation rate, DO consumption and ATP. Other biological monitoring tools (e.g., EPS, microbial community characterization) are presented for reference herein, but are recommended for specialty (e.g., non‐ routine) monitoring. 2.4.1 Biofilm Formation Rate Biofilm formation rate is an in‐line monitoring tool used to measure bioactivity as the quantity of biomass that accumulates on a given surface area over a given time. The biofilm formation rate is measured by installing a new coupon into a pipe loop where water flows across the coupon surface at a controlled flow rate. The coupons are held in place for a specified period of time, during which naturally occurring (indigenous) organisms attach, form a biofilm, and grow on the surface (see Figure 2‐6). The coupons are then collected and the quantity of ATP on the coupon surface measured. ATP is a bioenergy molecule used in all living cells for energy transfer further discussed in Section 2.4.3 (Evans et al. 2013a, 2013b, Lauderdale et al. 2011, Magic‐Knezev and van der Kooij 2004, Pharand et al. 2014, Velten et al. 2007). Figure 2‐7 shows an example pipe loop installation, coupon collection, and an example of ATP analysis; Figure 2‐8 shows an example process flow diagram schematic. Once the ATP concentration is measured, the biofilm formation rate is calculated as the mass of ATP divided by the coupon surface area and the days of incubation, as pg ATP/mm2/d. When measured at the filter influent, biofilm formation rate is an indicator of the biological growth potential on filter media. The amount of biomass that can accumulate on the coupon is directly related to water quality and is affected by multiple factors including temperature, quantity and quality of organic carbon,e pH, and th presence of disinfectant residual. It is important to note that if an oxidant residual is present the pipe loop should either be installed prior to oxidant addition or it should be quenched prior to the biofilm formation rate pipe loop. Alternatively, for GAC media where oxidant residual may be quenched in the upper portion of the filter bed, the biofilm formation rate pipe loop could be connected to a sample tap from within the filter profile. This tool may not be useful in pre‐ chlorinated systems as the presence of chlorine in the influent would considerably impact the biofilm growth on the coupon in both the upstream and downstream samples. When measured at the filter effluent, biofilm formation rate is an indicator of the biological stability (e.g., regrowth potential) of treated water. The change in biofilm formation rate is a direct measure of the filter biological activity. The biofilm formation rate may also be measured at the finished water point of entry into the distribution system or within the distribution system as an indicator of biostability (LeChevallier et al. 2015). The finished water and distribution system biofilm formation rate should be at least an order of magnitude lower in the biofilter effluent than the influent, and, after disinfection, be below 0.09 pg ATP/mm2/d for distribution systems with free chlorine and 0.017 pg ATP/mm2/d for chloraminated distribution systems to maintain disinfectant residual stability (LeChevallier et al. 2015).

Biofiltration Guidance Manual for Drinking Water Facilities 27 Hooper et al. (2019) tested a two‐week and one‐month incubation period for the coupons and determined that two weeks was sufficient to attain an established biofilm. Shorter incubation periods may be possible based on site‐specific conditions. A one‐log reduction in biofilm formation rate or greater across thes biofilter should be expected. If lower removal is observed, optimization strategies for enhancing biofilter performance should be explored. The biofilm formation rate is impacted by temperature, where significant increases were observed at temperatures greater than 15oC (Hooper et al. 2019).

Figure 2‐6. Biofilm Attachment and Growth on Coupons in a Pipe Loop. Source: Hooper et al. 2019.

Figure 2‐7. Biofilm Formation Rate Pipe Loop Example (a) installation, (b) Coupon Harvesting and (c) ATP Analysis. Source: Hooper et al. 2019.

28 The Water Research Foundation

Figure 2‐8. Biofilm Formation Rate Pipe Loop Example Schematic. Source: Hooper et al. 2019.

2.4.2 Dissolved Oxygen Consumption Dissolved oxygen (DO) consumption is a measure of biological activity, because oxygen is consumed during cellular respiration. Aerobic biofilters contain primarily heterotrophic bacteria, which use organic compounds as a substrate for growth. When organic carbon is degraded, DO is consumed and carbon dioxide (CO2), water, and energy are released (Figure 2‐9). The released energy is directly related to biological activity which can be stored into chemical potential energy in the form of ATP. However, other inorganic electron donors typically present in lower concentrations such as ammonium, may also be used as a substrate for growth and consume oxygen. DO is monitored at the filter influent and effluent and the DO consumption is calculated as the difference between the influent and effluent concentrations. DO consumption can also be used in conjunction with the biofilm formation rate by calculating the volumetric DO consumption rate as another indicator of biological activity. The volumetric DO consumption rate is calculated by normalizing the DO consumption by the EBCT (units of mg/L/d). The theoretical oxidant demand for carbon utilization has been shown to correlate with DO consumption in‐line with stoichiometric requirements for glucose removal, where one mole of carbon is consumed per mole of O2 (Hooper et al. 2019). Other reactions can occur within the biofilter that consume oxygen such as nitrification, but for aerobic biological filters the primary mechanism is heterotrophic biodegradation. If significant nitrification is expected, then the theoretical oxidant demand from nitrification needs to be included with the demand from carbon removal. Hooper et al. (2019) and Evans et al. (2013b) demonstrated DO consumption was a highly effective and responsive online tool for evaluating biological activity. However, for facilities with pre‐ozone,

Biofiltration Guidance Manual for Drinking Water Facilities 29 supersaturated concentrations of DO and air binding within the filter media can result in off‐gassing due to physical equilibration rather than biological consumption. Thus, DO consumption should only be used for facilities without ozone upstream of biofiltration.

Figure 2‐9. DO Consumption by Bacteria During Heterotrophic Cellular Respiration. Source: Hooper et al. 2019.

DO at the filter influent, and effluent should be monitored using an online luminescent DO (LDO) optical process probes. The LDO probes are preferable over membrane‐based probes, because oxygen is not consumed during measurement and they require less maintenance. Membrane‐based probes have significant issues with membrane fouling and sensor drift. For LDO probes, DO is measured using a sensor coated with an oxygen‐sensitive luminescent dye. Temperature is compensated for in the DO reading and temperature readings are recorded by a temperature sensor. The limit of detection for DO consumption is approximately 0.1 mg/L. Figure 2‐10 shows the probe installation methods, which can be pole‐mounted or placed in a flow‐ through cell. Placement of the instruments is important to collect representative data. The filter influent probe should be installed as close as possible to the filter influent, preferably in a stilling well directly above the filter media, at a location that is in equilibrium with the atmosphere and downstream of a weir or turbulent process that could release DO. If that is not possible, a flow‐through cell configuration may be installed. For probes installed in a flow‐through cell, the flow should be tested and confirmed weekly in the flow‐through cell and data should be removed if there are periods with no flow. If algae are expected, clean more frequently to prevent build up on the sensor cap and log when periods of algae occur in the raw water. The LDO probes must be connected to a controller to log online readings. Once installed, the facility elevation must be entered, and the probe should be calibrated in air‐ saturated water prior to use. The logging frequency should be hourly, and data should be censored for periods when the filter is offline, on standby, or backwashing.

30 The Water Research Foundation

Figure 2‐10. DO Probe Installation Methods as a (a) Flow Through Cell or (b) Stilling Well. Source: Hooper et al. 2019.

2.4.3 Adenosine Triphosphate (ATP) ATP is the highest energy state of the primary energy storage molecule used by all living organisms and is an indication of active biomass. ATP analysis is conducted using commercially available test kits. These kits offer a significantly less cumbersome alternative to time‐consuming tests such as HPCs. ATP should be monitored on the filter media, near the top of the filter bed, rather than aqueous grab samples. Filter media samples provide the quantity of sessile bacteria present on the biofilm who are responsible for contaminant degradation (Evans et al. 2013a). Detachment of biomass is a normal process that occurs on biofilters at an irregular frequency; filter effluent ATP data have a high degree of variability and are not recommended for regular monitoring. For media samples, biofilm on a filter media sample reacts with reagents which lyse the cells to release ATP into solution. The released ATP then reacts with another reagent, the luciferase enzyme, to produce light. The light released during this reaction is proportional to cell count and/or cell mass. Typical concentrations of ATP in a microbial cells range between 2 × 10‐5 to 4 × 10‐4 pg ATP per cell (Magic‐ Knezev and van der Kooij 2004, Velten et al. 2007, Velten et al. 2011). The sample is placed into a luminometer and the relative luminescence units (RLUs) are detected using a luminometer. An example of the ATP test kit protocol for the LuminUltra DSA™ test kit is provided in Figure 2‐11. Analysis takes less than 15 minutes. The lower detection limit is approximately 1 pg ATP/g filter media. Results should be normalized to the dry weight of filter media rather than wet weight, particularly for GAC media which can retain moisture in micropores. The dry weight should be determined by taking a 1 gram an aliquot of the sample collected for ATP analysis and dried separately at 180oC.

Biofiltration Guidance Manual for Drinking Water Facilities 31

Figure 2‐11. ATP Test Kit Protocol for Filter Media Samples. Source: Hooper et al. 2019. ATP on filter media tends to be relatively stable over time (e.g., less than an order of magnitude change over time) and does not correlate with other process or water quality parameters. ATP monitoring should be limited to start‐up to confirm biomass acclimation on new media, during filter conversion from conventional to biofiltration, during/after filter shutdowns, or troubleshooting. Inhibitory impacts to biomass from process changes, such as the presence of a disinfectant residual or low DO concentrations, can cause decreases in ATP. Thus, monitoring ATP can be a helpful for assessing impacts to bioprocesses during troubleshooting. Periodic samples (e.g., semi‐annually) can be included to provide a baseline for comparison during troubleshooting. The biofilm formation rate is a superior tool for monitoring biomass acclimation. During acclimation, the difference between the biofilm formation rate at the filter influent and effluent would be monitored over time. The biofilm formation rate removal would be initially low and increase over time until the biomass is fully acclimated. 2.4.4 Extracellular Polymeric Substances (EPS) EPS are excreted by bacteria and used to attach to filter media grains and create the matrix that supports the biofilm (Wingender et al. 1999). Figure 2‐12a and b shows stalked bacteria, biofilm filaments as well as cocci and rod‐shaped bacteria from a scanning electron micrograph (Evans et al. 2013a). The mass composition of biofilms is typically 90% extracellular substances and 10% microorganism. However, the composition can be influenced by factors such as shear stress, temperature, and nutrients (Flemming and Wingender 2010, Keithley and Kirisits 2018). There are multiple constituents which comprise EPS, including polysaccharides, proteins, and deoxyribonucleic acid (DNA). EPS is excreted by microorganisms to aid in the capture of substrate and nutrients, as well as protect against oxidative stress (Laspidou and Rittmann 2002, Wingender et al. 1999). EPS may be produced or consumed by microorganisms as needed. Utilization of the polysaccharide (carbohydrate) fraction occurs rapidly utilized, while the protein fraction may also serve as an energy source. Proteins are typically the greatest fraction of EPS on a mass basis (Zhang and Bishop 2003, Keithly and Kirisits 2018, Hooper et al. 2019).

a. b.

cocci

rods

stalked bacteria/ biofilm filament

Figure 2‐12. Scanning Electron Micrograph on a Sand (a) and Biologically Active Carbon (BAC) (b) Biofilter at Greater Cincinnati Water Works. Source: Evans et al. 2013a.

32 The Water Research Foundation Several methods for EPS extraction have been developed which are documented elsewhere (Zhang and Bishop 2003, Keithley and Kirisits 2015, Keithley and Kirisits 2018). EPS extraction should be conducted a minimum of five times prior to analysis. After EPS extraction, the protein fraction is recommended for monitoring as it is the dominant fraction by mass and is easy to measure with a colorimetric test kit, the modified Lowery Assay using the Pierce Bicinchoninic acid (BCA) protein kit and bovine serum albumin (BSA) standard. Quantification of carbohydrates is also not practical for use for most utilities because a hazardous volatile reagent, phenol, is required. Routine sampling during normal operations is not recommended, because EPS concentrations do not necessarily correlate with filter hydraulic parameters. However, baseline sampling may be helpful for comparison if hydraulic performance becomes compromised. Conditions that stress the microbial community, such as disinfectant residual or a deficiency of nutrients, may result in excess production of EPS. Excess production of EPS may cause issues with filter hydraulics, as pore space between filter media can become clogged and trap particulate more rapidly (Le Bihan and Lessard 2000). This concept is illustrated in Figure 2‐13.

Figure 2‐13. EPS Formation and Interference with Particle Collection in Biofilter Media. Source: Hooper et al. 2019.

Biofiltration Guidance Manual for Drinking Water Facilities 33 EPS monitoring can be used to determine if microbial stress may be a cause of excessive headloss. If a facility is experiencing issues with filter hydraulics, EPS sampling should be conducted every other week for troubleshooting purposes. EPS samples should be collected at the end of filter run, prior to backwashing. The percent pore space should be calculated to determine if EPS filter clogging is a cause (Equations 2‐7 and 2‐8).

𝑉𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝐸𝑃𝑆 𝑝𝑒𝑟 𝑔𝑟𝑎𝑚 𝑀𝑎𝑠𝑠 𝑜𝑓 𝐸𝑃𝑆 𝐸𝑃𝑆 𝐷𝑒𝑛𝑠𝑖𝑡𝑦 130 Equation 2‐7

% Pore Space as EPS = 𝑉𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝐸𝑃𝑆 𝑝𝑒𝑟 𝑔𝑟𝑎𝑚 𝑃𝑜𝑟𝑒 𝑆𝑝𝑎𝑐𝑒 𝑉𝑜𝑙𝑢𝑚𝑒 × 100% Equation 2‐8

The acceptable percent of pore space occupied by EPS may vary by utility and should be evaluated on a site‐specific basis. General guidance is to not exceed 5 to 7% of the pore space as EPS. EPS has been shown to reduce biofilter pore space by as much as %,7 to 13 whereas particle deposition reduced the pore space by no more than 7% (Mauclaire et al. 2004). Nutrient augmentation has been shown in some studies to lower EPS and mitigate issues with excess headloss on biofilters (Lauderdale et al. 2018, 2011). Hooper et al. (2019) did not observe correlations between headloss and EPS, but this may have been due to relatively low percent occupation of EPS in the pore space of the biofilters evaluated for this study. 2.4.5 Microbial Community Analysis Microbial communities are ubiquitous and often are beneficial because of their ability to biodegrade contaminants and remove organic matter (Li et al. 2017). Microbial community characterization is a method of biological forensics that identifies organisms present and typically requires a specialized laboratory familiar with preparing and analyzing DNA. The methods required, such as pyrosequencing, next generation sequencing, and reverse transcription, quantitative polymerase chain reaction (RT‐ qPCR) can be costly and have relatively long turnaround times. If utilities have capability to run these advanced microbiological techniques in‐house then turnaround times may be shorter. Generally, analysis involves DNA extraction from filter media biomass, amplification of a specific gene region, and sequencing to generate a gene library of microorganisms. This process is typically carried out at a specialized lab equipped to handle sensitive genomic techniques and complex statistical analysis. Sequencing results typically contain phylum to genus‐level information. For example, for the bacterium Escherichia coli, the phylum would be Proteobacteria (a major group of gram‐negative bacteria) and the genus would be Escherichia. Species‐level data are typically not included, as this level of identification requires significantly more effort and cost. Microbial community characterization in biofilters is still being explored to classify organisms responsible for key functions and understand implications on treatment (Kirisits et al. 2019, Steele et al. 2005). An evaluation of the microbial community structure and diversity can indicate differences in system resiliency (Li et al. 2017). The Simpson’s Index and Shannon Weaver Index are commonly used for evaluating biodiversity in microbial communities. However, unclassified or unidentifiable organisms may represent a significant fraction of the microbial community (Chouari et al. 2014, Gimbel et al. 2006). Functional gene analysis such as expression of specific enzymes used for contaminant degradation, could provide more useful data than the phylum or genus‐level identification. Studies have shown that filter design characteristics, such as the filter media type and depth, can have an impact on the microbial community. When GAC was compared to sand biofilters, the microbial communities were found to have different phylogenetic structures, functions and colonization patterns, indicating the importance of environmental conditions (Oh et al. 2018). With increasing filter depth there is a decrease in biomass due to microbial utilization of NOM (Moll et al. 1998). Depth‐based sampling can also elucidate at what EBCT specific biodegradation pathways occur (Khammar et al. 2005).

34 The Water Research Foundation Filter operations also have a significant effect on the microbial community. Disinfectants in backwash water can vastly impact the microbial community in the filter. The primary reason it is considered for biofiltration backwash operations is to control biomass production to minimize headloss, underdrain clogging, and to maintain overall uniformity in operation conditions (Brown et al. 2016). However, higher disinfectant residuals (e.g., 0.5 to 3 mg/L, depending on disinfectant demand and filter media type) may have detrimental impacts on the biological community and can lead to disinfectant resistance of colonized organisms (Lauderdale et al. 2011, Li et al. 2017, Pinto et al. 2012). This is of particular significance when considering opportunistic pathogens that already exhibit chlorine resistance, such as Pseudomonas, Mycobacterium, and Legionella, which have the potential to be selected for in a chlorinated filter, though this is not yet well understood (Falkinham et al. 2015, Zanacic et al. 2017). Additionally, the potential for seeding oxidant‐resistant bacteria from upstream unit processes needs to be recognized (Kotlarz et al. 2018). Microbial community analysis as a monitoring tool still has several hurdles, including lack of precision on field replicates, variability in results between laboratories, high degree of interpretation, and cost. These must be addressed before it can be implemented widely for utility biofilter monitoring. For example, Hooper et al. (2019) found field replicate samples analyzed by the same lab had a difference of up to 10% relative abundance of genus identified. Split samples sent to different laboratories had a variability of up to 30% relative abundance. This discrepancy was likely attributable to variances in the DNA extraction or amplification process. It is recommended that monitoring of microbial communities be limited to special research studies and should include multiple replicates (duplicate or triplicate samples) to evaluate precision. For these reasons, microbial community analysis is only recommended for research investigations, not for regular monitoring. 2.4.6 Other Biological Monitoring Tools Additional microbial tools can be used for characterizing microbial growth, diversity, and biomass. The following provides a brief overview of methods that are not recommended for immediate use but have been incorporated in some research studies. Enzymes are essential for microbial function and can be linked to specific cellular pathways, such as metabolism, microbial stress, and gene expression. Enzyme assays are analyzed using simple test kits. Adenosine monophosphate (AMP) and adenosine diphosphate (ADP) is also analyzed with a test kit and may be paired with ATP data to evaluate bioactivity, though additional validation testing is warranted. Phospholipid fatty acids (PLFA) and HPC are unpractical and highly labor‐intensive methods for evaluating biomass on filter media. HPC requires suspension of biomass from the filter media and enumeration on a culture medium at a set temperature and incubation time (Allen et al. 2004). Flow cytometry has been used in limited research studies to quantify planktonic bacteria in filter effluent water. A summary of these parameters and the reasons why they are not recommended for regular monitoring are provided in Table 2‐7.

Biofiltration Guidance Manual for Drinking Water Facilities 35 Table 2‐7. Other Biological Monitoring Tools Not Recommended for Regular Monitoring. Parameter Description Recommendations/Data Gaps Enzyme Activity Hydrolase is a class of household enzymes used by Hydrolase enzyme activity provides similar data most bacteria, which have been investigated as a as ATP on filter media (R2=0.78). Analysis is quantification tool for biological activity. Specific rapid but does not measure all types of cells hydrolase enzymes that have also been investigated (Evans et al. 2013a, Mauclaire et al. 2004, include esterase, phosphatase (PHO), and glycosidase Seredyńska‐Sobecka et al. 2006). (GLY). Esterase breaks down esters into acids and alcohols and is indicative of heterotrophic activity Esterase enzyme activity provides similar results (Nybroe et al. 1992). PHO is used in phosphate as total ATP (R2=0.94‐0.95) and analysis is rapid acquisition, an important component of cellular but has been limited to research studies (Berney metabolism and structure. GLY is an enzyme which et al.e 2008, McKi et al. 2019, Sharma et al. assists in the degradation of complex sugars (carbon) 2018). PHO and GLY are not commonly in biomass. investigated enzymes but have shown promise in a few recent research studies to assess When extracted, these enzymes react with a phosphate limitation (Keithley and Kirisits 2019, synthetic substrate which releases a fluorophore McKie et al. 2019). detected using a fluorometer or flow cytometry (Berney et al. 2008, Evans et al. 2013a, Keithley and Kirisits 2019, Sharma et al. 2018, Zhao et al. 2019). Adenosine AMP is the lowest energy state of the adenosine‐ Hooper et al. (2019) conducted a preliminary Monophosphate phosphate molecule, with just one phosphate group. investigation of this method but additional (AMP) ATP is the highest energy state with three phosphate testing is warranted to vet application to groups. AMP can be correlated to the mass of biofiltration. dormant cells in a sample. The ratio of AMP to ATP (AMP/ATP) can serve as an indicator of bioactivity. AMP is measured on the filter media, from near the top of the filter bed. Analysis of ADP in addition to AMP and ATP could be useful for characterizing active biomass. The ratio of (AMP + ADP)/ATP could be a useful indicator of bioactivity, but further research is warranted. Phospholipid PLFAs are components of cell membranes extracted Analysis is specialized and expensive. Use has fatty acids (PLFA) with a chloroform‐methanol‐buffer and analyzed by been limited to research studies (Emelko et al. GC‐FID/MS. 2006, Evans et al. 2013a, Findlay et al. 1989, Fonseca et al. 2001, Huck et al. 2000, Seredyńska‐Sobecka et al. 2006, Xiang et al. 2013). Heterotrophic Requires suspension of bacteria from filter media Method is labor‐intensive; biomass may be Plate Count (HPC) samples via sonication and homogenization. R2A agar underestimated due to incomplete suspension is used with the pour or spread plate methods for of biomass (Camper et al. 1985, Evans et al. manual enumeration. 2013a). Flow Cytometry Flow cytometry detects numbers of aqueous Aqueous‐phase bacteria have been enumerated suspended particles. in some studies (Hammes et al. 2008), but are not representative of filter media biomass (Evans et al. 2013a). Notes: DNA – deoxyribonucleic acid; GC‐FID/MS ‐ gas chromatography‐flame ionization detector and mass spectroscopy; INT ‐ 2‐para (iodo‐phenyl)‐3(nitrophenyl)‐5(phenyl) tetrazolium chloride; PCR – polymerase chain reaction; – quantitative PCR R2A – Reasoner’s 2A agar.

2.5 Recommended Monitoring Tools Table 2‐8 provides an overview of the methods, monitoring locations, sample types, frequency and typical ranges in values. Table 2‐9 provides quality assurance and quality control parameters including method number, holding time, and detection limit as well as qualitative costs.

36 The Water Research Foundation Table 2‐8. Summary of Monitoring Tools. Monitoring Online/ Frequency of Analyte Location Grab Monitoring Typical Values or Range of Values Biological Filter influent Coupon collected every Influent: 1.3 to 13 pg ATP/mm2/d Biofilm Formation Rate Grab and effluent two weeks Effluent: 0.9 to 12 pg ATP/mm2/d Adenosine For start‐up or trouble‐ Filter media Grab1 102 to 108 pg ATP/g filter media1 Triphosphate (ATP) shooting, collect weekly Measurable between 0.02 to 2.64 mg total glucose (free + bound)/g media2,3. Proteins Extracellular polymeric For troubleshooting Filter media Grab1 have been reported to range between 0.2 to substances (EPS) collect every two weeks 4.0 mg BSA/g media2,5 though higher concentrations are possible. Luminescent Dissolved Filter influent Online hourly, for facilities Online Net reduction across the filter of 0 to 4 mg/L1 Oxygen Probe (LDO) and effluent without pre‐ozone Microbial Community For research Filter media Offline Varies Analysis investigations only Natural Organic Matter (NOM)

Dissolved organic Filter influent Online Weekly grab samples 0.1 to 5 mg/L influent, net reduction across carbon (DOC) and effluent or Grab Online hourly filter of 0 to 2 mg/L1

Filter influent For optimization or 2 to 200 µg‐C/L influent with removal up to Carboxylic Acids Grab and effluent research investigations 171 µg‐C/L1,4,5 10 to 1500 µg‐acetate‐C/L influent with Assimilable Organic Filter influent For research Grab removals across filter ranging 0 to 220 µg‐ Carbon (AOC) and effluent investigations acetate‐C/L1 Biodegradable Filter influent For optimization or 0.2 to 1.14 mg/L influent with removals across Dissolved Organic Grab and effluent research investigations filter ranging 0 to 0.48 mg/L1 Carbon (BDOC) 0.01 to 0.08 abs/cm influent UV and 0.1 to 4 Filter influent 254 UV and SUVA Grab Weekly and L/mg‐m influent SUVA. Effluent level varies 254 and effluent with filter design1 Filter influent For research 0.01 to 0.08 abs/cm influent. Effluent level UV/VIS Online and effluent investigations only varies with filter design1 Other Water Quality For troubleshooting, optimization or research investigations collect Nitrate: 0.06 to 0.67 mg/L influent Nutrients (Ortho‐P, Filter influent monthly if filter hydraulics Nitrite: 0.01 mg/L influent6 Ammonia, Nitrite, and effluent Grab are an issue. Recent Total P: 0 to 5.8 mg/L influent1 Nitrate) water research has shown that Ortho‐P: 0 to 0.16 mg/L influent6 PHO:GLY ratio may be a Ammonia: 0 to 1.5 mg/L7 better indicator of phosphorus‐limitation. Temperature Filter influent Online Hourly 5°C to 30 °C 0.01 to 1.5 NTU influent1 Per SWTR requirements, 95% of samples must Turbidity Filter effluent Online Online every 15 minutes be less than 0.3 NTU and no combined effluent samples are to exceed 1 NTU pH Filter influent Online Online, hourly 6.0 to 9.0 SU Weekly if a pre‐oxidant Oxidant residual Filter influent Grab 0 to 0.5 mg/L1 is used Notes: 1. Evans et al. 2013b; 2. Keithley and Kirisits, 2018; 3. Lauderdale et al. 2011; 4. Evans et al. 2016; 5. Hooper et al. 2019; 6. Nyfennegger et al. 2016; 7. Lauderdale et al. 2018.

Biofiltration Guidance Manual for Drinking Water Facilities 37 Table 2‐9. Summary of Monitoring Tool Quality Assurance and Quality Control (QA/QC) and Relative Cost. Adapted from Hooper et al. 2019. ٭Qualitative Cost Method/ Holding Detection In‐house Analysis Offsite Analyte Method Number Time Limit Capital Reagents/Supplies Analysis Biological 0.01 pg ATP/ Biofilm Formation Rate ATP analysis of coupons 24 h1 $$$ $ NA mm2d 1 Adenosine Triphosphate 1 pg ATP/g Luciferin/Luciferase2,3 48 h1 $$$ $ NA (ATP) dry wt1 Proteins: modified Lowry 0.01 mg 24 h1 $ $ Extracellular polymeric Assay with BCA4 BSA/g TS $ substances (EPS) Carbohydrates: Phenol‐ 0.04 mg/g 24 h1 $$$ $$ sulfuric acid assay5 filter media6 Luminescent Dissolved ASTM D888‐097, proposed Analyze 0.2 mg/L $$ $ NA Oxygen Probe (LDO) EPA Method 360.3 Immediately Microbial Community Illumina Sequencing 28 d NA NA NA $$ ‐ $$$ Analysis Natural Organic Matter (NOM) Dissolved organic carbon SM 5310 B or C 28 d 0.05 mg/L $$$ $$ $ (DOC) Carboxylic Acids EPA 300.1 modified8,9 24 h 3‐5 µg/L $$$ $$$ $$ 10 µg SM 9217 $$$ $ $$ Assimilable Organic acetate‐C/L 24 h Carbon (AOC) Bioluminescent P17 and 10 µg $$$ $ $$ NOX method10,11 acetate‐C/L Biodegradable Dissolved Sand method12 48 h 0.2 mg/L $$$ $ $$ Organic Carbon (BDOC) UV‐VIS spectra, UV , SM Analyze 0.001 UV , grab 254 $$$ $ $ 254 5910B, SM 5910 Immediately abs/cm Variable Analyze UV/VIS Spectroscopy Online UV‐VIS spectra depending $$$ $ $ Immediately on analyte Water Quality For ammonia: EPA Method 14 d o‐ 300 or SM 4500‐NH3D; For Nutrients (Ammonia, phosphate 0.002 to 20 nitrite and nitrate: EPA Orthophosphate, and mg/L, $$ ‐ Method 350 or SM 4500‐ $ $ Nitrate, Nitrite, Total ammonia depending $$$ NO3 E and SM 4500‐NO2 B; Phosphate) 24 h nitrite on method For phosphates: EPA and nitrate Method 365.1 or 300 Analyze Temperature SM 2550 0.1 °C $ $ NA Immediately Analyze Turbidity SM 2130 0.001 NTU $ ‐ $$ $ NA Immediately Analyze pH SM‐4500‐H+ 0.1 SU $ ‐ $$ $ NA Immediately Analyze Dissolved ozone residual SM 4500‐O3 0.01 mg/L $$ $ $ Immediately Analyze Chlorine residual SM 4500‐Cl G 0.1 mg/L $$ $ $ Immediately Low, $$ = Medium, $$$ = High = $ ٭ Sources: 1. Hooper et al. 2019 2. Velten et al. 2007; 3. Magic‐Knezev and van der Kooij 2004; 4. Lowry et al. 1951; 5. DuBois et al. 1956; 6. Evans et al. 2013b; 7. Urfer and Huck 2001; 8. Peldszus et al. 1996a; 9. Peldszus et al. 1996b; 10. Weinrich et al. 2009; 11. Haddix et al. 2004; 12. Volk et al. 1994.

38 The Water Research Foundation 2.6 Developing a Monitoring Strategy Biofilter monitoring must incorporate an evaluation of biological, NOM, water quality, filter hydraulics, and operational parameters, in addition to any specific contaminants of interest for which the process is used. Monitoring data must also be evaluated against site‐specific filter treatment objectives. The treatment objectives should specify the contaminant(s) of interest, target contaminant(s) filter effluent concentration and/or the degree of removal, and the hydraulic performance requirements. These treatment objectives should be used to guide performance monitoring and decide when optimization testing or troubleshooting methods are needed. 2.6.1 Routine Operation During routine filter operations, a succinct monitoring strategy is needed to characterize the process and develop a baseline for comparison if troubleshooting or optimization is required in the future. Baseline testing should be completed for at least one year to evaluate seasonal variability on water quality and operations. Monitoring should include multiple categories of tools (e.g., biological, NOM, water quality, filter hydraulics, and other operational). More parameters can be included and tailored depending on site‐specific needs, but the tools in Table 2‐10 are the minimum required. Online monitoring data for flow rate, run time, UFRV/UFBV are recommended for monitoring filter hydraulics. DOC is recommended for water quality monitoring at the filter influent and effluent, but if sufficient data exist to demonstrate a direct correlation between DOC and TOC at the monitoring point, then TOC may be used. Other constituents with biofilter treatment goals should also be included at the filter influent and/or effluent. Additional water quality data for temperature, turbidity, and pH must be monitored on a regular basis, where online continuous measurements are preferred. UV254 and nutrients may also be included but are not required. Nutrients should be monitored monthly or less frequent. The biofilm formation rate and DO consumption should be used to evaluate biological activity. If a pre‐filter oxidant is used and an oxidant residual is expected at the filter influent, then the biofilm formation rate pipe loop should either be installed from a sample tap that is one ft below the top of the media or a quenching agent, such as calcium thiosulfate, should be metered into the pipe loop to quench oxidant residual before the coupons. Alternatively, the pipe loop couldd be installe prior to oxidant addition. Oxidant residual at the biofilter influent should be monitored weekly to confirm concentrations are near or below detection, particularly for anthracite and sand filters. If the facility uses ozone prior to filtration, then DO consumption should not be monitored. Table 2‐10. Recommended Biofilter Monitoring During Routine Operations. Analyte Location Level Operations Flow Individual and combined filters 1 EBCT Per Filter 1 Runtime Per Filter 1 UFRV Per Filter 1 Headloss Per Filter 1 Underdrain headloss during backwash Per Filter 1 Water Quality DOC Biofilter influent and effluent 1

Other constituents w/treatment goals (i.e., Mn, NH3) Biofilter influent and/or effluent 1 pH Biofilter influent or effluent 1 Temperature Biofilter influent or effluent 1 (Continued)

Biofiltration Guidance Manual for Drinking Water Facilities 39 Table 2‐10. (Continued) Analyte Location Level Turbidity Biofilter influent or effluent 1

UV254 (influent and effluent) Biofilter influent and effluent 2 Nutrients (i.e., P and N) Biofilter influent and effluent 2 Biological Biofilm formation rate Biofilter influent and effluent 2 DO Consumption* Biofilter influent and effluent 2 Upstream Operations Oxidant dose (i.e., raw or intermediate ozone, NA 1 chlorine, backwash oxidant, etc.) Oxidant residual (filter influent, backwash waste) filter influent, backwash waste 1** Notes: 1 = highly recommended 2 = recommended *DO consumption should only be measured at non‐ozone facilities **Measured if oxidant is applied prefilter or during backwash

2.6.2 Startup and Troubleshooting Table 2‐11 shows additional monitoring recommendations for these focused testing periods but should be augmented if specific troubleshooting issues are being addressed. During startup, the biofilm formation rate or ATP on filter media can be sampled to confirm biomass has established on the filter media and samples should be collected until steady‐state conditions exist. This may take a few weeks to a few months depending on temperature and water quality. The reduction in biofilm formation rate from filter influent to effluent is already listed as a routine monitoring parameter and is preferred over ATP. As bioactivity acclimates, the difference in the biofilm formation rate from the influent to effluent will increase and then stabilize once steady‐state conditions have been attained. Acclimation may also be needed if a filter is placed offline for an extended duration, if a filter is drained/dried, or if significant changes in water quality occur. During troubleshooting, if filter hydraulics have been compromised, then EPS should be included and monitored weekly. If filter effluent water quality has been compromised, then ATP should be included to evaluate whether loss in biomass has occurred. ATP may prove better than the biofilm formation rate for troubleshooting because analysis can be conducted quickly, whereas the biofilm formation rate requires two weeks of incubation. Additional tools may prove helpful to determine the cause of upset conditions, such as carboxylic acids, as well as operational tools like oxidant demand and the ratio of upstream oxidant dose to TOC. Table 2‐11. Additional Recommended Biofilter Monitoring During Startup and Troubleshooting. Additional Tools for Startup and Troubleshooting Level Biological ATP 2 EPS 2 Notes: 1 = highly recommended 2 = recommended 3 = recommended for optimization

40 The Water Research Foundation 2.6.3 Special Studies, Optimization, and Research Table 2‐12 shows additional recommended monitoring for special studies, optimization and research. Additional tools may be augmented, depending on the goals of the investigation. For example, if pre‐ filter oxidation is being optimized to enhance organic carbon removal, then the pre‐oxidant dose, pre‐ oxidant dose to TOC ratio, pre ‐oxidant residual, carboxylic acids, and UV254 should be included. If filter optimization testing is being conducted to optimize filter hydraulic performance, then EPS should be monitored on a weekly basis, nutrients should be monitored monthly, and differential pressure profiling should be considered. If removal of a specific contaminant of interest is a focus for optimization, then monitoring for that contaminant should be included at the filter influent and effluent, and enzyme activity may prove helpful. Microbial community characterization should only be incorporated in research studies and at least 50 gram of filter media should be collected in duplicate.

Table 2‐12. Additional Recommended Biofilter Monitoring During Special Studies, Optimization, and Research. Additional Tools for Special Studies/Research/Optimization Level Water Quality Carboxylic Acids 3 Assimilable Organic Carbon (AOC) 3 Biodegradable Dissolved Organic Carbon (BDOC) 3

Biological ATP 2 EPS 2 Microbial Community Analysis 3 Enzyme Activity 3 Notes: 1 = highly recommended 2 = recommended 3 = recommended for optimization

2.7 Data Management Utilities should have a clear and organized data management plan. Several of the tools described utilize online systems, which can be effectively monitored and recorded using SCADA for real‐time and long‐ term performance metrics. Other tools that require either grab sampling and in‐house testing or testing through an outside lab can be compiled using software such as Microsoft Access (recommended for large datasets), Microsoft Excel with templates to generate graphs, tables, and other figures that will assist in data interpretation. An established schedule and protocol for inputting and analyzing data from outside labs, in‐house analyses, and SCADA should be established according to the needs of the utility. Data management is discussed in further detail in the Tools Compendium in Appendix G.

Biofiltration Guidance Manual for Drinking Water Facilities 41

42 The Water Research Foundation

Optimizing Existing Biofiltration Plants

Minimizing or avoiding chlorine application to a filter will result in biofiltration. However, considerable research has been performed over the last decade showing that additional process and design strategies can be implemented that may improve biofilter hydraulic and treatment performance. The implementation of these optimization strategies has become known as “engineered biofiltration.” Appendix H summarizes biofiltration optimization studies and the various optimization strategies that were investigated. Detailed biofiltration design and operation are covered in Chapters 5 and 6, respectively. 3.1 Planning Despite the proven treatment results of biofiltration, it is important to assess biofilter systems for potential impacts to overall plant and filter operation to avoid unintended consequences. Unintended consequences of biofilters can include instrumentation fouling, unwanted biological growth in other processes, high chlorine demands, shorter FRTs, unwanted algae growth, and issues with maintaining chlorine and ammonia residuals. Planning for optimization of existing biofilters can benefit utilities by improving overall treatment efficiency and water quality as well as assist in avoiding unintended consequences. Optimization decision trees provided in Appendix I can be used to determine the suitability of a given optimization strategy eand th steps needed to evaluate and implement a given strategy. 3.1.1 Suitability Utilities should consider finished water quality goals prior to making changes for biofilter optimization. These strategies are typically considered when a biofilter is consistently underperforming with respect to hydraulic and/or water quality goals. Table 3‐1 lists the potential benefits of a given optimization strategy and therefore can be used to determine the suitability of a strategy. Indicator parameters for assessing the impacts of an optimization strategy are also listed.

43 The Water Research Foundation

Table 3‐1. Optimization Strategies. Optimization Strategy Potential Outcomes Indicator Parameters Media optimization  Improved hydraulics  Headloss accumulation  Improved filtration/particulate rate/UFRV removal  DOC/UV254  Improved biological activity,  DBPFP particularly under stressed conditions  Fe/Mn/NH3 (low temp, low EBCT, presence of  Disinfectant stability oxidant)  Turbidity Modified backwash  Improved hydraulics  UFRV protocol  Optimized biomass  Clean‐bed headloss  Mitigate fouling (i.e., underdrain and  Headloss accumulation rate media fouling)  ATP  Filter runtime  Backwash headloss Nutrient augmentation  Improved hydraulics  ATP  Improved biological growth and  EPS activity  UFRV  Improved water quality  Clean‐bed headloss  Improved removal of dissolved  Headloss accumulation rate constituents  Backwash headloss  DOC  Nitrogen species  Fe/Mn/NH3 pH adjustment  Improved biological activity  ATP  Improved water quality  Fe/Mn/NH3  Improved inorganic compound  TOC removal Pre‐oxidant addition  Improved organic compound removal  Headloss accumulation rate  Decreased biological activity  DOC/UV254  Improved hydraulics  UFRV  ATP  AOC/BDOC (site‐specific) Holistic optimization  Decreased upstream coagulant  DOC/UV254 demand  Coagulant dose  Decreased residuals solids production  Residuals generation rate  Decreased disinfectant demand  Disinfectant residual

3.1.2 Testing Evaluation of biofiltration optimization through bench‐, pilot‐, and/or full‐scale testing is necessary to assess whether long‐term performance goals can be realized. Testing can also provide a better understanding of site‐specific process requirements, limitations, and potential challenges associated with any optimization process, which allows utilities to identify, develop, and test potential mitigation strategies. Finally, testing provides an opportunity to develop and refine a monitoring plan for future full‐scale implementation. Biofiltration testing design and implementation details are discussed in Chapter 7 of this Biofiltration Guidance Manual. However, some key testing considerations specific to biofiltration optimization are discussed below:

44 The Water Research Foundation

(1) Evaluate the scales (i.e., bench‐, pilot‐, and/or full‐scale) that should be considered for biofilter optimization testing and what the outcomes and limitation are at each scale. Understanding plant objectives can help to determine which scale would be most important. (2) Benchmark performance for the existing plant (water quality and hydraulics) over various seasons and operating conditions. Before biofiltration optimization begins, it is important to understand and evaluate existing conditions. Benchmarking helps to establish performance goals and objectives and is critical for quantifying the impact of a given optimization strategy. (3) Select testing durations that are appropriate for a given optimization strategy. The duration of testing for each optimization strategy can differ significantly between optimization strategies. As observed in WRF 4555 (Lauderdale, Alito, Hooper, Dowdell, and Wert 2018), oxidant optimization can be observed within hours or days of chemical addition, while nutrient and pH adjustment strategies can take weeks or months to produce a measurable difference in operation. (4) When testing media replacement, existing media should be characterized through media analysis to determine ES and distribution. Media analysis results should be compared to fresh media characteristics provided during commissioning. (5) When testing nutrient addition, first evaluate the current carbon:nitrogen:phosphorus ratio in the biofilter feed to determine nutrient dosing targets. Recent research has shown that PHO:GLY ratio may be a better indicator for phosphorus limitation (Keithly and Kirisits 2019). (6) When testing the use of pre‐oxidants, dosing should start low and be increased carefully and gradually so as to avoid broad inactivation of the biofilter community. (7) When testing the impact of increasing biofilter feed pH, it is important to monitor coagulant metal concentration downstream of the biofilter, as coagulant metals tend to resolubilize at higher pHs. (8) If testing multiple optimization strategies sequentially, testing should be staggered to the extent possible to minimize the influence of the previous optimization strategy. (9) Once multiple optimizations are evaluated individually, testing should also consider applying multiple optimization strategies simultaneously (e.g., pH adjustment + nutrient addition), which can provide synergistic benefits. 3.2 Optimization Strategies Biofiltration optimization leverages design and operational strategies to enhance contaminant biodegradation kinetics, particle removal, and/or filter hydraulics. Biofilter optimization strategies that have been explored in drinking water treatment and will be discussed in this chapter include:  Media selection.  Backwash protocol.  Nutrient augmentation.  pH adjustment.  Pre‐oxidant addition.  Holistic optimization. Optimization strategies are tested empirically to ensure that all biofilter performance objectives will be met. 3.2.1 Media Selection Most biofilters are dual media with anthracite or GAC over sand (Brown et al. 2016). The sand layer improves filtration characteristics to enhance particle removal. Drivers for selecting anthracite or GAC

45 The Water Research Foundation

and size include cost, durability, adsorption needs, and resultant biological activity. Media size can also impact biofilter hydraulic and treatment performance. Table 3‐2 summarizes the characteristics of media type and size options.

Table 3‐2. Media Selection Characteristics. Media Pros Selection Cons Type  GAC  GAC: - Provides adsorption capacity, though - More expensive this is temporary without regular media - More susceptible to abrasion, which changeout may lead to media loss - More conducive to maintaining higher  Anthracite: biomass levels; this can be - Supports lower biomass levels advantageous where biofilters compared toC; GA this may limit experience one or more environmental biological activity in biofilters that stressors (e.g., cold temperatures, experience one or more environmental regular presence of chlorine) stressors - Holds contaminants for longer on the media through adsorption, which may allow contaminant biodegradation over a longer time span, subsequently freeing adsorption sites  Anthracite - Less expensive - Less susceptible to abrasion Size  0.5 to 1.3 mm effective size:  0.5 to 1.3 mm effective size: - Greater specific surface area results in - Smaller size results in greater headloss higher biomass levels; this may  >1.3 mm effective size: improve contaminant removal and - Increased pore space may also robustness under stressed conditions decrease particle removal (lower temperature and/or presence of - Lower specific surface area results in chlorine) lower biomass levels; this may diminish  >1.3 mm effective size: contaminant removal and robustness - Improved filter hydraulics due under stressed conditions increased pore space: this would likely result in lower pumping costs and greater UFRVs Source: Lauderdale et al. 2018.

46 The Water Research Foundation

3.2.2 Backwash Protocol Particle accumulation and biomass growth over the course of a filter run consume void space and generate headloss. Thus, biofilters must be taken offline regularly and backwashed to remove excess particles and biomass. If the backwash protocol is insufficient, filter clogging and channeling can eventually occur, allowing particle and contaminant breakthrough.e If th backwash protocol is overly aggressive, filter efficiency decreases unnecessarily, and the filter may require extensive ripening times to reestablish sufficient particle removal and biological activity. An optimized backwash protocol removes sufficient particles and biomass to reestablish target hydraulics without removing too much biological activity. Parameters that should be tracked to monitor the effectiveness of a given backwash protocol are summarized in Table 3‐3.

Table 3‐3. Parameters for Tracking Backwash Effectiveness. Parameter Comments Clean‐bed Typical clean‐bed headloss will vary based on several factors including media size, loading headloss rate, water temperature, etc. It is important that clean‐bed headloss for a given plant and filter remains fairly steady. Headloss Headloss accumulation rate varies based on several factors and it should remain fairly accumulation consistent from run to run. rate UFRV Different plants have different targets for minimum UFRV (e.g., >10,000 gal/ft2). An optimized backwash protocol will result in sufficient and fairly consistent UFRVs among runs. Backwash Backwash pressure should remain fairly constant from backwash to backwash. Steadily pressure increasing backwash pressure under similar backwash loading conditions could be indicative of filter and/or underdrain clogging. Contaminant If a backwash is overly aggressive, a target contaminant (DOC, MIB, ammonia, etc.) may removal break through the biofilter at unusually high levels while biological activity is being reestablished. Alternatively, if a backwash is insufficient, channeling may occur, which also may lead to elevated contaminant breakthrough.

Ultimately, demonstration testing (see Chapter 7) is required for backwash optimization. The discussion below summarizes various tools that can be considered when optimizing a biofilter backwash protocol. 3.2.2.1 Fluidization Rate and Duration Backwash fluidization rates should be established to achieve 30 to 50% bed expansion, which results in good mixing of the media. Backwashing duration generally ranges from four to eight minutes (Crittenden et al. 2012). 3.2.2.2 Air Scour Addition of air, as an initial step in the backwash protocol or in a subsequent combined air‐water wash, can be particularly effective for removing particles and biomass (Emelko et al. 2006,d Soucie an Sheen 2007). Air scour enhances media cleaning by causing substantial agitation within the media bed. The use of air scour during backwashing is generally used for the purposes of media cleaning only, and does not provide water quality benefits (Emelko et al. 2006, Liu et al. 2001). 3.2.2.3 Oxidant Addition While not common practice, research has shown that adding minimal amounts of oxidants can provide improved media cleaning and removal of excess EPS and associated biomass without diminishing biomass activity (Lauderdale et al. 2018). However, success using an oxidant in backwashing is site‐ specific, and depends on filter size, media type, and oxidant demand. Oxidants that have been studied include chlorine, chloramines, and hydrogen peroxide. Chlorine doses up to 2 mg/L, chloramine doses

47 The Water Research Foundation

up to 5 mg/L, and hydrogen peroxide doses up to 10 mg/L have been evaluated (Lauderdale et al. 2018). Regardless of the oxidant used, the residual should be monitored to determine if an appropriate dose is applied. Chlorine and Chloramine. Research investigating the use of chlorine in filter backwash has shown benefits and challenges with this optimization strategy. Chlorinated backwash has been shown to reduce accumulated biomass at the top of granular media filters, and decrease effluent turbidity and lower headloss during normal operation (Liu et al. 2001). Alternatively, chlorinated backwash has also inhibited biological activity in the biofilter and reduced organics removal (Ahmad et al. 1998, Krasner et al. 1993). Additionally, chloramine in backwash water has not been shown to substantially impact organics removal during normal operation (Liu et al. 2001). It is generally not recommended to use chlorine in backwash water when temperatures are low (≤ 5°C), although if chlorine must be used during low temperature operation, GAC media should be utilized instead of anthracite/sand (Liu et al. 2001). Hydrogen Peroxide. Hydrogen peroxide use in backwash water has shown a reduction in the rate and extent of underdrain fouling as well as improved hydraulic performance during normal operation without diminishing biological activity. However, research indicates that peroxide use during backwash does not improve water quality (Lauderdale et al. 2014). 3.2.3 Nutrient Augmentation Nutrients, including biodegradable organic carbon, nitrogen, and phosphorus, are essential for promoting and maintaining biological activity within a biofilter. Seasonal changes in surface water quality can increase or decrease nutrient loading or availability to biofilters, subsequently affecting biological health. Decreased nutrient (i.e., nitrogen and/or phosphorus) concentrations limit biological activity and maye cause th production of EPS, which, in conjunction with particle loading, can lead to filter clogging. Therefore, hydraulic parameters (e.g., headloss trends, UFRV, backwash pressure), microbial parameters (e.g., EPS, ATP), and target contaminants are prospective indicators of nutrient limitations. While nutrient augmentation upstream of the biofilters can often improve biological activity and hydraulic performance, benefits are site‐specific, and it is possible that no benefits will be realized. Recent research (Keithly and Kirisits 2019) indicates that the PHO:GLY ratio could be a better indicator of phosphorus limitation. There are also challenges with its implementation as the nutrient requirements vary from site to site. Optimal nutrient conditions should be tested at the pilot‐ or full‐scale prior to making system‐wide changes. It is important to avoid overdosing nutrients, which can lead to excessive biomass growth and nutrient presence in finished water. While too much phosphorus may not be an issue, especially if the utility is feeding phosphorus for corrosion control, excessive ammonia can lead to disinfection and breakpoint chloramination issues. 3.2.3.1 Dose As discussed in Section 2.3.4, in general, a 100:10:1 molar ratio of biodegradable organic carbon consumed (C): ammonia‐nitrogen (N): orthophosphate‐phosphorus (P) is suggested to adequately support biological growth. This correlates to a concentration ratio of 1 mg C/L, 0.117 mg N/L, and 0.026 mg P/L. The amount of nutrient addition necessary will depend on a confirmation of a nutrient limitation/deficit, which should be investigated prior to chemical addition. To translate this ratio into a target dose for testing and/or full‐scale implementation, the following protocol is suggested (Lauderdale et al. 2018): • The bioavailable carbon concentration could be based on DOC removal from the biofilter influent to effluent. BDOC concentrations in the biofilter influent could provide a rough estimate, if available.

48 The Water Research Foundation

• Using the 1 mg C/L: 0.117 mg N/L ratio, calculate the ammonia nitrogen requirement. Usually WTPs exceed the necessary nitrogen concentrations in their influent, therefore nitrogen is typically not a limiting nutrient. • Using the 1 mg C/L: 0.026 mg P/L ratio, calculate the orthophosphate‐phosphorus requirement nitrogen requirement. Upstream coagulants tend to strip phosphate from biofilter influent; therefore, phosphate is usually the limiting factor. While a molar ratio of 100:10:1 is a good starting point for utilities that wish to optimize biofilters through nutrient augmentation, the ratio may need to be adjusted for a given utility as a result of the specific microbial communities present in their biofilters and characteristics of their treatment processes (Nyfennegger et al. 2016). Furthermore, PHO:GLY ratio may be a better indicator of phosphorus limitation (Keithly and Kirisits 2019). Table 3‐4 summarizes a range of nutrient ratios tested during previous biofiltration pilot studies. In general, phosphorus tends to be the limiting nutrient, and required supplemental orthophosphate doses are low. Furthermore, many utilities already use orthophosphate as a corrosion inhibitor, so it might be possible to leverage existing chemical storage facilities in some cases.

Table 3‐4. Summary of C:N:P Ratios Tested. Biofiltration Pilot Study Nutrient Ratios Tested WRF 252 (Huck et al. 2000) 15:5:1

WRF 4555 (Lauderdale et al. 2018) 100:10.2:0.23 to 100:14.3:0.11 100:0.4 (C:P) to 100:1.4 (C:P) WRF 4346 (Lauderdale et al. 2014) 100:37:1 100:33:2 100:34:3 100:10:2 100:32 (C:N) 100:19 (C:N) 100:20:>2 WRF 4215 (Lauderdale et al. 2011) 100:10:1 100:6:1 100:4:4 100:2:8 100:2:3 100:3:2 100:14:2 Lauderdale et al. (2012) 100:10:1

49 The Water Research Foundation

3.2.3.2 pH Impact If coagulated floc (i.e., aluminum or iron containing floc) carries over into biofilter feed water, it may adsorb dosed phosphorus, thus limiting or eliminating any potential benefit of nutrient augmentation. If biofilter feedwater pH is lower than the isoelectric point (pH resulting in zero surface charge) of the coagulated floc, then the surface of the floc will carry a net positive charge, which would facilitate the adsorption of negatively charged orthophosphate. Thus, for plants that experience consistent floc carryover to their filters, increasing the pH of the biofilter feedwater may improve the photoavailability and thus the benefit of dosed orthophosphate (Lauderdale et al. 2014). This can be overcome by pH adjustment upstream of the biofilter but downstream of the flocculation/sedimentation process. However, it is important that utilities consider finished water quality goals such as corrosion control or DBPs prior to making full‐scale changes. 3.2.4 Pre‐Oxidation Pre‐oxidation involves the addition of ozone, ozone/peroxide, chlorine, chloramines, peroxide, or permanganate at low doses upstream of the biofilters to remove excess biomass for maintaining efficient hydraulic operation. Pre‐oxidants can break down larger, more complex organic compounds into smaller, more biodegradable compounds, which could improve DOC removal across downstream biofilters. Additionally, pre‐oxidants can improve hydraulic performance by reducing biological fouling at the filter surface (de Vera et al. 2019). Oxidants improve hydraulics by pushing biology deeper into the filter bed, leaving the top filter layer for particle removal, while nutrients improve hydraulics by promoting biological growth and reducing EPS formation. Pre‐oxidants can also improve the filterability of particles, thereby decreasing biofilter effluent turbidity. Since pre‐oxidants are also disinfectants, use of pre‐oxidants must be considered carefully with the intention of minimizing pre‐oxidant breakthrough in biofilter effluent. Thus, along with typical hydraulic parameters (headloss, backwash pressure, UFRV, etc.) and pertinent water quality parameters, it can be particularly useful to also track ATP to detect and avoid over oxidation of a biofilter’s microbial community. Filter configuration has a large impact on oxidant reactivity, where oxidant reactivity with anthracite and sand is relatively low, allowing oxidants to breakthrough and persist in effluent. Alternatively, oxidant reactivity with GAC is higher, and subsequently depleted much faster (de Vera et al. 2019). Table 3‐5 presents oxidant doses that have been evaluated in biofiltration studies to improve operation. There is a wide range of potential oxidant doses that can be used in biofiltration, and the optimal oxidant dose and combination of oxidants is site‐specific. In Table 3‐5, oxidants including ozone, peroxide, chlorine, and chloramine are shown. Oxidant doses are expressed in mg/L. Peroxide:ozone ratios are also shown, when studied, since the peroxide dose selected for evaluation is generally a function of ozone dose. Additionally, Table 3‐5 shows the ratio of ozone to DOC concentration, when possible, as dose is primarily dependent on influent organics concentrations.

50 The Water Research Foundation

Table 3‐5. Summary of Pre‐Oxidant Doses Tested. Author / WRF Project Number Chemicals Evaluated Ozone:DOC Ratio Evaluated Singh Sidhu et al.  Ozone doses of 2.9 and 3.4 mg/L  0.84:1 and 1:1 (2018)  Peroxide and ozone doses of 1.45 and 2.9 mg/L, respectively (H2O2:O3 ratio of 0.5:1)  Peroxide and ozone doses of 2.9 mg/L each (H2O2:O3 ratio of 1:1)

Beniwal et al.  H2O2:O3 ratios of 0.1, 0.2, 0.35, and 0.5, using  Ranged from 0.5:1 to 1:1 (2018) an ozone dose of 2 mg/L

McKie et al.  Ozone 1 mg/L and peroxide 0.2 mg/L (H2O2:O3  1:2 (2015) ratio of 0.2:1)  Peroxide 0.2 mg/L McKie et al.  Ozone doses of 1 mg/L  1:2 (2016) Lee et al. (2012)  Ozone doses of 2, 4, and 8 mg/L  Ranged from 0.5:1 to 1:1 Xing et al. (2018)  Ozone, dose not provided, chlorine residual  2.18:1 of 2.8 mg/L Sun et al. (2018)  Ozone doses ranging from 0 to 4.5 mg/L  Ranged from 0:1 to 1.1:1 Ross et al. (2019)  Not provided  0.4:1 de Vera et al.  Chlorine 0.5 to 2 mg/L  None (2019)  Chloramine 0.5 mg/L  Hydrogen peroxide 2 to 5 mg/L Black and Berube  Ozone, doses not provided, peroxide 10 mg/L  1:1 and 2:1 (2013) Lauderdale et al. Peroxide 1 mg/L None (2012) WRF 504 (Price et Ozone doses of 1.5 and 3.0 mg/L 0.75:1 and 1.5:1 al. 1995) WRF 252 (Huck et  Ozone 1.3 to 2.0 mg/L Ranged from 0.5:1 to 1:1 al. 2000)  Peroxide Chlorine 0.5 to 3.25 mg/L WRF 2775 Ozone doses of 2.8 to 11 mg/L Not provided (Westerhoff et al. 2005) WRF 4215 Peroxide 1.0 mg/L None (Lauderdale et al. 2011) WRF 4525 Peroxide 0.2 to 1.0 mg/L None (Nyfennegger et al. 2016) WRF 4346 Peroxide 0.1 to 2.0 mg/L None (Lauderdale et al. 2014)

51 The Water Research Foundation

3.2.4.1 Ozone, Peroxide, and Ozone/Peroxide Ozone is a successful pre‐oxidation strategy that is often coupled with biofiltration due to its ability to convert non‐biodegradable organics. Research has shown that as ozone dose increases, organics conversion to biodegradable fractions increases, providing improved organics removal through the biofilter. However, the extent of removal and the required ozone dose is site‐specific, and dependent on the influent TOC concentration (Lauderdale et al. 2018). One major disadvantage of using ozone upstream of a biofilter is the potential for the formation of bromate, a regulated DBP, which is formed during a reaction between ozone, bromide, and organics. Bench‐ or pilot‐scale testing of ozone is recommended prior to full‐scale implementation to assess bromate formation if bromide is present. Pre‐oxidants including ozone and peroxide, as well as a combination of the two (termed “peroxone”) have shown great success in biofiltration hydraulic improvements. Hydrogen peroxide may lower EPS concentration by up to 65% (Nyfennegger et al. 2013). 3.2.4.2 Chlorine and Chloramines Chlorine concentration subsides as it produces AOC from DOC, promoting microbial growth (Fisher et al. 2017, LeChevallier et al. 1992). However, research has also shown that the use of chlorine upstream of biofilters lowered organics removal compared to biofilters that did not receive chlorine (Lauderdale et al. 2018). Another study found that the addition of chlorine and chloramines neither significantly improve turbidity and organics removal nor lowered particle counts (de Vera et al. 2019). Chlorine is a pre‐oxidant that can be added for improved headloss through the biofilter bed. While traditional biofiltration does not usually involve chlorine addition upstream to avoid killing the biomass, research has shown that minimal amounts of chlorine will push biology deeper into the biofilter bed (de Vera et al. 2019, Lauderdale et al. 2018). This decrease in headloss occurs due to the inhibition of filter‐clogging biomass at the filter surface, which leaves the top filter layer predominantly for particle loading. It is generally recommended to apply a chlorine dose that will result in no or minimal residual in biofilter effluent. Chloramine can be used as a pre‐oxidant and can be implemented upstream of the biofilter. Chloramine addition serves the same purpose upstream of a biofilter as chlorine, and one study showed headloss decreased up to 90% with chloramine addition (Lauderdale et al. 2018). The use of chloramine upstream of a biofilter can provide ammonia removal without losing nitrifying functionality. Research has shown ammonia removal and nitrate accumulations were, on average, 73% and 25%, respectively (Lauderdale et al. 2018). 3.2.4.3 Permanganate Research has shown that permanganate addition upstream of biofiltration did not improve the removal of TOC, DOC, AOC, and BDOC. This did not provide any noticeable decrease of DBP ‐FP potential (Lauderdale et al. 2018). The addition of permanganate upstream of biofiltration has not been shown to provide substantial hydraulic improvements. Therefore, until further research shows otherwise, permanganate should not be considered as a biofilter optimization strategy.

52 The Water Research Foundation

3.2.5 pH Adjustment pH can affect multiple aspects of biofiltration, including biological activity, depth filtration, and nutrient availability. The ideal biofiltration pH or pH range is site‐specific, driven by the treatment and operational goals for a given biofiltration facility. pH optimization should be considered carefully since a sudden or dramatic increase in pH can potentially cause scaling and/or particle breakthrough during filtration (Nyfennegger et al. 2016). An increased pH can also lead to the resolubilization of coagulant metals. Key monitoring parameters used to track the impact of pH adjustment on biofiltration include standard hydraulic parameters (headloss trends, UFRV, backwash pressure), turbidity, dissolved coagulant metal concentrations, and target contaminant concentrations. Table 3‐6 presents a summary of studies that have focused on the impact of pH on biofilter performance.

Table 3‐6. Summary of pH Increase Tested in Previous WRF Studies. Author / WRF Project Number pH Levels Tested WRF 4555 (Lauderdale et al. 2018) Increased 1.0 SU from 5.8 to 6.8 WRF 4346 (Lauderdale et al. 2014) Increased 0.5‐1.0 SU from 7.5 to 8.0‐8.5

3.2.5.1 Microbial Growth and Activity Microorganisms generally prefer pHs between 6.5 and 8.5 for optimal growth and reproduction (Lauderdale et al. 2018), although pHs as low as 6.0 and as high as 9.0 can also promote biological activity (Evans et al. 2013a). pH adjustment can also be customized to specific organisms of interest for targeted biodegradation (see Table 3‐7). It is worth noting that although modification of pH has been greatly studied in wastewater treatment systems and pure culture biological evaluations, targeted biodegradation in biofiltration is still in its infancy.

Table 3‐7. Optimal pH Ranges for Ammonia, Nitrite, and Manganese Oxidizers. Targeted Microorganism/Function Target pH Range Ammonia and Nitrite Oxidation 7.5 to 8.5 S.U. (Arnaout & Gunsch 2012) Manganese Oxidation 7.8 to 8.0 S.U. (Lauderdale et al. 2016)

3.2.5.2 Depth Filtration Some full‐scale biofiltration plants have experienced improved hydraulic performance (i.e., dampened headloss accumulation) by increasing the pH of the biofilter feed water (Lauderdale et al. 2014, Lauderdale et al. 2016). While mechanistic studies still need to be performed, it is likely that the dampened headloss is the result of improved depth filtration across the biofilter. Increasing pH drives up the surface charge on filter media, which can decrease electrostatic particle removal near the top of the filter bed, spreading particles deeper into the bed and decreasing any localized clogging and headloss as a result. 3.2.5.3 Nutrient Availability As discussed in Section 3.2.3.2, plants that experience consistent floc carryover to their filters may need to increase biofilter feedwater pH to improve the availability and thus the benefit of dosed orthophosphate (Lauderdale et al. 2014).

53 The Water Research Foundation

3.2.6 Holistic Optimization Given the multiple benefits and capabilities of biofiltration, there may be opportunities to optimize upstream treatment to achieve enhanced organics removal through the biofiltration process, thereby reducing overall operating costs and decreasing the volume of plant waste streams. When optimizing the biofiltration process, it is important to consider impact to finished water biostability to avoid the potential for downstream biological growth. Lauderdale et al. (2014) demonstrated that a reduction in the ferric sulfate coagulant dose by 50% did not significantly disrupt the overall TOC removal for the plant, as biofiltration picked up additional TOC removal (Figure 3‐1).

70%

60%

50% Floc/Sed TOC Removal

40% Biofiltration TOC Removal 30% Removal TOC 20%

10%

0% 60 mg/L Ferric Sulfate*9H2O 30 mg/L Ferric Sulfate*9H2O Coagulant Dose Coagulant Dose

Figure 3‐1. Demonstration of Holistic Optimization of Coagulant Dose. Source: Lauderdale et al. 2014.

As noted in the Nutrient Augmentation section (3.2.3), coagulants have the potential to sorb phosphorus species out of solution, thus reducing the nutrient balance going to downstream biofilters. Reducing coagulant dose has the added benefit of improving supplementation phosphorus conditions for nutrient augmentation, as long as other treatment goals such as turbidity and TOC removal are not greatly impacted. To optimize a plant holistically, two primary methods can be used to pinpoint optimal coagulation conditions that could be coupled with downstream biofiltration. First, jar testing can rapidly assess the coagulation efficacy for various water quality conditions by using a series of jars to test chemical dosages and sequencing, while observing formation and settlement of flocs. The second method for assessing coagulation capabilities is through the use of a zetameter to measure zeta potential. Zeta potential measures the surface charge of particles to determine their particle velocity. This is especially useful in coagulation experiments, as zeta potential impacts the density and rate of floc formation. Turbidity removal has been demonstrated to be most stable in zeta potential ranges of +3 mV (millivolts) to ‐22 mV (“Water Treatment and the Role of Zeta Potential in Water Treatment Process Control,” 2005). Implementing holistic optimization will not likely require design or equipment modifications unless a new coagulant is selected. Most modifications will be able to adjust with existing pumps and chemical systems. Sudden changes to coagulant dose can cause upsets to coagulation processes, therefore adjustments should be made slowly and carefully.

54 The Water Research Foundation

Converting Conventional Filters to Biofilters

Potential for improved treatment performance have encouraged many utilities to consider converting their conventional filters to biofilters. Many of these utilities employ conventional filtration using high‐ rate chlorinated filters as the primary method for removing certain pathogens and turbidity. Converting conventional filtration to biofiltration has been shown to improve water qualityh throug the removal of contaminants (e.g., T&O‐causing compounds, manganese, and ammonia), and enhancement of disinfectant stability and effluent biostability. Conversion from conventional filtration to biofiltration is the result of a singular strategy: decreasing the chlorine dose to a filter (in service and possibly during backwash events) to the point that microbial communities can colonize and grow on the filter media. Utilities can experience operational/hydraulic challenges and water quality deterioration during biofilter conversion. As such, the probability for process upsets and unacceptable contaminant breakthrough is highest immediately after the system is converted to biological mode, but these can often be mitigated or avoided with proper planning and evaluation. Therefore, it is critical to evaluate and understand the conversion process, potential outcomes, and the various strategies available to improve the conversion process. Most of the technical and practical concepts applicable to biofilter conversion also apply to existing and greenfield biofiltration plants, and therefore, are covered in detail in Chapters 3 and 5. This chapter covers guidance specific to biofilter conversion applications, walks through the biofilter Conversion Assessment Tool, highlights biofilter testing components that are pertinent to conversion applications, and describes various strategies that can be considered to improve the outcome of a biofilter conversion process. A summary of example biofiltration conversion case studies can be found in Appendix J. 4.1 Planning Converting conventional filters to biofilters can improve treatment performance across a wide range of parameters. However, since it does carry some risks, biofilter conversion requires careful consideration and planning before biofiltration testing or implementation. 4.1.1 Suitability A list of suggested tasks to be completed during the conversion planning phase is presented in Figure 4‐1. Early identification of all benefits, challenges, opportunities, and concerns is a critical first step in considering biofiltration conversion. Specific treatment and operational goals can then be identified and the suitability of biofiltration conversion assessed. The last two steps in the biofilter conversion planning phase include a facility assessment and the determination of any necessary process modifications. Biofiltration conversion planning details can be found in the WRF 4496 final report and related appendices published on WRF’s website (Upadhyaya et al. 2017). An overview of determining conversion suitability through the biofilter conversion assessment tool is provided below.

Biofiltration Guidance Manual for Drinking Water Facilities 55

Figure 4‐1. Tasks to be Completed During Conversion Planning. Source: Upadhyaya et al. 2017.

The suitability of a plant for biofiltration can be evaluated using a wide variety of parameters (e.g., upstream water quality, filter media, EBCT, influent nutrients, etc.). Therefore, to facilitate the evaluation process, The Conversion Assessment Tool was developed to help utilities narrow down what information would be most useful when determining a plant’s suitability for biofiltration. The Tool allows utilities to identify factors that can affect their plant’s biofilter performance during and after conversion. It can also guide utilities in identifying and mitigating common challenges that may occur while converting to biofiltration. The Tool converts the survey responses and facility information to semi‐quantitative biofiltration suitability factors (low, medium, and high) for each of the five categories and generates a Conversion Assessment Report. In addition to summarizing the suitability factors, the Conversion Assessment Report also describes recommendations for improving suitability per category and per question. This provides utilities with a road map for increasing the likelihood of biofilter conversion success while limiting unintended consequences. See WRF 4496 Final Report Appendix D for 1) details on how the tool converts responses to suitability factors, and 2) a listing of conversion suitability improvement recommendations. Finally, the Conversion Assessment Tool also allows utilities to compare their facility to other North American utilities that have converted or are in the process of converting to biofiltration. This may provide a level of confidence in design and operating decisions or may identify concerns that should be further assessed (e.g., if a key design or operating parameter falls well outside what is typical for converted biofilters). However, the Compare My Facility data are static as of June 2015, which will ultimately limit the applicability of a given comparison.

56 The Water Research Foundation

4.1.2 Testing Evaluation of the biofiltration process through bench‐, pilot‐, and/or full‐scale testing is necessary to assess whether long‐term performance goals can be realized. Testing can also provide a better understanding of site‐specific process requirements, limitations, and potential challenges associated with the conversion process, which allows utilities to identify, develop, and test potential mitigation strategies. Finally, testing provides an opportunity to refine the monitoring plan developed during conversion planning. Biofiltration testing design and implementation details are discussed in Chapter 7 of this Biofiltration Guidance Manual. However, some key testing considerations specific to biofiltration conversion are discussed below: (1) Benchmark performance for the existing plant (water quality and hydraulics) over various seasons and operating conditions. Before biofiltration conversion begins, it is important to understand and evaluate existing conditions. Benchmarking helps to establish performance goals and objectives, and outline whether optimization strategies should be considered for testing. (2) Evaluate the scales (i.e., bench‐, pilot‐, and/or full‐scale) that should be considered for conversion testing and what the outcomes and limitation are at each scale. Understanding plant objectives can help to determine which scale would be most important. (3) Depending on the age, condition (e.g., presence of legacy contaminants), or treatment goals, it may be advantageous to test and possibly replace filter media. (4) Understand typical testing durations required to see potential impacts of conversion. The acclimation period (i.e., time required to reach steady‐state biological activity) can be very site‐ specific. (5) Assess the impact of decreasing or eliminating upstream chlorination on the filterability of particles. Since prechlorination can improve particle removal (Becker et al. 2006), it is critical that effluent turbidity trends are closely monitored during biofilter conversion. At the conclusion of testing, it is important to refine monitoring parameters, evaluate the effectiveness of the optimization strategies, and select design and operational parameters for full‐scale implementation. 4.2 Biofilter Conversion Strategies At its most fundamental level, conversion to biofiltration involves creating an environment within an existing filter that is conducive to establishing and maintaining biological activity. Beyond biological activity, several related process and design considerations must be factored to achieve successfully operating biofilters over the long term. 4.2.1 Decreasing Chlorine Dose Conversion from conventional filtration to biofiltration is the result of a singular strategy: decreasing the chlorine (free or combined) dose to a filter to the point that microbial communities can colonize and grow on the filter media. Decreasing the chlorine dose to a filter may involve: 1) decreasing the upstream chlorine dose, 2) relocating the chlorine injection point from upstream to downstream of a filter, 3) switching from inert filter media to GAC, which can quench chlorine, 4) decreasing the concentration of chlorine in backwash supply water either by converting to a non‐chlorinated supply or quenching the chlorinated supply, and/or 5) switching to an alternative primary disinfectant such as ozone.

Biofiltration Guidance Manual for Drinking Water Facilities 57

4.2.2 Addressing Pre‐Loaded Manganese Manganese has two important oxidation states: particulate manganese (Mn+4) and soluble manganese (Mn+2). Particulate manganese is easily removed through sedimentation and filtration processes, whereas soluble manganese requires a dedicated mitigation strategy. The most common strategies for removing Mn+2 are: 1) oxidation (i.e., with permanganate, chlorine dioxide, ozone) followed by particle removal (i.e., sedimentation, flocculation, etc.) and 2) the use of filter media specifically designed to remove Mn+2 through adsorption and oxidation (Lauderdale et al. 2016). In the presence of chlorine, filter media may become coated with manganese oxides (phenomenon known as the “natural greensand effect”), after which continuous application of an oxidant would be required to continue to oxidize the adsorbed Mn+2 and regenerate adsorption sites (Lauderdale et al. 2016). Therefore, the removal of pre‐chlorination would diminish the “natural greensand effect,” reducing the ability of a filter to remove manganese. In addition, interruption of free chlorine application reduces the ORP in the filter which may lead to the reduction of Mn+4 to Mn+2 and the release of Mn+2 in the finished water (Tobiason et al. 2008). Thus, converting to biofiltration (i.e., decreasing the chlorine dose to a conventional filter) can result in the unintended release of manganese from the filter media surface. There are essentially three options to address this concern: (1) Media Replacement. Purchasing fresh media can be costly but provides the surest option for avoiding legacy metals release. (2) Media Washing. Chemically based media cleansing services are available that dissolve legacy metals from media surfaces. This option is rarely chosen, as it involves harsh chemicals, can be costly, and may not completely remove legacy metals. (3) Chemical Enhancement Strategies. Applying various chemical enhancements during and after conversion to biofiltration can mitigate the release of manganese. Phosphorus addition may accelerate the requisite biological activity needed for biological manganese oxidation if phosphorus is limited. Further, increasing pH can help maintain manganese in their precipitated form while requisite biological manganese oxidation activity develops. Lauderdale et al. (2016) performed a biofilter conversion pilot study using anthracite media from full‐scale filters that had historically removed manganese through the natural greensand effect. When prechlorination was terminated without applying any chemical enhancements, they observed rapid manganese breakthrough to above the secondary MCL (0.05 mg/L). In parallel, two filters were operated with a low dose of phosphorus and the influent pH was adjusted from 7.2 to 7.8. Not only did manganese breakthrough decrease, staying at or below 0.05 mg/L, but biofilter hydraulics improved considerably, with 24‐ hour terminal headloss 57% lower relative to the non‐enhanced biofilter. Much more detail can be found on these concepts in the WRF 4448 final report on Optimizing Filter Conditions for Improved Manganese Control during Conversion to Biofiltration (Lauderdale et al. 2016). Depending on the condition of a plant’s existing media (e.g., presence of pre‐loaded manganese) and what that, media is a utility might consider replacing the filter media as a part of the conversion process. However, proper testing should take place before a final decision in made. Details about media selection and change‐ out drivers, such as cost and media characteristics, are discussed in greater detail in Section 3.2.1. 4.2.3 Upgrading Backwash Capabilities Maintaining an optimal biofilm concentration on the filter media is a critical factor for ensuring stable contaminant removal and hydraulic performance. Backwashing is the principal method used for biofilm control and is thought to be the most important operation to maintaining a healthy biofilter. Inefficient backwashing can lead to excess growth, resulting in biofouling and increased head loss, whereas over backwashing can remove too much biofilm, limiting biological activity and necessitating a “re‐acclimation”

58 The Water Research Foundation

period after each backwash. Common challenges that may arise during conversion include changes in hydraulic performance, changes in turbidity trends, and underdrain clogging; where the underlying problem stems from excess biogrowth. Several strategies, specific to conversion, have been developed to help mitigate these potential challenges, thus minimizing uncontrolled growth: (1) Adding air scour or increasing air scour times during a backwash event. Air scour rate and duration can be site‐specific. Table 5‐3 in Chapter 5 provides some guidelines for typical air scour rates and times. (2) Increasing backwash rate and/or duration during a backwash event. Backwash rate and duration can be site‐specific. Table 5‐3 in Chapter 5 provides some guidelines for typical backwash rates and times. (3) Adding an oxidant to the backwash supply. Both hydrogen peroxide and chlorine are used to enhance biomass control during backwash events. Oxidants may also have a detrimental impact on biological activity, particularly chlorine and especially with inert media‐based biofilters. Thus, testing of oxidant‐enhanced backwashing is recommended prior to broad full‐scale implementation. The purpose of the strategies is to create more aggressive backwash conditions where biomass can be removed more easily. These strategies are discussed in greater detail in Chapter 3 of this Biofiltration Guidance Manual. Understanding and monitoring biological growth is critical for these strategies to be effective. For additional information on biological monitoring see Section 2.4 and Table 2‐8 in Chapter 2. 4.2.4 Installing Preoxidation Pre‐oxidation, the process of adding an oxidant before coagulation or filtration, commonly accompanies biofiltration and is often implemented as a part of the biofiltration conversion process. Pre‐oxidation has been known to increase the efficiency of coagulation and filtration processes (Schultz 2014, Ma et al. 2001), improve turbidity, color, UV absorbance, and NOM removal (Becker et al. 2006, Chang and Singer 1991, Jekel 1986, Ma et al. 2001, Amirsardari et al. 1997), and help with the breakdown of NOM into lower molecular weight organic fractions that are readily available to microorganisms, ultimately lowering DBP‐FP (Carlson and Amy 1998, Emelko et al. 2006, Juhna and Melin 2006). Typically, pre‐ oxidants used include chlorine, chlorine dioxide, chloramine, ozone, potassium permanganate, and ferrate (VI). Treatment objectives and anticipated water quality should be considered before the selection and/or implementation of a pre‐oxidant. The application of these preoxidants is discussed in detail in Section 3.2.4. 4.2.5 Modifying Filter Design Depending on existing conditions at a plant, filter modification might also be included as part of the biofiltration conversion process. Historically, there are three filter modifications that a plant might consider: (1) Raising the troughs so that the media bed expansion can increase during a backwash. During biofilter conversion, biomass willn begi to build up on the filter media, which could potentially lead to biofouling issues. Additional media bed expansion may be necessary to improve media bed cleaning, thus mitigating too much biofilm growth. Careful analysis of existing plant hydraulics should be evaluated before modifications are made. (2) Changing the underdrains to a type more suitable for biofiltration, such as those discussed in Chapter 5. Biomass may also accumulate on a filter’s underdrain system. If left unchecked, the underdrain may clog, resulting in an underdrain system fail. Installing a more robust underdrain system, or one less prone to clogging, may avoid underdrain system failures. Underdrains with sintered bead caps are not suitable for biofiltration.

Biofiltration Guidance Manual for Drinking Water Facilities 59

(3) Adding underdrain pressure transmitters that measure headloss across individual filters. If replacing the existing underdrain system is not feasible, monitoring the underdrain backwash pressure and headloss across the underdrain can quickly detect hydraulic issues related to the underdrain system. For example, after a Texas WTP converted to biofiltration, filter hydraulic performance diminished over subsequent years. As shown in Figure 4‐2, UFRVs decreased by approximately 50% from 2001 to 2005, where the largest decreases occurred between 2001 and 2002, coinciding with the ozone/biofiltration system going online. Clean‐bed headloss had also increased over time, resulting in higher water elevations within sthe filter at the start of filter runs.

Figure 4‐2. A WTP in Texas Converted to Biofiltration in 2001 and Observed Rapid Decreases in UFRVs. Source: Lauderdale et al. 2012.

The headloss accumulation across the biofilters eventually led to mechanical failures of the plant’s plastic underdrains with beaded cap media retainers. An autopsy of the failed underdrains/caps suggested that the primary foulant was microorganisms and associated biological materials. Ultimately, the City elected to convert to gravel underdrains. For additional information on biofilter design and applications see Chapter 5.

60 The Water Research Foundation

Greenfield Biofiltration

This chapter discusses planning and design of greenfield biofiltration. When converting an existing conventional filter to a biofilter, the treatment train is adjusted to make a suitable environment for biofiltration (see Chapter 4). When optimizing an existing biofilter, one or multiple enhancement strategies are employed to improve overall biofilter hydraulic and/or water treatment performance (see Chapter 3). Greenfield biofiltration provides the unique opportunity to design around a wider range of parameters that impact biofiltration since detailed filter design as well as upstream processes can be considered. Appendix K provides some examples of greenfield biofiltration plants. 5.1 Planning As with any greenfield WTP, it is essential that sufficient planning and evaluation occur so that raw water quality, water demand, and treatment objectives align with the design and operation of a biofiltration facility. Furthermore, benefits of biofiltration, potential unintended consequences, public perception, and regulatory support must also be considered during the planning stage. 5.1.1 Suitability An assessment of biofiltration suitability begins with a characterization of raw water quality, taking into account seasonal variations. The resulting treatment objectives should then be matched up with the Contaminant Drivers and Removal Potential Table provided in Section 1.2 to determine whether biofiltration could be considered. Next, biofiltration performance drivers (Section 1.5) should be reviewed to determine whether site‐specific conditions may limit the applicability of biofiltration. Lastly, it is recommended that utilities identify other biofiltration plants with similar raw water characteristics and treatment objectives to understand design considerations and operational challenges. 5.1.2 Testing Once biofiltration is selected for full‐scale consideration, testing can be performed to help develop design criteria, identify any process limitations and unintended consequences, and familiarize operations staff with process concepts and protocols. Details on defining testing objectives, selecting and appropriate testing scale, developing testing designs and monitoring plans, managing and interpreting data, overcoming common testing challenges, and understanding expected outcomes can be found in Chapter 7. 5.2 Biofiltration Design Numerous references are available that provide rigorous filter design guidance. This section provides a process and design overview of treatment plant and biofilter elements that impact biofiltration performance and outcomes, including pre‐treatment, filter underdrains, filter trough height, media type and configuration, EBCT, chemical feed, backwash system, chlorine/oxidant addition, and residuals handling. The designer of a greenfield biofiltration WTP is encouraged to pair the information in this section with modern filter and treatment plant design principles found elsewhere. 5.2.1 Pre‐Treatment Upstream processes or pre‐treatment can be designed to promote biofiltration in the filters. Processes can be designed to determine which treatment process a specific contaminant removal takes place.

Biofiltration Guidance Manual for Drinking Water Facilities 61

This optimization can entail adjusting coagulant and pre‐oxidant dosages and types. Table 5‐1 summarizes chemicals that may be added to the treatment process for filters with biofiltration and typical dose ranges. Table 5‐1. Chemicals Added in Biofiltration Applications. Chemical Use Typical Dose Range Filter Aid Polymer Added upstream to reduce filter ripening periods with 0.00‐0.04 mg/L, varies by Type. respect to turbidity removal. Consideration for dosage Goal is to use low dosages and due to headloss accumulation with biological activity when possibleg only durin and challenges cleaning the filter bed when dosages filter ripening. are too high. Chlorine May not be added at all to promote biofiltration or Prechlorination: 0.2 to 1 added intermittently and at low dosages to control 0.5 mg/L Cl2 (acceptable biogrowth and headloss. It can be added to backwash concentration is dependent on water to help maintain uniform hydraulic conditions filter media type, i.e., GAC vs. in the bed and minimize underdrain clogging. anthracite filter media)

1 Backwash: 0.2 to 2.0 mg/L Cl2 (range represents routine backwash with chlorine, higher dosages can be used to recover clean bed head loss)

Ozone Added to feed water to breakdown organic 0.2 to 1.2 O3:TOC; typically 1 compounds into more readily biodegradable 1.0 O3:TOC compounds and increase dissolved oxygen levels. At the top of the filter, the ozone residual should be negligible.

Hydrogen Peroxide Added to feed water or backwash water to help 0.1 to 5.0 mg/L H2O2; typically 2,3,4 maintain uniform hydraulic conditions in the bed and 1.0 mg/L H2O2 minimize underdrain clogging by oxidizing EPS. Phosphorus Depending on filter feed water quality, phosphorus 0.025 to 0.1 mg/L P 2,3,5 may be added to improve hydraulic and water treatment performance. Sources: Lauderdale et al. 2018, (2) Lauderdale et al. 2011, (3) Lauderdale et al. 2014, (4) Nyfennegger et al. 2016, (5) Lauderdale et al. 2016.

5.2.1.1 Coagulation, Flocculation, and Sedimentation A primary goal of direct or conventional filtration is to create a filterable floc that can be efficiently removed through filtration. In addition to solids, coagulation, flocculation, and sedimentation can also remove organics and inorganics, such as influent phosphorus, which can be a limiting nutrient to the biofilters. The ratio of bioavailable organic carbon (C) to nitrogen (N) to orthophosphate (P) is an important parameter for determining the feasibility of biomass synthesis and biological respiration, with a typically referenced C:N:P molar ratio of 100:10:1 (LeChevallier et al. 1991). If nitrogen or phosphorus is limiting, microorganisms may produce more EPS and the conditions may also favor the growth of filamentous bacteria. This, in conjunction with particle loading, can clog filters and filter underdrains, lowering FRTs (Lauderdale et al. 2011). Enhanced coagulation can achieve organic carbon removal through a combination of pH and coagulant dose adjustments. However, coagulation has limitations in its ability to remove smaller molecular weight compounds and uncharged particles (Edzwald 1993). Thus, a combination of coagulation, flocculation, and sedimentation to remove higher molecular weight organics and biofiltration to

62 The Water Research Foundation

remove low molecular weight fractions has been shown to improve water quality (Juhna and Melin 2006). Research has shown in some waters that reducing the coagulant dose in half provided the same NOM removal benefit if the process was paired with biofiltration (Lauderdale et al. 2014). Therefore, where feasible, shifting someC of the TO removal burden to the biofiltration process can be operationally less expensive than coagulation, flocculation, and sedimentation, and can provide cost savings to the utility. 5.2.1.2 Pre‐Oxidation Details on implementing an oxidation step upstream of biofilters can be found in Section 3.2.4. 5.2.1.3 Downstream Free Chlorine Contact Biofilter effluent can contain biomass detached from the filter media and soluble microbial products, especially early in a run following a backwash. This can result in increased disinfection demand and affect disinfectant stability if monochloramine is used (Marda et al. 2008). To address this, a short free chlorine contact time (CT;) e.g., 1 min can be implemented downstream of the biofilters prior to final disinfection. As noted in Section 3.2.1, utilizing a lower layer of media such as sand (with a higher specific gravity and lower effective size) can decrease the amount of biomass leaving the biofilter.

5.2.2 Filter Underdrains The filter underdrain is an important design element for biological filtration. The purpose of the filter underdrain is to retain and support the filter media during filtration and to provide an effective means of backwashing the filter. Without a well‐functioning underdrain, the biofilter’s performance will quickly degrade. Important factors in selecting and designing an underdrain system are:  Resistance to biofouling/plugging at the media retention layer.  Backwash features such as air scour, necessary for proper backwash of modern deep media designs.  Robust, capable of operating for 20+ years between rehabilitation cycles.  Excellent air and water flow distribution during backwash.  Pressure monitoring and over‐pressure protection.  Tolerant of moderate levels of chlorine or other oxidants, should chemical cleaning become necessary.  Corrosion resistant to environment present in biofilter. Historically, engineers designed filter underdrains using pipe laterals and gravel as shown in Figure 5‐1. These underdrain designs have fallen out of favor as they do not accommodate air scour and are susceptible to gravel displacement over time. The underdrain can fail after the gravel becomes displaced.

Biofiltration Guidance Manual for Drinking Water Facilities 63

Figure 5‐1. Example of a Historical Gravel Underdrain System. Two common types of modern underdrains used for biofiltration are presented below: • Block Underdrains. The block underdrain system consists of a series of plastic blocks that form channels to collect filtered water and distribute the backwash water and air. This type of underdrain is generally used with a media retaining cap. Otherwise, a specifically designed system of 12‐18 inches of layered gravels can be used to retain the media. Gravel media retention is less susceptible to biological fouling (Alito and Wangner 2018). However, if designed or operated incorrectly, the gravel can easily become displaced, quickly resulting in underdrain plugging and failure.

Block underdrains with media retaining caps (Figure 5‐2) have a lower profile than most underdrain types. Block underdrains typically provide excellent backwash water and air distribution across the entire filter area. It is preferred to feed the block underdrain through the end of the laterals rather than through the bottom via a flume, to avoid weaknesses and failures inherent in some bottom‐ fed design configurations.

Figure 5‐2. Example of a Block Underdrain System. Source: Leopold. • Nozzle Underdrains. Similar to block underdrains, nozzle underdrains are typically configured without the use of gravel. Figure 5‐3 shows a typical nozzle underdrain system. Nozzle sleeves are typically cast into concrete floor and become part of the filter structure. Similar to block underdrains with media retaining caps, nozzle underdrains typically do not use gravel for media retention. As such, they can be susceptible to biological clogging.

64 The Water Research Foundation

Figure 5‐3. Example of a Nozzle Underdrain System. Source: Orthos Liquid Systems.

Because underdrains can be susceptible to biofouling or scale formation, careful attention to underdrain selection and construction is important. Manufacturers have improved the integrity of anchoring systems due to concerns over biofouling. Moreover, pressure transmitters should be installed to measure headloss across individual filters and alarm if headloss trending shows fouling is occurring. An increase in clean‐bed or backwash headloss over time can be an early indicator of underdrain fouling.

Figure 5‐4 shows an example of a clean and clogged underdrain media cap. Figure 5‐5 shows a failed porous plate underdrain cap. Manufacturers typically no longer offer porous plate underdrain caps such as these for biological filtration due to the potential of clogging and failure. Additional protocols in Chapter 2 detail appropriate process monitoring to avoid a failed underdrain system. When selecting a commercially available underdrain, it is important to 1) select an underdrain that has been successfully used in biological applications, and 2) contact references of similar installations.

Figure 5‐4. Cleaned (left) and Clogged (right) Porous Plate Underdrain Cap.

Biofiltration Guidance Manual for Drinking Water Facilities 65

Figure 5‐5. Failed Porous Plate Underdrain Caps: Stripped Anchors (left) and Blown Mastic (right). 5.2.3 Trough Height Above Media The design for the trough height above media is based on the desired filter media expansion during backwash. The typical design criteria for the trough height is for 30 to 50% bed expansion during the high‐rate backwash plus 6‐18 inches. The higher bed expansion design value provides future flexibility to add media if desired and to provide longer simultaneous air scour/hydraulic backwash. The filter design for trough height is not specific or unique to biofiltration but matches typical design practices for rapid rate granular media filters. 5.2.4 Media Type and Configuration Filter media serves two purposes in biological drinking water treatment: particle removal (filtration) and physical support for the growth of microbial communities. Optimizing filter media design criteria, such as media type, depth, and structure should be taken into consideration to achieve performance objectives. Refer to Chapter 3 for details on optimization strategies. This chapter focuses on media design for greenfield applications, which will impact filtration performance, biological activity, adsorptive capacity, and cost. 5.2.4.1 Media Type The most common medium types of filter media are GAC, anthracite, and sand. The majority of biofilters are comprised of dual media, with GAC or anthracite over sand, though several biological plants have operated successfully using mono‐media (Upadhyaya et al. 2017). Common biofilter media characteristics are summarized in Table 5‐2, and drivers for media selection are provided in Section 3.2.1. The designer should select the proper media size, uniformity coefficient, and configuration based upon pilot test results or a nearby facility operating with similar water quality characteristics and treatment approach. Table 5‐2. Common Biofilter Media Design. Uniformity Apparent Density Media Effective size (mm) Coefficient (UC)1 (gram/mL) Porosity (%) Sand 0.4 – 0.8 1.3 – 1.7 2.65 40 – 43 Anthracite 0.8 – 2.0 1.3 – 1.7 1.4 – 1.8 47 – 52 GAC 0.8 – 2.0 1.3 – 2.4 1.3 – 1.7 N/A (1) Lower UCs are typically preferred, as they result in lower clean‐bed headloss

66 The Water Research Foundation

5.2.4.2 Configuration The primary function of filtration is to remove turbidity; therefore, particle removal must drive design. In general, the design parameter to determine the filter depth is the ratio of depth of the biofilter to ES of the filter media. Typical L/d ratios are provided by AWWA (1998) for mono‐d, dual‐ an tri‐media filters to achieve acceptable particle removal. These ratios are provided as guidance for conventional filtration systems but are also applicable for biofiltration systems. Generally, conventional practice is to design a filter with a minimum L/d of 1,000. As noted in Section 3.2.1, utilizing a lower layer of sand is often desirable as it can improve filtered water turbidity. 5.2.5 Empty Bed Contact Time and Filter Loading Rate EBCT is a metric that describes the length of time water is in contact with the filter media. It is calculated by dividing the total volume of media bed (including pore space) by the flow rate across the bed and is expressed in terms of minutes. It is well documented that EBCT is one of a few critical variables controlling the degree of contaminant biodegradation (Krasner 1993, Nugroho, 2010). In many cases (e.g., biofilter conversions or optimizations), EBCT is not an actual design parameter due to the fact that the existing filter box structure limits bed depth changes and thus changes in EBCT. However, when designing a greenfield biofiltration plant, pilot testing is typically used to identify an optimal EBCT (i.e., achieves treatment goals at minimized cost), which can then be used to design the new filter box structure. Contaminant surface loading rate (i.e., the mass loading of a given contaminant per unit time per unit medium surface area is a more fundamental biofilter design parameter and is described in detail in Chapter 17 of the American Water Works Association's/American Society of Civil Engineer's Water Treatment Plant Design book (Rittmann et al. 2020). Similar to EBCT, there is limited opportunity to adjust this parameter for existing biofilters. EBCT is strongly correlated with the rate of biological removal. In general, EBCT ranges from 2 to 30 minutes. Short EBCTs (<2 minutes) do not provide sufficient contact time for removal of influent contaminants and longer EBCTs (>30 minutes) provide diminishing returns on removal of influent contaminants. Typical EBCTs are 5 to 15 minutes. Methods to evaluate the impact of EBCT on biofilter performance are discussed in Section 2.2.4. It is important to note the inverse relationship between EBCT and filtration rate. Doubling the filtration rate will half the EBCT. Therefore, there can be a temptation to reduce the size of the filtration facility by increasing both the filtration rate and the media depth (to achieve the desired EBCT). This can result in filters that are excessively deep. Designers should be cautious of high filtration rates when high EBCTs are also required, otherwise backwashing may be frequent and efficiencies low as the top portion of the media quickly reaches terminal headloss. 5.2.6 Chemical Feeds In general, an important benefit of biofiltration is that typically, little to no chemical addition is required, it is a naturally occurring process. However, as discussed in Chapter 3, biofiltration performance may be enhanced through pre‐oxidant addition, nutrient augmentation, and/or pH adjustment upstream of the biofilter, which may require dedicated chemical storage and feed systems. These feed systems may already be a part of the treatment plant design for other purposes, such as corrosion control or disinfection. Filter aid polymer may also be dosed upstream of biofiltration to promote removal of fine particles and to minimize filter ripening periods. Biofiltration often will increase the amount of headloss formation in a filter. This is the trade‐off of the benefit of additional organic carbon removal. Adjustment of filter aid polymer dosing may be required to reduce the amount

Biofiltration Guidance Manual for Drinking Water Facilities 67

of headloss accumulation. Filter aid polymer can cause mudball formation within the filter. Therefore, dosages should be minimized when used. 5.2.7 Backwash System Backwashing is essential conventional filter maintenance to remove accumulated particles. For biological filters, backwashing removes accumulated particles and biomass. Optimal backwashing removes particles and biomass to clean the bed and reduce headloss but preserves enough biomass that steady‐state operation can resume afterwards to meet treatment objectives. Example backwash parameters are provided in Table 5‐3, and a detailed discussion of backwash protocol considerations is provided in Section 3.2.2. Table 5‐3. Example Backwash Protocol. Backwash Step Timeframe and Flowrate Air scour 1‐4 minutes at 2‐5 standard cubic feet per minute (scfm)/sf Simultaneous air scour/ hydraulic backwash 2‐4 minutes of air at 2‐4 scfm/sf and water at 4‐9 gpm/sf Low‐high transition 1 minute High‐rate backwash 6‐12 minutes at 15‐25 gpm/sf (to achieve 30‐50% bed expansion) Backwash ramp down 1‐2 minutes Filter‐to‐waste or rinse to waste Duration based on filtrate turbidity and filter efficiency goals. Filter‐to‐waste flow rate should match operational flow rate. Filters are typically equipped with supplemental media washing systems to improve the efficiency of the backwash. Air scour is the most common form of supplemental backwash for biofiltration. Air scour is typically used for media designs of 48 inches or deeper. This is because the air travels the entire depth of the bed. It is important to note that air scour alone is not as good at cleaning the media as simultaneous air scour/hydraulic backwash. In challenging biological designs, additional freeboard between the launders and media may be needed to provide an extended simultaneous air scour/hydraulic backwash. Alternatively, media retaining launders can be used, which allow for indefinite simultaneous air scour/hydraulic backwash. Although not as common, fixed‐grid surface wash is also an excellent backwash enhancement when the media depth is less than 36 inches. 5.2.8 Chlorine and Oxidant Addition Chlorine and oxidants have been added to backwash water to improve backwash efficiency. It is recommended to design the backwash system with injection ports for the flexibility to dose chlorine or oxidants to the backwash. An example backwash system that can supply chlorinated or unchlorinated water is shown in Figure 5 ‐6. In Figure 5‐6, a gate is located prior to the final chamber of the clearwell which can be opened for a chlorinated backwash or closed so the backwash bypasses the intermediate chambers where chlorine is mixed.

Figure 5‐6. Process Flow Diagram of Backwash System with Unchlorinated and Chlorinated Backwash Capability.

68 The Water Research Foundation

Chlorinated backwash has been reported to be used frequently or only when there is observed high headloss (Brown et al. 2016). Of the reporting utilities in the Biofiltration Knowledge Base, 24% use a chlorinated backwash either when biofouling occurs or regularly. If chlorination is implemented in the full‐scale system, it is recommended for it to be thoroughly tested in pilot testing, as research indicates mixed results for the effect of chlorinated backwash water on biofilter performance. Hydrogen peroxide is not common practice and has mixed performance during testing (see Section 3.2.2.3); thus, careful testing and monitoring of oxidant type and concentration should also be conducted if dosed in the backwash. Further discussion of chlorine or oxidant addition to backwash is discussed in Chapter 3. 5.2.9 Residuals Handling Compared to ion exchange or membrane filtration, biofiltration will produce less waste for residuals handling because contaminants are typically converted to innocuous end products (with the exception of some trace organic contaminants). Biofiltration can reduce the amount of primary coagulant or powdered activated carbon needed to achieve a similar organics reduction, which directly impacts solids handling. If using biofiltration for oxidation of inorganics, such as iron or manganese, the residuals management handling process is an important consideration. Bacteria can oxidize dissolved iron and manganese to form iron and manganese precipitates that are filtered out in the biofiltration process. Depending on the solids handling process, anoxic conditions can be introduced during solids thickening and storage, and iron and manganese can be released back into the dissolved state. If supernatant is returned to the head of the WTP, such as in a zero‐liquid discharge (ZLD) facility, then this recycle stream can be a source of manganese in the influent. 5.2.10 Hydraulics Hydraulics for a green field biofiltration facility are similar to other filtration facilities. However, due to the potential additional headloss contribution from the biological activity, additional head may be needed. This is where pilot data can be of significant assistance. When designing the filter depth and media placement within the hydraulic profile, it is important to place the media low enough in the hydraulic profile that the top of the media always has positive head, even when the filter is at the end of the filter run, and headloss is the greatest. The downstream hydraulics should not allow negative head below the top of the media.

Biofiltration Guidance Manual for Drinking Water Facilities 69

70 The Water Research Foundation

Operation and Maintenance

This chapter summarizes key operational and maintenance considerations for biofiltration. The majority of operational and maintenance considerations are also applicable to conventional filters. Chapter 2 contains detailed information on biological filter monitoring and recommended tools including filter integrity, hydraulic, water quality, and biological monitoring techniques. Important complementary tools for this chapter can be found in Appendix C (Biofiltration Calculations), Appendix D (Biofilter Operations Checklist), and Appendix E (Biofilter Troubleshooting Guide). 6.1 Monitoring There are no regulated monitoring requirements specific to biofiltration beyond those in place for conventional filtration (Nieminski and Perry 2015). However, to achieve effective biofilter performance while minimizing negative unintended consequences, specific biofilter monitoring strategies are strongly recommended. These strategies are discussed in Chapter 2. 6.2 Start‐Up and Acclimation Acclimation is the time required to establish a mature and active biomass in the biofilter after start‐up. Acclimation first includes colonization of the biofilter media followed by adaptation of the microbial community’s metabolism to degrade BOM or other target contaminants (e.g., MIB, geosmin, ammonia, manganese). While the colonization may occur within weeks, establishing a mature and active biomass capable of degrading the contaminants of interest may take several months. In general, faster acclimation can be expected for BOM degradation compared to inorganic contaminant (e.g., ammonia, iron, and manganese) removal. Initially, microorganisms may not express enzymes specific to the contaminants of dinterest an may require considerable exposure time for producing the enzymes. A fully acclimated biofilter will demonstrate steady‐state contaminant removal and relatively stable hydraulic performance. The length of time required to reach steady‐state removal is also dependent on the influent water characteristics and operating conditions, such as:  Filter loading rates/EBCT.  Backwashing procedures.  pH.  Disinfectant/oxidant residual.  Upstream process operation (e.g., ozone, coagulation and pre‐chlorination).  Filter media characteristics.  Availability of nutrients.  Temperature. 6.3 Steady‐State Operation Once a mature and active biomass is continuously degrading target contaminants, steady‐state operation is achieved. However, even under steady operational conditions, changes in water quality or upstream process operations can impact biofilter performance. Thus, a strategic monitoring plan is essential. Chapter 3 discusses hydraulic, water quality and biological monitoring tools and a method to develop a Monitoring Strategy or Filter Monitoring Plan.

Biofiltration Guidance Manual for Drinking Water Facilities 71

6.4 Shutdown, Idling, and Restart Similar to most drinking water processes, biological filters operate best under steady conditions. However, required maintenance, low flow conditions, or process upsets may require moderate or extended biofilter shutdown periods. In these cases, the health of the microbial community may diminish due to lack of fresh substrate (i.e., food); this may necessitate a reacclimation period upon restart. Furthermore, the in‐bed redox potential will decrease during shut‐downs, as oxygen is consumed by the microbial community. Depending on the shutdown duration, this may resolubilize previously precipitated metals (e.g., iron and manganese) and/or generate unpleasant odors, thereby necessitating an extended backwash flushing period prior to restart and/or an extended filter‐to‐waste period upon restart. A few parameters that can affect the degree to which a shut‐down period can impact biofilter performance are listed below.  Duration. Without a proactive biofilter idling strategy (discussed below), the longer the shut‐down period the greater the impact on the microbial community and in‐bed redox potential.  Water Temperature. Warmer temperatures (e.g., > 15oC) will likely improve reacclimation kinetics, thereby shortening the reacclimation period. However, warmer temperatures may also result in more rapid degradation of dissolved oxygen, thereby decreasing the redox potential and increasing the rate of any metals leaching from the media bed.  Water Quality. There are several ways that water quality can impact biofilter performance after a shut‐down: 1) Increased substrate loading to a biofilter while in service will generally result in a more active and robust microbial community in the bed, which can shorten the reacclimation period following a shutdown, 2) Metals loading to a biofilter while in service introduces the concern over metals leaching upon restart, especially after an extended shutdown, and 3) If the biofilter influent water quality/source upon restart is different than it was before the shut‐down, the microbial community may shift, thereby taking more time to reach pre‐shutdown performance.  Media. Relative to inert media, GAC tends to support higher biomass concentrations and its surface roughness can promote the establishment of particularly resilient biofilm communities. Thus, similar to increased substrate loading, GAC‐based biofilters may require a shorter reacclimation period relative to inert media‐based biofilters. Various biofilter idling strategies can be considered that may shorten reacclimation periods and help minimize negative unintended consequences associated with biofilter shutdowns: (1) Minimize Shutdowns. Since biofilters operate best under steady conditions, they should be operated continuously to the extent possible. When water availability or production demand limits flow, operating more filters with lower flow rate is generally recommended rather than shutting down a portion of the filters. This will help maintain redox conditions and biological activity in the bed. (2) Record Detailed Baseline Performance Prior to a Shutdown. A detailed characterization of water quality, active biomass levels (e.g., ATP), and hydraulic performance prior to a shutdown is key for quantifying any shutdown impacts and for determining when biofilter performance is back to ‘normal.’ (3) Consider the Impacts of Wet vs. Dry Storage. After shutting down a biofilter(s), some utilities drain and store the media essentially dry while others leave water in the filter box above the media height. Dry storage may prevent the development of anoxic conditions and the associated metals dissolution. However, complete desiccation will effectively inactivate the microbial community,

72 The Water Research Foundation

thereby extending the reacclimation period. During wet storage, metals release may be an issue if aerobic conditions are not maintained, but reacclimation periods are expected to be shorter. Furthermore, during wet storage algal growth may occur, necessitating additional mechanical (e.g., backwashing) or chemical (e.g., oxidant addition) strategies for removal prior to restart. The optimal wet/dry storage plan for each utility should be identified based on the above considerations along with full‐scale testing. (4) Maintain Aerobic Conditions. During a wet‐storage shutdown, consider monitoring DO in the water column just above the bed. Strategies for maintaining aerobic conditions during a biofilter shutdown include: 1) Practicing daily aeration through the air scour system (~15 minutes, and may need to be done more than once daily), 2) Draining the filter box to waste and refilling with fresh settled water; this not only increases redox potential in the biofilters, but also replenishes substrate and nutrients, or 3) Practicing regular (e.g., daily) low‐rate backwashes (without any oxidant); depending on the backwash supply source, this may also increase redox potential and provide fresh substrate and nutrients. (5) Evaluate Restart Options. The optimal backwash aggressiveness and filter‐to‐waste duration following a biofilter shutdown will be site‐specific. It will depend on 1) the parameters discussed in Section 6.4 (shutdown duration, water temperature, water quality, and type of biofilter media used), 2) the idling strategy used during the shutdown, and 3) biofilter treatment objectives If metals dissolution is expected or detected, then a more aggressive backwash and/or filter‐to‐ waste duration may be prudent; a higher pH may also be considered in this case to help maintain metals in a precipitated state (see Chapter 4). Monitoring metals in the backwash wastewater can help assess potential metals breakthrough upon resuming filter operation. If metals concentration in the backwash wastewater is above a threshold level (e.g., 0.3 mg/L Mn) determined through testing, then additional backwashing should be performed. However, if the microbial community is exposed to limited substrate and oxygen during an extended shutdown, an overly aggressive backwash may unnecessarily flush out most of the community, substantially extending the required reacclimation. period Full‐scale demonstration testing is recommended to determine an effective and efficient biofilter restart protocol. Testing should also factor potentially increased levels of metals in the backwash wastewater and what impact that might have on recycle and discharge options.

Biofiltration Guidance Manual for Drinking Water Facilities 73

74 The Water Research Foundation Biofiltration Testing

Well‐designed filter testing facilitates the optimization of design and operation, develops more accurate capital and operating cost estimates, and improves WTP performance. This chapter is structured as a guide to best practices for designing and implementing a robust biofiltration testing program. Each phase of testing (i.e., desktop evaluation, bench testing, pilot testing demonstration testing) provides utilities an opportunity to answer key questions regarding biofiltration performance and design requirements and each scale offers unique benefits. Figure 7‐1 summarizes key question(s) a well‐ designed testing plan should answer and refers to the relevant Chapter 7 sections.

7.1 Define Testing 7.2 Benchmarking Water 7.3 Selecting Testing Scale(s) Objectives Quality and Treatment • Is a desktop evaluation, • What water quality and/or Characteristics bench‐, pilot‐ or operational goals am I trying • What water quality and demonstration‐ or full‐scale to achieve? treatment characteristics do testing most appropriate for I need to account for? my goals?

7.4 Designing a Desktop 7.5 Designing Bench, Pilot, 7.6 Overcoming Common Evaluation or Demonstration Tests Testing Challenges • Is biofiltration suitable for • What should my testing plan • How can I further improve my facility? include? my testing plan? • How do I develop a business • What equipment do I need? case for conducting • How do I assess biofiltration testing? performance?

7.7 Understanding Expected 7.8 Resource Planning Outcomes • What resources do I need to • When should I move to the allocate towards testing to next phase of testing? ensure I meet my goals? • When is testing complete? • How do I manage cost?

Figure 7‐1. Key Testing Plan Questions Addressed in This Chapter.

7.1 Defining Testing Objectives Well‐designed testing programs are based on specific and measurable objectives for biofiltration (refer to Section 1.2). These objectives will often help define full‐scale performance goals when testing is complete. Figure 7‐2 can be used to help craft a list of objectives. Start by defining the key questions to be answered, and then select the factors and parameters under investigation. For each factor or parameter, define a specific performance objective. Consider objectives for both conventional filtration and biological degradation. Lastly, include relevant operational challenge conditions that represent the range of water quality and operational conditions under which each objective must be tested. Sample biofiltration testing objectives are provided in Appendix L.

Biofiltration Guidance Manual for Drinking Water Facilities 75

Biofiltration Testing Generally Answers the Following Questions

How can I improve Why does this work this way? Will this work? performance? (Research Studies)

Evaluating Biofiltration Optimization of Biofiltration Mechanistic Understanding

Suitability Performance of Biofiltration

By Testing One or More of the Following Factors

Biofilter Design Biofilter Operation Contaminant Removal

 Media type   Filter hydraulics (e.g.,   Particulates (e.g.,  Media depth headloss) turbidity)  Backwash protocols  Water production (e.g.,  Natural Organic Matter (rates and durations, surface loading rate, run (e.g., TOC, DOC, AOC, surface wash, air scour, time, backwash water UV254) use of disinfectants) volume)  Inorganic compounds  Filter‐to‐waste  EBCT (e.g., dissolved metals)  Filter influent chemical  T&O compounds addition  Emerging contaminants (e.g., nitrosamines)  Disinfection byproducts (DBPs) Figure 7‐2. Common Biofiltration Testing Objectives.

7.2 Benchmarking Water Quality and Treatment Characteristics Benchmarking is the process of understanding historical water quality and current or planned treatment characteristics. This information will inform the testing plan, help establish or refine testing goals and define the correct scale(s) of testing. For new WTPs, benchmarking may include modeling of potential biofilter performance and estimating the potential for organic carbon biodegradation. 7.2.1 Historical Water Quality and Performance Table 7‐1 summarizes recommended water quality and operational data analysis prior to beginning any testing. Critical parameters are recommended for basic biofiltration testing, as well as additional parameters that may be necessary depending on testing goals. If critical data are not available or not available in sufficient quantity (e.g., two years of data covering anticipated operational conditions and seasons), then consider conducting targeted full‐scale sampling for a minimum of six months to capture both winter and summer seasons. Refer to Chapter 2 for details on the impact of each of these parameters on biofiltration and suggested analytical techniques. Historical filter performance data (e.g., filter headloss over run time) may also be reviewed to determine a baseline for filter performance goals.

76 The Water Research Foundation Table 7‐1. Recommended Historical Water Quality and Operational Data. Minimum Optional Additional Recommended Data Parameters (Depending on Category Critical Parameters Frequency Testing Goals) Location Operations Surface loading rate Online Filters Headloss – Run time UFRVs General Water Temperature Weekly Total dissolved solids Filter influent Quality pH Zeta potential Alkalinity Nitrite Hardness Nitrate Ammonia Disinfectant residual Conductivity ORP Color Orthophosphate Particulates Turbidity Daily Particle counts Filter influent Total suspended solids and effluent Organics TOC and DOC Monthly UVA254, AOC, BDOC (Calculate removal) Taste and Odor MIB Monthly TON (T&O) compounds Geosmin Inorganics (total Iron Monthly Aluminum and dissolved) Manganese Lead Copper Emerging As available Nitrosamines Contaminants – Algal toxins PPCPs Disinfection TTHMs Quarterly Bromate (critical if plant Distribution system Byproducts HAAs uses ozone) Biological ATP ATP: Monthly EPS Filter media Microbial community analysis Enzyme activity Biofilm formation Biofilm formation rate: Filter influent and rate OR Every other week effluent (Calculate DO consumption*, DO consumption: change from influent DOC removal Online to effluent) DOC removal: Monthly Note: * DO consumption is most valuable for facilities without intermediate ozonation (i.e., ozone directly upstream of biofiltration that may result in supersaturated DO conditions in the biofilter influent).

At a minimum, benchmarking should include analysis of temporal and seasonal trends. Analysis should also consider the range of water quality and performance (e.g., 5th percentile and 95th percentile) over the past two or more years. Figure 7‐3 highlights several tools available to aid in this (and future testing data) analysis.

Biofiltration Guidance Manual for Drinking Water Facilities 77 Biofilter Description: Excel‐based tool that assesses the relative suitability of biofiltration based Conversion upon 24 multiple choice or select all that apply questions (refer to Appendix G) Assessment Tool Best Application: Initial business‐case evaluations of biofiltration, or identification of optimization strategies for testing.

Statistical Description: Software that conducts statistical analysis of data and generates a variety Analysis Software of graphics (Microsoft Excel, Matlab, R Studio, Proformance (refer to Appendix G), etc.) Best Application: Quick analysis of static data, such as historical data review, or when data files are to be transferred to another user with the same software. Note that some of these tools may auto‐update once initial programming is completed.

Data Analytics Description: Advanced software that allows for visualization of data through Software interactive, live dashboards and reports (Tableau and Microsoft Power BI, etc.)

Best Application: Ongoing data analysis where new data are generated and analyzed frequently, or when stakeholders need the ability to interact and manipulate data.

Figure 7‐3. Data Analysis Tools.

New treatment facilities or facilities considering upgrading upstream treatment processes in concert with implementing biofiltration may not have historical biofilter influent water quality data and may not be able to conduct a sampling campaign. In these cases, it may be necessary to first estimate the biofilter influent water quality prior to considering potential biofilter performance. The following tools are available to estimate the biofilter influent water quality: • U.S. EPA WTP Model – The WTP Model is a program that estimates NOM removal and DBP formation through drinking water treatment. • EPA Drinking Water Treatability Database – The Treatability database provides data on removal of contaminants by treatment process. • RTW Model – The RTW model estimates changes in water quality and corrosion indices with different chemical treatment strategies. • Jar Testing – Upstream treatment through coagulation and sedimentation can be evaluated using bench jar tests. • Alternative Source Water Assessment Tool (ASWAT; WRF 4665) – The ASWAT is an Excel‐based tool that estimates finished water quality based upon entered raw water quality and facility design factors. • AWWA Water Quality and Treatment: A Handbook on Drinking Water (Edzwald 2011, Rittmann and McCarty 2001; Rittmann et al. 2020) – Includes essential engineering and operations guidance on drinking water treatment processes.

78 The Water Research Foundation

7.2.2 Estimating Potential Biological Organic Carbon Removal New biofiltration facilities may run analyses to characterize the quality of organic carbon that can be assimilated into biomass or biodegraded. These include the AOC and BDOC assays, respectively, as described in Section 2.3.2.2. AOC is a two‐species bioassay used to determine the potential for biological regrowth (e.g., biostability). Samples are pasteurized to inactivate indigenous organisms and then inoculated with Pseudomonas fluorescens strain P17 and Spirillum strain NOX. These bacteria were selected because they are heterotrophic (use organic carbon for energy) and oligotrophic (can survive in water with low concentrations of carbon). The test organisms are allowed to grow to maximum density and then enumerated by the spread plate method. BDOC measures the DOC removed over time in a batch reactor over a five‐ to seven‐day period (Mogren et al. 1990; Allgeier et al. 1996; Joret et al. 1991, refer to Appendix G). While the results of the BDOC test do not capture expected biodegradation under hydraulically relevant residence times across a full‐scale biofilter (Terry et al. 2018), the BDOC test could serve to confirm that biological treatment, in general, is a feasible treatment strategy. If BDOC is detectable (i.e., above the approximately 0.2 mg/L detection limit noted by Escobar and Randall 2001), further bench, pilot and/or demonstration testing of biofiltration would be recommended. 7.2.3 Filter Design and Operation Prior to designing an experiment, it is important to understand the current design and operational characteristics of the filters (or planned filters). Key design and operational factors that should be considered include (refer to Appendix D for biofilter calculations): Plant design and average flow - Duration of filter‐to‐waste, if  Filter design available - Number of filters - Rate and duration for each - Surface area per filter backwash phase (e.g., low and - Media type(s) high) - Media depth per layer - Bed expansion - Media ES - Terminal headloss trigger for - Distance from top of media to backwashing bottom of troughs - Effluent turbidity trigger for  Select filter design calculations backwashing - Design and average filter - Run time trigger for loading rate backwashing (provide per - Design and average EBCT season, if applicable) - L/d10 ratio  Items to collect - Design available bed expansion - Full‐scale process flow diagram space - Doses of upstream oxidation  Filter backwash design1 processes and oxidant residual - Air scour or surface wash - Filter design drawings - Air scour rate, if available - Filter photos - Air scour/surface wash duration - Filter operations SOP - Backwash pressure or pressure - Backwash SOP through the underdrain during - Descriptions of any planned a high‐rate backwash modifications

1 – Filter backwash design can only be tested at demonstration‐scale.

Biofiltration Guidance Manual for Drinking Water Facilities 79

7.3 Selecting Testing Scale(s) The most successful testing strategies utilize multiple testing scales, described in Table 7‐2, in progressive testing phases. Desktop evaluations help to define testing objectives and identify testing conditions likely to meet those goals. Testing of multiple design or operational parameters is best suited for bench scale. This allows for a more controlled environment and testing of multiple strategies or variables. Bench testing design may focus on determining broadly applicable principles (i.e., research‐ based bench studies) or evaluating performance for a specific water system (i.e., treatability studies). As optimal strategies are identified, testing should progress to scales more reflective of current or anticipated full‐scale process design and operating conditions (i.e., pilot and demonstration studies). The accuracy of scaling‐up testing results increases as you proceed from laboratory bench testing to demonstration testing. Photos showing the physical differences between the testing scales are presented in Figure 7‐4.

Table 7‐2. Description, Objectives, and Limitations of Each Testing Scale.

Order of Increasing Relevancy to Full‐Scale Operations

Scale Desktop Bench ‐ Research Bench ‐ Treatability Pilot Demonstration  Literature Laboratory conditions: Relevant operating Operational conditions: reviews  Conditions are strategically controlled to conditions:  Conditions match  Modeling allow for testing of specific variable(s)  Conditions full‐scale operations directly  Tests or matches represent full‐scale current or backwashing anticipated full‐

protocols scale process design Description

 May represent  May represent  Tests seasonality typical process site‐specific  Tests water production and compliance design, but likely process design (e.g., effluent turbidity) not site‐specific and/or use site‐ specific water quality  Project  Often not  Prior to conversion or implementation Conception connected to  Transitioning during conversion

implementation or  Business Case  Optimizing existing biofiltration (refer to Chapter 4) Development optimization of biofiltration at a Potential Timing specific WTP

80 The Water Research Foundation

Table 7‐2. (Continued)

Order of Increasing Relevancy to Full‐Scale Operations

Scale Desktop Bench ‐ Research Bench ‐ Treatability Pilot Demonstration  Evaluating  Developing  Testing many  Testing a select  Conduct potential for mechanistic different number of optimization testing biofiltration to understandings strategies strategies over a of a single strategy meet specific (i.e., provides a independently longer duration to existing biofilters goals highly‐controlled and quickly (e.g., one year) (e.g., addition of a  Identifying and easily  Investigating to capture the specific chemical or potential manipulated site‐specific impact of varying EBCT) for a optimization environment) contaminant uncontrolled lower cost, if

strategies biodegradation variables, such modifications can be  Potential as seasonal made full‐scale at a variables include water quality low cost to temperature, and operating complete the testing EBCTs, pre‐ conditions and isolate a filter or Objective(s) oxidant dose or  Establishing set of filters. oxidant:TOC expectations for  Increase operator ratio, and media full‐scale familiarity with the type. contaminant design and removal and operation of the hydraulic selected strategy performance  Developing design criteria  Actual  Does not match full‐scale filter  Generally, only test a limited number of conditions may production or hydraulic performance strategies to control costs (more expensive) not match  Not useful for turbidity removal or  Limited control over changes in the influent models or full‐ backwash approach assessment water quality (need a control for scale data in  May not match full‐scale contaminant comparison) ‐ potential for confounding literature removal performance (e.g., as water factors to confuse the clarity of results

 Limited ability quality changes seasonally)  May not capture all water quality variations to estimate the  Not scalable, except for biodegradation (if duration is a year or less duration) impact of  May require a permit for discharge of  Space, flow,  Highest costs changing water effluent energy, etc.  Space, flow, energy, quality or Limitations requirements etc. requirements operating  May require a for side‐stream Key conditions permit for demonstration‐scale discharge of filters effluent  If effluent is sent to distribution, must ensure all regulatory requirements are met

Biofiltration Guidance Manual for Drinking Water Facilities 81

0.16 gpm Bench Biofilters used in WRF 4749 (Evans et al. 2018) Filters: 12 Footprint: 1“ diameter each

Courtesy: Alameda County Water District

6 gpm Pilot Biofilters used in WRF 15‐11 (Funk et al. 2018) with Human Machine Interface (HMI) Automation (left) and Filter Pilot (right) Filters: 4 Footprint: 6“ diameter HMI each

Courtesy: Gwinnett County

1 MGD Demonstration Biofilters used in WRF 4155 (Chowdhury et al. 2010) Filters: 4 Footprint: 50’x50’ each

Courtesy: Birmingham Water Works Board

Figure 7‐4. Photos of Bench, Pilot, and Demonstration Testing Facilities.

82 The Water Research Foundation

Testing scales should be selected based on testing goals (Section 7.1). Common testing objectives are summarized Table 7‐3 with the relative capacity of each testing scale to assess each objective. Table 7‐3 can be used as a tool to select the appropriate testing scale(s) by highlighting each row with an applicable testing goal.

Table 7‐3. Testing Scale Selection Based Upon Common Testing Objectives. Common Testing Objectives Desktop Bench Pilot Demonstration Evaluate the suitability of biofiltration to meet treatment ◑ ◑ ● ● objective(s) Optimize media type and depth ◑ ● ● Optimize backwashing strategy Type and concentration of disinfectant ● ● Surface wash rate and duration ● ◑ ● Air scour rate and duration ◑ ● Backwash water rates and durations Test hydraulic performance and filter production (e.g., ● ● headloss and flow) Evaluate contaminant removal (consider seasonality) Particulates (e.g., turbidity) ◑ ◑ ● ● Organic Contaminants ◑ ◕ ● ● ◑ ◕ ● ● Inorganic Contaminants ◑ ◕ ● ● Emerging Contaminants Test optimization strategies for improving biodegradation of ◑ ● ● organic and inorganic contaminants (e.g., improve manganese removal) Develop a mechanistic understanding of biological ● ◑ ◑ contaminant removal or formation (e.g., nitrosamine precursors) Notes: ● – Addresses objective; ◕ ‐ Mostly addresses objective; ◑ ‐ Somewhat addresses objective; – Does not address objective 7.4 Designing a Desktop Evaluation A desktop evaluation is a good place to start when developing a business case for conducting biofiltration testing and/or to identify potential design or optimization strategies to test. The two main options for a desktop evaluation include a literature review or modeling biomass development and/or biodegradation of contaminants. A key conclusion should be whether additional biofiltration testing is recommended. 7.4.1 Literature Review Literature reviews are a great way to capture the status of the industry and benchmark design and water quality criteria for the facility being evaluated to other full‐scale biofiltration facilities or studies. A literature review may include peer‐reviewed and grey literature. The review may capture potential contaminant removal by biofiltration, optimization strategies and water quality data available on the water source(s) to be treated. Some example sources include: • Peer‐reviewed scientific journals. • American Water Works Association journals, handbooks, and manuals of practice. • The WRF reports. • Conference proceedings. • Thesis or dissertation reports.

Biofiltration Guidance Manual for Drinking Water Facilities 83

• Published data on specific water supplies (current or planned) found through regional metropolitan development centers or state development boards, geologic survey, etc. • Data from other full‐scale biofiltration facilities (e.g., neighboring utilities, the Biofiltration Knowledge Base, or Appendices J and K). 7.4.2 Modeling Another option for forecasting the potential for biofiltration to be effective for a specific WTP is to model potential biomass development and/or biodegradation of contaminants. Modeling can be completed in as little as a couple months, but generally considers multiple years of historical water quality and operational data. However, key outcomes of any modeling effort should be confirmed with bench‐, pilot‐ and/or demonstration‐scale testing. A few options for modeling include: • BIOFILT: The BIOFILT Model was developed by Hozalski and Bouwer (2001). This model simulates removal of biodegradable organic matter under varying operational conditions. The BIOFILT Model also simulates the initial acclimation process and the effects of a sudden loss in biomass. • BIOFILM MODELS: The Transient‐State, Multiple‐Species Biofilm Model (TSMSBM) developed by Rittmann et al. (2002a), the Integrated Biofilm Model (IBM) developed by Rittmann et al. (2002b), and the index of contact time, X*, developed by Huck ad Sozanski (2008) simulate biofilm changes and analyze various biofilm factors that impact biofiltration performance. • LITERATURE VALUES: Apply percent contaminant removals reported in literature or by peer full‐ scale biofilter facilities to biofilter influent water quality data for the facility under investigation. The Conversion Assessment Tool (Upadhyaya et al. 2017, WRF 4496, Appendix F) summarizes contaminant removal ranges for common biofiltration goals including removal of organic carbon, disinfection byproduct precursors, metals, and T&O compounds. Dickenson et al. (2018, WRF 4559) summarized trace organic contaminant removal. • OTHER MODELS: Additional quantitative biological tools are available in Environmental Biotechnology (Rittmann and McCarty 2001). Other non‐commercially available models are under development, but not covered here. 7.5 Designing Bench, Pilot, or Demonstration Tests Detailed biofiltration testing plans result in higher quality data and more efficient studies. Testing plan samples are provided in Appendix L. The sections below summarize best practices for each critical part of a well‐designed testing plan. 7.5.1 Duration Typical durations for each scale of testing are summarized in Table 7‐4. The duration should be selected to allow sufficient time to meet all project objectives. Testing should be scheduled to allow multiple testing scales to inform full‐scale implementation. Key considerations for selecting an appropriate duration are provided below. Table 7‐4. Comparison of Typical Testing Durations for Different Testing Scales. Bench – Research Bench – Treatability Pilot Demonstration 3 to 6 months 6 to 12 months 6 to 18 months* * Duration should capture relevant seasonal variations. In some cases, goals may be met with a shorter duration.

• Acclimation – Acclimation is the time required to reach steady‐state operation based on the objective of the study and parameter in question. This may include achieving stable biomass concentrations (e.g., ATP) and steady‐state contaminant removal performance (e.g., organics and

84 The Water Research Foundation

inorganics removal). Acclimation may require weeks or months but could be shortened by using spent GAC and/or previously acclimated media. • Baseline – Before beginning testing of challenge conditions or alternative operating scenarios, develop a performance baseline under anticipated typical operating conditions. • Seasonality – Consider seasonal changes in source water (e.g., algal blooms, lake turnover, or runoff) that may impact the filter influent water quality. Include testing that represents range of water quality anticipated full‐scale during at least one testing scale prior to implementation of biofiltration or biofiltration modifications. • Performance Demonstration – Allow time to assess performance under steady‐state conditions after variables (i.e., controlled variables, such as flow, or uncontrolled variables, such as source water changes or seasonal events) are modified. Steady‐state biofilter performance can be assessed by study‐specific objectives but typically includes the following variables: filter effluent turbidity, TOC removal, target contaminant removal, and run time headloss or headloss accumulation rate. Temporal changes of less than 20% are a typical range for confirming steady‐state has been achieved (Lauderdale et al. 2018). However, this rule of thumb may vary depending on study‐ specific objectives.

7.5.2 Design Potential design criteria for biofiltration are discussed in Section 5.2. These full‐scale criteria should serve as the basis for designing testing systems but will be scaled‐down or modified as described in Table 7‐5 through 7‐7 to allow for testing of different variables at a feasible scale. Testing design should also consider conventional filtration design criteria, such as depth over ES (i.e., L/d), which may explain water quality changes and differences between studies.

Table 7‐5. Comparison of Filter Design for Different Testing Scales. Design Factor Bench – Research Bench ‐ Treatability Pilot Demonstration  Synthetic water with a  Site‐specific water  Site‐specific water specific NOM fed from a  Fed continuously directly from the source (typically composition (could be a reservoir, refilled settled water) NOM extract) in batches (i.e.,  Site‐specific water fed collected from a Water from a reservoir, refilled single time point Source in batches (i.e., collected representative of a from a single time point season) or fed representative of a continuously season) or fed continuously  Typically, in the mL/min range  Typically, in the gpm (or L/min) range  Targeted to match anticipated full‐scale EBCTs  Targeted to match anticipated full‐scale loading rates Flow per per unit area (e.g., gpm/ft2) and EBCTs filter  Allows for testing hydraulic conditions (e.g., headloss accumulation rate) and particle removal Number of  3 to 12, ideally in a factorial design (refer to  Typically, 4 to 6 with  Typically, 1 to 3 (could Filters Table 7‐14; each variable is tested independently) one control include a control)  Filter column diameter to media diameter ratio  Column inner  Use of an existing full‐ Filter should be > 15:1; Column inner diameter typically diameter typically in scale filter or ~50 sf if a Diameter ≤ 2 inches (i.e., in the 1 to 5 cm range). the 4 to 8 inches (10 side stream filter to 20 cm) range  Column height typically >  Column height  Column height  Use of an existing full‐ 5 cm typically > 50 cm typically up to 12 ft scale filter or side stream Filter Height range (3.7 m) filter representing full‐ scale filter height

Biofiltration Guidance Manual for Drinking Water Facilities 85

Media selection (Table 7‐6) is a critical component of testing design and should match specifications for full‐scale implementation as closely as possible. If an alternative media type or size is used (e.g., for bench testing using small column sizes), the testing plan should address how results will be interpreted. Virgin media may be acquired from a vendor or acclimated and/or, for bench or pilot testing, spent media may be obtained from full‐scale biofilters (i.e., either from the existing full‐scale biofilters or from another biofiltration facility). Considerations for each are described below.

Considerations for acclimated/spent media Considerations for virgin media: (for bench or pilot testing):  Proper washing to remove excess fine media  Adsorption is less likely to be a significant particles and chemical residuals. removal mechanism for organic carbon.  Extended acclimation time.  If acclimated to a different water quality,  Adsorption of contaminants if GAC is used – re‐acclimation to site‐specific water quality. GAC will retain adsorption capacity for  Implications of testing media with several months to years and elucidating the accumulated biofilm and precipitants (e.g., removal mechanism, from physical precipitated metals or scaling) accumulated adsorption or biological degradation, may be over time from another water source. difficult.

Table 7‐6. Comparison of Media Design for Different Testing Scales. Design Bench – Research Bench – Treatability Pilot Demonstration Factor Granular filter media Uses granular filter Uses granular filter media; often includes or glass beads; often media; may include dual‐media profiles to match full‐scale Filter Media includes only one dual‐media profiles to anticipate design Type media type per column match full‐scale anticipate design ≤ 1.5 mm; selected considering column diameter Uses granular filter media matching full‐ Filter Media (Filter column diameter to media diameter ratio scale design (e.g., 1 to 1.5 mm) Size should be > 15:1; may match anticipated or typical full‐scale effective size Filter Media Typically, virgin media May use virgin media, or may use media collected from the full‐scale Source from a vendor WTP under consideration or another biofiltration WTP Depth typically selected based upon the column Media depths match anticipated full‐scale Filter Media diameter and target EBCT (McKie et al. 2019) design (e.g., 24 to 72 inches) and are Depth selected considering EBCT and L/d10 Biomass on media must be acclimated; testing may allow for a period Testing should allow of acclimation to understand performance dynamics during startup or for a period of media may use media from another WTP to decrease acclimation time acclimation and to Filter Media understand Acclimation performance dynamics during startup

86 The Water Research Foundation

One of the defining differences between each testing scale is the ability to test backwashing strategies (Table 7‐7). Bench testing typically only includes backwashing as needed to continue to meet the target flow; backwashing hydraulic performance is not representative of full‐scale. Pilot testing allows for backwashing, may be automated,d an may allow for testing the impact of different chemicals (e.g., chlorine, ammonia, coagulants) during backwashing, but is often limited to a standard backwashing system with limited potential to scale‐up to full‐scale implementation. Thus, to fully test different backwashing systems, a demonstration test must be conducted.

Table 7‐7. Comparison of Backwash Design for Different Testing Scales. Design Factor Bench ‐ Research Bench – Treatability Pilot Demonstration  Adjust to achieve 20‐50% bed expansion and  Match full‐scale rates per sf spent filter backwash is visibly clear  Adjust to achieve 20‐30% bed expansion  May be in the range of 0.1‐0.4 gpm for 2 to and spent filter backwash turbidities Backwash 3 minutes (~1.5 to 2 bed volumes) below 10 NTU Rates/Durations  Short experiments (e.g., less than 6 weeks)  May be in the range of 5‐30 gpm/sf for 10 within the overall testing plan may be to 20 minutes conducted without backwash  Synthetic or tap Filtered water from the Filtered water from the test filters or full‐scale water bench or full‐scale backwash source Backwash Water  May be operated Source without backwash for shorter duration tests May use lab bottles (e.g., 2 L) or a tank (e.g., 75 L) 150 gal, typically Supplied from the to collect filtered water from the bench columns filled with filtered full‐scale backwash Backwash Tank water source or by Volume dechlorinating finished water Typically, none, but manual media breakup may be Air scour (3 to Surface wash or air Agitation required (e.g., shaking or stirring with a metal rod) 5 scfm/sf) scour (3 to 5 scfm/sf) Typically, none  May include a filter aid (e.g., coagulant or polymer) or disinfectant (e.g., chlorine, Backwash Chemical chloramines or peroxide) Feed  Typically, automated based upon backwashing set points Backwashing Headloss, including water height if gravity fed, run Terminal run time, effluent turbidity and Triggers time, and/or percent effluent valve open headloss matching full‐scale None (typically, all flow is sent to waste)  Match full‐scale and/or set based upon a turbidity limit  Generally, turbidity data recorded during Filter‐to‐Waste filter‐to‐waste to allow for analysis of ripening profiles  If all flow is sent to waste, used to control quality of backwash water

Biofiltration Guidance Manual for Drinking Water Facilities 87

7.5.3 Testing Conditions After determining the overall test approach, the experimental plan (i.e., testing conditions) must be established. Testing conditions should be selected based upon water quality or filter hydraulic goals. These conditions will inform the overall testing duration and may require modifications to the system design (e.g., the addition of more chemical feed pumps). Table 7‐8 summarizes common testing conditions for each testing scale.

Table 7‐8. Comparison of Testing Conditions for Different Testing Scales. Conditions Bench – Research Bench ‐ Treatability Pilot Demonstration  Typically held constant  Typically includes testing of design flow,  For some studies, the flow may be reduced  May consider testing at other flows (e.g., Flow or increased to evaluate different EBCTs typical or seasonal full‐scale flows)  Flow spiking and reduced flow during filter‐ to‐waste may be conducted  Uses lab  Uses ambient  Typically uses  Typically uses temperature or temperature or ambient ambient temperature‐ temperature‐ temperature (i.e., temperature (i.e., adjusted water adjusted water capturing a range capturing a range (e.g., to test (e.g., 5th and 9th by testing during by testing during performance at percentile of multiple seasons) multiple seasons) Water specific historical  May use Temperature temperature conditions to temperature‐ ranges relevant to bracket adjusted water for biofilter facilities) temperature special testing  May be located in ranges) (e.g., to match a temperature‐  May be located in historical minimum controlled room a temperature‐ temperatures) controlled room Target contaminant Target contaminant concentration is typically ambient (e.g., organic concentration spiked to compounds), but contaminants may be spiked or pulsed to a desired Target desired concentration concentration (e.g., historical maximums) for a relatively short duration (e.g., Contaminant(s) determined based up to several weeks) to simulate a short periodic water quality excursion upon research goals (e.g., T&O or manganese) or if testing cannot be conducted during the season when historical maximum concentrations are observed  Typically, no filter aid used  Typically, only added if anticipated for full‐ Filter Aid  Filter aids may lead to faster clogging of scale application filters and turbidity is generally not scalable  May include performance testing of different filter aids In addition to filter aids or target contaminants, filter influent chemical feed may include (refer to Chapter 4 for additional discussion of optimization strategies): Filter Influent  Oxidant addition (e.g., peroxide or chlorine for headloss management) Chemical  Nutrient addition (e.g., for management of extracellular polymeric substances) Enhancement  pH adjustment (e.g., for management of manganese)  Other chemicals, based upon testing goals Chemicals may be added for the duration of the study or for short durations (e.g., two weeks).  May test settled water  Commonly tested in combination with  May include bench‐scale upstream upstream ozonation treatment processes, such as ozonation,  For testing of new treatment plants, Pre‐Treatment coagulation and sedimentation upstream treatment may replicate the full upstream treatment processes  May also be tested without an additional upstream treatment (i.e., using applied water from the full‐scale treatment plant)

88 The Water Research Foundation

For each testing condition identified, limits should be established to challenge the biofilters based upon historical conditions and/or to meet project goals. For testing contaminant removal performance, it is valuable to test both ambient conditions and challenge the biofilters beyond historical conditions, to anticipate potential performance should water quality become more challenging in the future. In the absence of historical data for a contaminant of interest, testing should ensure multiple seasons and anticipated operational conditions are captured. Example challenge testing strategies are summarized in Table 7‐9.

Table 7‐9. Challenge Testing Strategies Potential Applications – Achieved by Testing Over Challenge Testing Strategies Multiple Seasons or by Manipulating the Influent Ambient conditions  General water quality (e.g., hardness and alkalinity) (i.e., no imposed challenge testing)  Organic compounds  Turbidity  Dissolved oxygen (may have an established minimum acceptable concentration)  Nutrients  ORP Minimum design, average historical and maximum  Flow design conditions  EBCT  For site‐specific testing, based upon full‐scale  Filter aid doses treatment plant or anticipated design  For research, based upon industry typical designs Given parameter at 5th and 95th percentiles based  Seasonal source water changes / blends (e.g., algal upon historical data and cyanobacteria blooms)  Organic compounds  Turbidity  Temperature  pH  Filter aid doses Given parameter is 150 or 200% higher than If a specific goal for testing: historical conditions  Metals  T&O compounds  Nutrients (e.g., ammonia or orthophosphate)  Emerging contaminants (e.g., PPCPs) Range established based upon literature and prior  Chemical enhancement doses studies  Pre‐treatment conditions  Contaminants where historical data is unavailable Operational stressors  Chemical feed failures  Preoxidant overdosing  Biofilter shutdowns

Biofiltration Guidance Manual for Drinking Water Facilities 89

7.5.4 Equipment and Instrumentation Select equipment and instrumentation to match the testing goals. Some key considerations for selecting equipment are summarized below. Review the testing design and conditions when selecting equipment.

 Equipment location  Power requirements - Protection from weather and sunlight - Total power needed for all - Distance from water and/or equipment/instrumentation backwash source - Location of power outlets relative to - Availability and location of sufficiently equipment location sized drains for overflow or spill - Backup power / procedures for - Operator comfort power outages - Access large enough for moving - Power surge protection equipment to/from  Chemical feed equipment - Clearance on all sides for sample - Pump types and flow rates collection, maintenance and - Tubing materials and sizes operation - Chemical compatibility - Mitigation of any vibrations or - Chemical feed points disturbances that might interfere - Number of chemicals needed with instrumentation or columns - Chemical feed mixing requirements - Personal protective equipment (PPE)  Discharge plan / environmental  Online monitoring instruments compliance - Water quality/operational - Hazardous waste disposal for parameters chemicals - Monitoring locations - Discharge permits - Data collection frequency - Chemical storage - Cleaning, calibration and detection - Double containment for chemicals limits - Spill response equipment - Data management and backup - Eye wash or shower station - Power supply

90 The Water Research Foundation

Table 7‐10 summarizes typical levels of automation for each testing scale. In general, higher levels of automation improve the quality of data obtained, decrease labor time for operators, and increase response time by allowing adjustments to testing conditions based upon real‐time data. However, increasing levels of automation also require larger project budgets, and may not be practical for bench‐ scale testing involving many columns and low flows.

Table 7‐10. Comparison of Monitoring and Automation for Different Testing Scales. Instrumentation Bench ‐ Research Bench ‐ Treatability Pilot Demonstration  Limited to no‐automation  High‐level of automation Overall  In‐line/online instrumentation is limited  In‐line/online instrumentation utilized  Influent, backwash, and chemical pumps set  Flow control may be manual or through manually programming in a human machine  Flow may be controlled by changing the interface (HMI) percent open of the effluent valves  Online analyzers may monitor and Automated  Headloss (i.e., filter level) typically monitored datalog flow, filter level, headloss, run Operational Control manually if gravity fed, but could include time, etc. pressure sensors  Chemical pump setpoints may be changed  Chemical pump rates set manually manually or through the HMI based upon a concentration setpoint  Typically, none  Individual filter effluent turbidity  Water quality analyzed through grab samples  Filter influent temperature  May include additional online analyzers for key parameters (e.g., pH) Online Water   Quality Analyzers Typically, grab Utilize existing sample full‐scale measurements instrumentation for filter influent if testing full‐ turbidity scale biofilters Recorded by hand  Data logging by  Data logging by a a manufactured manufactured system system or Data Acquisition through full‐ scale SCADA  Grab sample measurements recorded by hand

Biofiltration Guidance Manual for Drinking Water Facilities 91

Once a list of all equipment and instrumentation required is developed, the final question is how the equipment and instrumentation will be procured. Three main options include building new equipment either on‐site or off‐site (e.g., at a university), leasing equipment from a specialty manufacturer or university, or purchasing equipment from a specialty manufacturer. Table 7‐11 summarizes key considerations for each procurement options based upon testing scale.

Table 7‐11. Testing Equipment Procurement Considerations. Instrumentation Bench ‐ Research Bench – Treatability Pilot Demonstration  Effective when constructed by a  Labor costs for construction and troubleshooting of University or specialist equipment issues generally exceed costs for leasing / Build  May be constructed off‐site and shipped buying specialized equipment  Online automation will be limited  Not applicable  Available through  Typically, most economical if a year or less of testing  Equipment Universities or  Available customizations may be limited used in specialized  May include equipment maintenance/replacement research consultants (i.e., a costs (i.e., minimizing unexpected costs) studies is contract for  May include on‐call support (i.e., minimizing generally built construction of shutdowns) Lease/Rent or already equipment)  If leased through a company specializing in owned by the  Equipment is equipment, generally provides equipment with high researcher most often reliability (i.e., minimizing labor required for returned after the equipment troubleshooting and increasing quality of study concludes data) (i.e., similar to a  Availability may be limited during high demand lease) as  Becomes more economical if multiple years of equipment is testing labor intensive to  May be required if customized equipment is needed continually  Allows for continued long‐term testing (e.g., for operate optimization of biofiltration after full‐scale Buy implementation)  Requires utility staff training and availability to operate or maintenance of contract services  Equipment may require maintenance over time  Generally, requires at least 16 weeks for design/construction

7.5.5 Basic Monitoring Monitoring plans for testing will be very site‐specific and should be developed based upon testing goals. Chapter 2 describes various parameters and methods that could be considered. Table 7‐12 describes monitoring recommended for biofiltration testing. However, monitoring frequencies of some or all of these parameters may vary during specific testing phases, and additional parameters may need to be included depending on the testing goals. Additionally, monitoring should include relevant upstream and downstream water quality (not included in Table 7‐12). The best approach to developing and modifying a monitoring plan is to utilize industry experts to balance the value obtained against the resources required to conduct additional monitoring. One of the benefits of testing is the ability to collect samples at more locations than may be available full‐scale. Additionally, it may be easier to collect media samples and control the depth of media sample collection. Thus, additional monitoring that may be advantageous during testing, depending on testing goals and preliminary results, includes:  Contaminant removal over filter runs (e.g., every 6, 12, or 24 hours).  Headloss over filter depth (e.g., with a manometer or pressure gauge).

92 The Water Research Foundation

 Contaminant removal over filter depth, which can be used to rapidly evaluate the impact of EBCT on contaminant removal.  Zeta potential if a cationic filter aid polymer is used.  Online microbial analyses, such as dissolved oxygen uptake or pH reduction.  Advanced microbial analyses conducted by partner universities, including microbial community sequencing, qPCR, and EPS.

Table 7‐12. Recommended Basic Monitoring Plan for Biofiltration Testing. Recommended Sample Frequency Category Critical Parameters Location Online Daily Weekly Monthly Loading rate  Operations Headloss  Filters Run times  Temperature ** Filter influent pH  Alkalinity  General Water Quality Hardness, Total and  Calcium Disinfectant Residual (if  applicable) Turbidity  TOC and DOC  UV  Organic Carbon 254 AOC  Filter Influent and effluent Carboxylic Acids  Nitrite  Nitrate  Nutrients Ammonia  Orthophosphate  Taste and Odor (T&O) MIB  compounds Geosmin  Inorganic Compounds Iron x2 (total and dissolved) Manganese x2 TTHMs * Distribution system or Disinfection Byproducts HAAs * formation potential tests Biofilm formation rate x2 Filter influent and effluent In situ ATP x2 Filter influent and effluent Microbial DO Consumption (non‐   Filter influent and effluent zone facilities only) Notes ‐ * = Monthly to quarterly (as applicable); ** = To confirm online measurements.

7.5.6 Data Management and Interpretation The data management best practices discussed in Section 2.6 are also recommended during testing. At minimum, temporal trends in water quality and performance data should be evaluated, per season and per test condition. There are several data analysis tools available (Figure 7‐3). Common methods are used for data interpretation, including, but not limited to: • General statistical analysis (e.g., 5th percentile, 95th percentile, and average). • Time‐series graphs (e.g., temperature over time). • Bar graphs showing results for each test condition as the average with error bars as the standard deviation. • Box and whisker plots (e.g., to compare the same parameter over different testing conditions). • Scatter plots of one variable as a function of another (e.g., organic carbon versus temperature).

Biofiltration Guidance Manual for Drinking Water Facilities 93

When interpreting data, it is critical to isolate changing variables (e.g., both uncontrollable water quality changes and intended testing changes). Although basic analytical tools are sufficient for many studies, when determining cause and effect relationships, it may be advantageous to consider more robust statistical analyses, such as multivariate regression or principal coordinate analysis. 7.5.7 Staffing Options Insufficient staffing of testing systems is one of the greatest challenges to biofiltration testing. Operating bench, pilot and demonstrating testing systems often requires 20 to 40 hours per week or more and may require additional hours with minimal notice if equipment requires troubleshooting or maintenance. Operation without dedicating appropriate resources may result in significant schedule delays and unreliable data. Thus, it is valuable to consider partnerships that allow for utility involvement, expert review and input, possible university involvement and dedicated operators (i.e., at minimum one regular and one backup operator) on‐site for the duration of testing. Table 7‐13 highlights considerations that should be evaluated in determining the best team to staff a biofiltration testing program. Table 7‐13. Comparison of Options for Staffing. Operator Options Considerations Utility Staff  Provide training to operators prior to full‐scale operation  May be difficult to schedule around existing priorities  May be more expensive than university students or consultants depending on the grade level of staff utilized and overtime conditions  Potential to get operator licenses using testing experience University  May be most cost‐effective if a university is nearby with relevant experience Students  University students have flexible working hours, but may have limits to total hours worked  Could be operated on‐site or in a university lab Junior to Mid‐  Typically, well‐trained with dedicated time to pilot operations Level  In frequent communication with experts Consultants  May have relevant historical knowledge of the WTP  May be more expensive hourly, but could be cost competitive due to operational efficiencies Experts  May include senior consultants, professors, or utility staff with significant experience in pilot testing and/or the contaminants of interest  Most expensive per hour  Expertise reviewing plans and data may save significant resources  Recommended to be utilized in combination with other operators at key decision points duringd testing an to provide quality reviews throughout testing

94 The Water Research Foundation

7.5.8 Quality Control Quality control is critical to maximizing the value obtained from testing. Further, inaccurate data could lead to invalid conclusions that are not repeatable full‐scale and may increase design and operation costs. Table 7‐14 summarizes some common quality control measures implemented during testing. Testing plans should include a section describingw ho quality data will be ensured throughout testing.

Table 7‐14. Common Quality Control Measures. Quality Control Measures Description Control Column A control column matching the full‐scale media profile and chemical addition. Operational conditions may match full‐scale conditions during select baseline testing phase(s) to evaluate the scale‐up potential of the pilot results; however, during most testing phases the control column will be operated under the same challenge conditions as the other columns to allow comparison of media profiles or strategies. Factorial Design A factorial design (Collins et al. 2014, Lavrakas 2008) ensures that only one variable is different between each set of columns (i.e., that each variable is tested independently) Baseline Testing Experimental plans often include different phases of testing (e.g., for spiking contaminants or operating at different flow rates). The baseline testing phase should set a performance expectation under typical operating procedures without any specific amendments or challenge conditions. This phase should also confirm that steady‐state performance has been achieved as defined in 7.5.1 (i.e., the biofilters are acclimated). Replicate Samples Replicating 5‐10% of samples ensures that measurements are precise and repeatable. Calibration Calibration of all instruments per manufacturer recommendations ensures instruments are reading accurately. Regularly check that standards used in calibration have not expired. Standards, Sample Spiking and Depending on the method used, the accuracy of measurements should also Blanks be verified by measuring a blank sample, spiking a sample with a known concentration, or measuring the value of standard solutions of known concentrations. Instrument Maintenance All instruments and equipment must be maintained per manufacturer recommendations. Creating a calendar for maintenance work is a best practice. Data Management A data management plan should be developed to ensure data is stored securely and in an accessible way that allows for continual review while preventing/correcting any data entry errors or unintended manipulation of data (e.g., accidental deletion). Routine Expert Review Frequent (e.g., weekly) review of operating conditions and data by industry experts allows for modifications to testing plans to maintain focus on meeting testing goals. Routine Utility Review Frequent (e.g., every other week) review of operating conditions and data by utility staff allows for identification of atypical data and/or conditions that may not be practical.

Biofiltration Guidance Manual for Drinking Water Facilities 95

7.5.9 Safety Testing is often conducted by staff that do not have prior experience with the testing equipment and may be located in an area not designed for the specific testing to be undertaken. Thus, developing a detailed safety plan and providing safety training to all operators of testing equipment is critical. Table 7‐15 summarizes a list of common safety considerations. This is not an exhaustive list of all potential safety hazards but is provided as a tool to begin safety hazard assessments.

Table 7‐15. Common Safety Considerations. Common Safety Measures Description Personal  Lab coats, long sleeves and long pants Protective  Close‐toed shoes (consider steel‐toed boots, chemical resistance, slip resistance, etc.) Equipment  Gloves (selected based upon task and hazards)  Safety glasses or goggles Operator Safety  Prevention of awkward twisting or repetitive movements  Proper lifting techniques  Frequent breaks  Water availability  Weather conditions (if not located in a room controlled with HVAC) Chemical  Double containment should be provided; review chemical compatibility for any chemicals Storage stored in the same area  All chemical tanks should be labeled legibly with chemical names and concentrations  Material data sheets and safety data sheets should be nearby and reviewed by all operators prior to beginning testing  Spill kits  Hazardous chemical waste Housekeeping  Glass disposal  Hazardous chemical disposal  Mats to prevent slipping in areas that may get wet (e.g., around testing equipment)  Prevention of tripping hazards (e.g., ensure pathways are clear of cords, pipelines, buckets, etc. for walking)  Secure use of ladders with three points of contact  Absorbent pads for benchtops General Safety  First aid kits Supplies  Eye shower / eye wash stations or kits  Fire extinguishers  Air quality monitoring

96 The Water Research Foundation

7.6 Overcoming Common Testing Challenges Investing in high quality biofiltration testing often reduces costs during design, start‐up and operation and ensures that performance goals are met. However, this value may be diminished if shortcuts are implemented to reduce the costs, schedule or resources dedicated to testing. Common challenges include: During Testing: During Implementation:  Unclear conclusions due to: • Optimal design and operational conditions tested - Inconclusive data (e.g., too many cannot be implemented full‐scale and/or are cost‐ variables changing at the same time) prohibitive - Inaccurate data • Full‐scale design does not meet performance Insufficient data collection frequency - objectives (e.g., if full‐scale conditions are (i.e., no statistical significance in trends) encountered that were not tested) - A missing baseline to allow • Untrained full‐scale operators that did not receive comparison hands‐on training on biofilter operation and site‐  Increased costs specific performance expectations during testing  Schedule delays • Increased costs for optimization following  Missed critical testing periods, including implementation acclimation/start‐up dynamics and historically most challenging water quality  Safety hazards leading to onsite injuries, chemical spills, etc.  Additional testing required to meet goals Table 7‐16 summarizes best practices for overcoming these challenges and increasing the value obtained from testing. Refer to Appendix F for further operational guidance.

Biofiltration Guidance Manual for Drinking Water Facilities 97

Table 7‐16. Overcoming Common Biofiltration Testing Challenges. Testing Component Overcoming Common Challenges Defining Testing  Define specific testing goals prior to selecting a testing scale (Figure 7‐2) Goals Benchmarking  Conduct benchmarking prior to testing design (Section 7‐2) Testing Scale  Conduct multiple scales of testing selected based upon testing goals (Tables 7‐2 and 7‐3) Selection Duration  Do not shorten the study duration (determined based upon goals) to reduce costs (Section 7.5.1)  Allow for acclimation of biomass on filter media  Collect a performance baseline before beginning to modify conditions  Allow for testing all seasons / operating conditions  Include time for testing with challenging water quality and operational conditions Design  Differentiate between adsorption and biodegradation (if GAC is used) (Section 7.5.2)  Consider full‐scale design limitations (e.g., space available for media)  Collect operator feedback Testing Conditions  Test conditions across multiple seasons (Section 7.5.3)  Test conditions independently Challenge Testing  Exceed conditions anticipated full‐scale (e.g., spiking of contaminants above historical levels) Conditions  Challenge the system operationally (e.g., testing at and exceeding the design loading rate) (Section 7.5.4) Equipment and  Use online instrumentation Instrumentation  Implement appropriate frequency of O&M for instruments (cleaning, calibrating) (Section 7.5.4)  Do not select manual equipment due to lower capital costs (the operational costs will be higher, and this will likely have schedule implications due to troubleshooting) Monitoring  Capture start‐up dynamics (Section 7.5.5)  Monitor changes in the biological community  Collect frequent enough samples to meet project goals confidently  Collect sufficient sample replicates to assess precision Data Analysis  Turn statistical analysis of trends into a performance baseline for operators (Section 7.5.6)  Consider limitations of the testing scale when analyzing results (Table 7‐2) Operator Selection  Take advantage of operational experience / training opportunities (Section 7.5.7)  Bring in the appropriate level of expertise  Sufficiently staff testing systems to meet project goals even if it increases costs Quality Control  Include a control matching the full‐scale plant (Section 7.5.8)  Change one variable at a time (i.e., independent variables)  Allow for the system to reach steady‐state as defined in 7.5.1. before changing another variable  Check standards, calibrations, and expiration dates and clean analytical equipment  For full‐scale demonstration studies, existing SOPs should be updated as needed to reduce confounding variables; SOPs should be followed consistently. Safety  Develop a health and safety plan (Section 7.5.9)  Enforce health and safety procedures

7.7 Understanding Expected Outcomes A final piece of a well‐designed testing plan is an understanding of what conclusions can be expected and how next steps following testing will be determined. Table 7‐17 summarizes commonly expected outcomes from each testing scale. At the end of testing, actual versus expected outcomes should be evaluated to determine whether additional testing is warranted at the current scale, or whether testing should proceed at a different scale prior to full‐scale implementation. Figure 7‐5 subsequently summarizes key considerations for moving to larger or smaller testing scales.

98 The Water Research Foundation

Table 7‐17. Expected Outcomes of Each Testing Scale.

Expected Outcomes Desktop Bench ‐ Research Bench ‐ Treatability Pilot Demonstration Feasibility of biofiltration for Site‐Specific biodegradation of Estimate General Indication Site‐Specific Indication Confirmation Prediction organic or inorganic contaminants Potential for Site‐Specific optimization strategies Identification General Indication Site‐Specific Indication Confirmation Prediction to meet project goals Potential formation of byproducts (i.e., Site‐Specific Estimate General Indication Site‐Specific Indication Confirmation regulated, nonregulated Prediction and microbial) Expected full‐scale Site‐Specific contaminant removal, No Indication No Indication No Indication Confirmation Prediction including turbidity Expected full‐scale Site‐Specific hydraulic performance No Indication No Indication No Indication Confirmation Prediction and production Impact of seasonal Site‐Specific water quality and No Indication No Indication General Indication Confirmation Prediction upstream treatment Optimization of No Indication No Indication No Indication Partial Indication Full Indication and Confirmation backwashing protocols Design criteria Identification No Indication Partial Indication Selection Confirmation

Final Outcome  Selection of  General indication of  Selection of 3‐5 strategies and  Selection of 1‐3  Selection of optimal strategies and/or potential biofiltration design criteria for pilot testing strategies and strategy/design for full‐ design criteria that performance under testing  Recommendation to test design criteria for scale implementation may meet project conditions successful strategies at pilot demonstration goals  Recommendation for site‐ and/or demonstration‐scale testing  Recommendation specific testing if results to be  Recommendation to conduct either applied to a specific to test successful bench‐ or pilot‐ treatment plant strategies at scale testing demonstration‐ depending on the scale number of strategies selected

Biofiltration Guidance Manual for Drinking Water Facilities 99

Figure 7‐5. Next Steps After Evaluating Testing Outcomes.

100 The Water Research Foundation

Figure 7‐5. (Continued)

Biofiltration Guidance Manual for Drinking Water Facilities 101

7.8 Resource Planning Not appropriately budgeting to meet the project objectives has historically been a major obstacle to testing. Budgets should be designed to meet the testing goals, defined testing conditions, and estimated schedule. Table 7‐18 provides a checklist for the many different factors that must be considered in developing a budget for testing. However, this table is not a comprehensive list of all potential cost factors, and as all cost factors may not be predictable during the planning phase, a contingency should be included to address unexpected events (e.g., water quality changes or equipment malfunction) that are common during testing. Each phase of testing has a progressive level of investment required to produce quality results that can be used to make decisions. While desktop evaluations are overall the least expensive, the information gleaned is limited and additional testing is recommended based upon the results. Bench testing is typically the least expensive per column from a capital cost perspective but may require significant labor resources as bench testing usually has limited automation. Pilot testing requires a significant capital investment per column if piloting facilities do not currently exist onsite. Demonstration testing could be the most expensive, if custom sidestream filter(s) are procured. However, demonstration testing could also be the least expensive if one or two strategies are to be tested and existing full‐scale biofilters can be isolated for testing. Table 7‐18. Resource Planning for Each Scale of Testing. Cost Factors Desktop Bench Pilot Demonstration Duration Duration of Testing    Design Labor for Modeling  Labor for Development of Test Design/Plan    Labor for engineering design    Installation Costs Civil, construction, piping, mechanical, electric,    I&C, permitting, utility routing Utility routing (flow/power/drains))    Equipment and Supplies Testing Equipment (Manual) Columns/filters (incl. underdrain, seals, etc.)    Ancillary equipment (valves, pumps, etc.)    Skid/frame to hold valves, columns, etc.    Tanks (influent and backwash water)    Testing Equipment (Automated) PLC / HMI / SCADA   Remote access modules / internet   Pumps (influent and backwash)    Chemical Feed Equipment (tanks, pumps, etc.)    Chemicals (for dosing and/or monitoring equipment)    Filter media    Labor for discharge/environmental compliance    Monitoring and Reporting Sampling and Analysis (Grab Sampling) Organic carbon compounds    Biological Parameters (biomass and bioactivity)    Inorganics (metals, nutrients, etc.)    General Water Quality (pH, turbidity, etc.)    Manual Operational Monitoring (headloss, run time, filter    level, flow, etc.)

102 The Water Research Foundation

Table 7‐18. (Continued) Cost Factors Desktop Bench Pilot Demonstration Online Instrumentation Water Quality (turbidity, temperature, pH, etc.)   Operational (run time, headloss, filter level, flow)   Data Analysis Labor for Analyzing Data on a Continuous Basis     Data Management Tools (e.g., PowerBI, Tableau)   Labor for Reporting     Operating Costs Labor (Operators, Consultants, Universities)     Chemicals    Power    Replacement parts    Quality Control Labor to Develop Quality Control Plan    Duplication of 10% of Samples Measured    Sample Quantities that Allow for Statistical Analysis    Control Filter    Safety Labor to Develop Safety Plan    Personal Protective Equipment (goggles, gloves, etc.)   

Biofiltration Guidance Manual for Drinking Water Facilities 103

104 The Water Research Foundation APPENDIX A

Biofiltration Terminology (Adapted from Brown et al. 2016)

Term Definition

Acclimation Period The time that passes before a process is ready to perform at its best; sometimes called lag time.

Adenosine‐5’‐tri‐ A compound found in all living cells in which energy is stored in high‐energy phosphate (ATP) phosphate bonds. Its components are the purine, adenine, ribose, and three phosphoric acid groups.

Aerobic bacteria See aerobe.

Aerobic biofilters Biofilters that rely on the growth and activity of microorganisms that use oxygen as the electron acceptor.

Aerobic condition An environmental condition in which oxygen is available. In water this means that dissolved oxygen is present.

Aerobic treatment A biological treatment technique, in which oxygen is present, used to oxidize and remove from water soluble or fine material, usually organic in nature.

Air pocket A location within a pipeline or filtering medium in which air has collected.

Air scour The practice of admitting air through the underdrain system of a filter to help complete cleaning of media during filter backwash.

Air stripping A process that removes volatile compounds from a liquid phase by passing air through the liquid.

Aliquot A representative portion of a sample, often an equally divided portion.

Alkalinity A measure of the capacity of a water to neutralize strong acid.

Anaerobic bacteria See anaerobe

Anoxic/anaerobic Reactors that rely on the growth and activity of microorganisms that use bioreactors electron acceptors other than oxygen (e.g., nitrate and perchlorate).

Anoxic treatment A treatment used to remove organic and inorganic contaminants using electron acceptors other than oxygen

Biofiltration Guidance Manual for Drinking Water Facilities 105 Term Definition

Anoxic/anaerobic An environmental condition in which oxygen is not available as a terminal condition electron acceptor.

Analyte The substance for which an analysis is performed.

Analytical method An analysis for which the description is sufficiently detailed to be set up in a laboratory.

Anionic Having a negative ionic charge.

Anionic polymer A negatively charged polymeric compound used to assist in removing particles from water.

Anthracite A particular form of coal that is used in granular media filters to remove particles from water.

Anthracite‐sand filter A granular filter in which a layer of crushed anthracite coal of a specified size is placed over a layer of sand of a specified size.

Aquifer storage and A management strategy in which excess water is treated and recharged to an recovery (ASR) aquifer system for later withdrawal and use.

Aseptic Free from living organisms causing fermentation or putrefaction; sterile.

Assimilable organic The fraction of dissolved organic carbon (DOC) that can be easily assimilated by carbon (AOC) microorganisms and converted to cell biomass. AOC is a measure of organic material available for microorganisms to grow on and also as bacterial regrowth potential in distribution systems.

Autotrophic bacteria Bacteria that obtain energy through the oxidation of inorganic compounds and (autotrophs) have the ability to fix carbon dioxide (CO2) as their sole source of carbon.

Available chlorine A measure of the amount of chlorine in chlorinated compounds, such as chlorinated lime, hypochlorite compounds, and chloramines that are used for disinfection as compared to the amount in elemental (liquid or gaseous) form.

Available pressure head The pressure, or head, available to drive water through a granular media filter.

Backwash The process of reversing flow through a (bio)filter to remove accumulated particles and biomass.

Backwash rate The flow of water used during backwash per unit of a (bio)filter area, expressed as volume per unit time per unit surface.

106 The Water Research Foundation Term Definition Backwash stage A specific action in the sequence of actions that form the backwash process. Examples of backwash stages include filter drawdown, air scour, fluidization, and filter‐to‐waste.

Backwash volume The volume of water used to backwash a filter during the filter cycle.

Bacteria Microscopic unicellular organisms having a rigid cell wall.

Bacterial regrowth The presence of a persistent population of bacteria in a basin or water distribution system.

Bacterial regrowth An assessment of the potential for bacteria present in treated drinking water to potential increase in density.

Bed The mass of granular media through which water passes during (bio)filtration.

Bed depth The height of the granular media, excluding support material, in a bed.

Bed expansion The effect produced during backwashing when filter media becomes separated and rises above its resting position.

Bioavailability A measure of the extent to which a chemical is available for systemic absorption by an organism.

Biochemical Pertaining to chemical change resulting from biological action.

Biochemistry The branch of chemistry that deals with the chemical reactions involved in the life processes of plants and animals.

Biodegradability The susceptibility of a compound to decomposition by microorganisms.

Biodegradable Subject to degradation into simpler substances by biological action.

Biodegradable organic A portion of the organic carbon in water that can be degraded by heterotrophic carbon (BDOC) microorganisms.

Biodegradable organic A portion of the organic matter in water that can be degraded by matter (BOM) microorganisms.

Biodegradation The breakdown of organic matter by microorganisms.

Biofilm A layer of microorganisms held together and found at the interface between water and a solid substrate in an aquatic environment.

Biofilm formation Effect of various parameters that influence the formation, promotion, and potential stability of biofilms and the subsequent impact on water quality in the distribution system.

Biofiltration Guidance Manual for Drinking Water Facilities 107 Term Definition

Biofilter Granular media, such as GAC, anthracite, or sand that has developed a biofilm capable of degrading organic matter or any inorganic electron donor (e.g., NH3).

Biofouling A phenomenon in which the performance of a unit process is compromised by biological growth.

Biohazard An infectious agent presenting a risk or potential risk to human health, either directly through infection or indirectly through disruption of the environment.

Bioindicator An organism that produces an observable response on exposure to a given substance.

‐ Biological denitrification The transformation of nitrate nitrogen (NO3 ‐N) to nitrogen gas (N2) by microorganisms in an oxygen‐free (anoxic) environment and in the presence of an electron donor to drive this reaction.

Biological deposits Deposits of organisms or the products of their life processes.

Biological filtration Biological filtration is an operational practice of managing and maintaining (biofiltration) biological activity within rapid rate, aerobic, granular media filters that remove organic and inorganic contaminants from water.

Biological filtration could be enhanced by an application of a pre‐oxidant prior to filtration. Further optimization, referred to as "engineered biofiltration," employs the use of nutrient chemicals such as phosphorus or nitrogen that can enhance microbial activity in filter media.

Biologically active carbon Granular activated carbon (GAC) used as a biogrowth support medium.

Biological stability A biochemical condition in which the nutrient status of treated drinking water is (biostability) such that the water will not support (or will minimally support) the growth of microorganisms.

Biomass The total weight of biological matter, including any attached extracellular polymeric materials.

Bioreactor A vessel in which a biological process takes place.

Bioregeneration A process in which exhausted adsorption sites are biologically regenerated or made available for future adsorption.

Buffer A chemical substance that stabilizes the pH value of solutions.

Cationic Having a positive ionic charge.

108 The Water Research Foundation Term Definition Collapsed pulsing A method of cleaning a filter by introducing air and subcritical fluidization backwash backwash water, causing the formation and collapse of air pockets within the bed. This condition enhances the abrasion between (bio)filter media grains to assist in particle removal.

Cometabolism Process by which microorganisms transform substrates into organic products but do not obtain energy, carbon, or nutrients from the conversion.

Dechlorination The process of removing chlorine from solution.

Decomposition The conversion of chemically unstable materials to more stable forms by chemical or biological action.

Degasification The removal of dissolved gases from water to reduce their impact on water treatment (e.g., water quality, filter operation (via air binding), pump cavitation, corrosion).

Deoxygenation The depletion of dissolved oxygen in a liquid either under natural conditions associated with biochemical oxidation of organic matter or through the addition of chemical reducing agents.

Desorption The movement of a previously adsorbed constituent into the liquid phase.

Diffuser Device of varied design that transfer gas into a liquid.

Disinfectant An agent that destroys or inactivates microorganisms.

Disinfectant decay The loss or decline of disinfectant concentration or residual over time.

Disinfectant demand The amount of disinfectant required to sustain a disinfectant residual.

Disinfectant stability The ability of a disinfectant to resist degradation.

Disinfection by‐product A chemical by‐product of the disinfection process. (DBP)

Disinfection by‐product The potential of a given water to produce DBPs when exposed to a disinfectant formation potential over a given time period. (DBPFP)

Disinfection by‐product A substance that can be converted into a disinfection by‐product during precursor disinfection.

Dissolved organic carbon The portion of total organic carbon that passes through a 0.45‐micron filter. (DOC)

Biofiltration Guidance Manual for Drinking Water Facilities 109 Term Definition Dissolved oxygen (DO) The concentration of oxygen in aqueous solution.

Dual media filter A filter containing two types of granular filtering media with different sizes and specific gravities to maintain media stratification during backwashing.

Effective size The granular media particle diameter (d10) for which 90% of a sample, by mass, has an equivalent or larger diameter.

Empty bed contact time A measure of the time when water is in contact with the granular media bed in (EBCT) a (bio)filter. EBCT is calculated by dividing the total volume of the media bed (including all pore space) by the flow rate across the bed.

Extracellular polymeric Material released outside the cell by microorganisms. These substances are substances (EPS) primarily polysaccharides and exist as highly hydrated gels and fibers surrounding, encapsulating, or connecting a consortium of microbial species in an aquatic environment (i.e., biofilm).

Filter core Vertical media sample collected from a (bio)filter.

Filter draw down The process of lowering the free water surface above filter media. Filter drawdown typically is performed prior to filter backwash.

Filter run The time interval between backwashes.

Filter‐to‐waste The practice of discharging filtered water directly to disposal immediately following a backwash. Filter‐to‐waste is typically performed for a period of 5‐10 minutes following backwash, or until the filtered water is of acceptable quality.

Filtration rate The flow of per unit of (bio)filter area, expressed as volume per unit time per unit surface area.

Fixed bed column A treatment unit containing media that remain stationary during the course of treatment.

Floc Collection of smaller particles that have agglomerated into larger particles as a result of the coagulation‐flocculation process.

Flocculant A water‐soluble organic polyelectrolyte that is used alone or in conjunction with inorganic coagulants, such as aluminum or iron salts, to agglomerate solids present in water to form large, dense, floc particles that settle rapidly.

Flow equalization The use of storage tanks to control a changing flow of water and make it nearly uniform with time.

Granular activated A form of particulate carbon manufactured with increased surface area per unit carbon (GAC) mass to enhance the adsorption of soluble contaminants.

110 The Water Research Foundation Term Definition Granular activated A layer of GAC on top of a sand/anthracite filter, used for adsorption. carbon cap

Headloss A reduction of water pressure (head) in a hydraulic or plumbing system. Head loss is a measure of 1) the resistance of a medium bed (or other water treatment system), a plumbing system, or both to the flow of the water through it, or 2) the amount of energy used by water in moving from one location to another.

Heterotrophic bacteria Bacteria that use organic carbon for cell synthesis and energy.

Hydraulic loading The amount of water applied to a give treatment process, usually expressed as volume per unit time or volume per unit time per unit surface.

L/d ratio Ratio of media depth to the effective size (d10) of the media, commonly used in establishing minimum filter media depths.

Metabolism The biotransformation of various chemicals by an organism.

Metabolite The end product of the biotransformation of various chemicals by an organism.

Microbial activity The activities of microorganisms resulting in biological, chemical or physical changes.

Microbial community Microbial populations present in a biological system.

Micronutrient A beneficial element that is needed only at trace concentrations.

Microorganism A microscopic organism, either plant or animal, invisible to the naked eye.

Mixed media A combination of two or more media products in a single loose‐media filtration bed where the products are intermixed rather than stratified in layers.

Natural Filtration The removal of particles or contaminants through in situ soil treatment.

Nitrification The oxidation of ammonia that produces nitrite and nitrate.

Nitrogen fixation The use of free nitrogen gas (N2) in the formation of nitrogen compounds during some forms of biological activity.

Rapid granular filter A filter in which particulates are removed by granular media through which water flows, typically by gravity. A rapid granular filter typically operates at design filtration rate of approximately 2 gpm per square foot or higher.

Riverbank filtration A process of collecting water in an infiltration gallery located within a bank along a river to allow the river water to pass through the soil in the riverbank.

Biofiltration Guidance Manual for Drinking Water Facilities 111 Term Definition Slow sand filter A biologically active filter characterized by a slow rate of filtration, commonly 0.015‐0.15 gallons per minute per square foot filter area.

Soluble microbial By‐products of biological activity that are released during (bio)filtration. products (SMP)

Surface wash A supplementary method of washing filter media by applying water under pressure at or near the surface of the media by means of a system of stationary or rotating jets.

Total organic carbon A measure organic carbon in water that includes both dissolved and particulate (TOC) carbon.

Turbidity A measure of the cloudiness of water or a measure of how much suspended material inhibits the passage of light through water.

Underdrain A support underneath (bio)filter media that collects treated water and distributes backwash water and air during backwashing.

Uniformity coefficient A measure of the distribution of granular media particle sizes, calculated as d60/d10

Where, d60= the grain size, in millimeters, for which 60% (by weight) of the media’s grains are finer

d10= the grain size, in millimeters, for which 10% (by weight) of the media’s grains are finer

Unit filter backwash The volume of backwash water used during a backwash per unit area filter volume media.

Unit filter run volume The volume of water filtered per run per unit area of filter media.

112 The Water Research Foundation

APPENDIX B

Frequently Asked Questions (Adapted from Brown et al. 2016)

This appendix covers biofiltration related questions and is broken down into questions related to planning, evaluation, design, and operation. Planning Phase 1. What are the drivers for biofiltration? Common drivers for biofiltration include the need to reduce disinfection byproduct formation potential for regulatory compliance, to improve distribution system biostability, to remove manganese, and/or to remove taste and odor causing compounds. 2. What contaminants can biofiltration address? Biofiltration can effectively reduce total organic carbon (TOC), manganese, iron, ammonia, MIB, geosmin, and a wide range of biodegradable micropollutants. 3. What are the benefits of using biofiltration? Biofiltration uses naturally occurring microorganisms to enhance current treatment schemes at a relatively low cost with minimal to moderate impacts on facility operations. In some cases, biofiltration may tbe a cos ‐effective alternative to more costly treatment processes such as post filter GAC adsorption or ion exchange. 4. What are the key factors to successfully implementing biofiltration? Biofiltration may not be as effective as other, more costly treatment processes (e.g., GAC) so care must be taken to ensure that the desired treatment objectives are appropriate. Further, biofiltration relies on water quality that will support the growth of microorganisms and treatment of target contaminants. Not all water sources, regardless of geographic location, will support biofiltration. Pilot testing for a full year prior to full‐scale demonstration is strongly recommended. 5. eWhat are th regulatory requirements associated with using biofiltration? There are no specific regulatory requirements in the USA or Canada related to biofiltration. The primary objective of a biofilter is filtration, for which the requirements vary from state to state. Please check with your primary agency for requirements in your area. 6. What process modifications are needed to accommodate biofiltration? In retrofit applications, most utilities stop carrying an oxidant residual through their filters. Other possible modifications may include providing a backwash supply without an oxidant, installing an air scour system, or changing filter media. More significant facility modifications could include the installation of pre‐ozonation or chemical feed systems for nutrient addition. Lastly, utilities must install or purchase appropriate monitoring equipment for observing ongoing biofilter performance. 7. What unintended consequences are typically experienced with biofiltration? Most utilities have not experienced unintended consequences; however, some have reported shortened filter runs, increased headloss, increased monitoring equipment maintenance, and increased chlorine demand.

Biofiltration Guidance Manual for Drinking Water Facilities 113

8. What does it cost to implement biofiltration? Costs are relative to the facility and treatment objectives. For new construction, the cost of implementing biofiltration need not differ significantly from the cost of implementing conventional filtration. In retrofit applications, implementation costs can range from minimal to significant depending on the planned facility modifications. 9. How do I assess long‐term operational cost for leveraging biofiltration? Operational costs beyond those associated with conventional filtration may include nutrient addition, cleaning of monitoring equipment, and backwash water usage. If your monitoring strategy changes as per Chapter 3, there may be additional monitoring requirements and associated costs. 10. What factors make biofiltration a sustainable technology? In most applications, treatment performance targets are met without the use of chemical addition. This is accomplished by harnessing the power of microorganisms indigenous to the raw water supply. This resource‐efficient process also addresses a wide range of contaminants and often converts contaminants to harmless end products instead of sequestering or concentrating them into a waste stream. Biofiltration also improves biostability of the treated water. 11. Is the application of biofiltration limited to certain geographies? No. Although water temperature is a factor in biofiltration performance, it has proven effective in full‐scale facilities across the U.S. and Canada. Evaluation Phase 12. How important is it to perform bench‐, pilot‐, and/or demonstration‐scale evaluations? While bench‐scale evaluations may have value for establishing treatability, pilot‐ and demonstration‐scale testing is a critical precursor to full‐scale implementation. These tests will support optimization of design and operational parameters. 13. What aree th costs associated with performing these evaluations? Evaluation costs are generally a function of scale, duration, equipment procurement, and analytical requirements Overall costs can range from the low thousands to hundreds of thousands of dollars. Typically, these costs are well worth the investment through savings in full‐scale design and operation. 14. What are target performance criteria for these evaluations? Typical performance criteria include maintaining filtered turbidity below 0.1 NTU, filter run time greater than 24 hours, unit filter run volume greater than 10,000 gal/ft2, and target contaminant removal (e.g., greater than 20% TOC removal). In all cases, these performance criteria should be tailored to the specific needs of the facility. 15. Are there performance data available in the knowledge base? Full‐scale operational data are summarized in the knowledge base report and in select case studies. 16. How long should I run an evaluation? It is recommended that evaluations be run long enough to achieve biological operation mode followed by a period that includes seasonal fluctuations in water quality. Pilot‐ or full‐scale evaluation durations can be as short as six months but are recommended to be at least one year. 17. How do the evaluation results translate to full‐scale? Some utilities have had difficulty in achieving full‐scale performance predicted by bench‐ or pilot‐ scale tests. Therefore, it is valuable to run a demonstration‐scale evaluation on a single full‐scale filter once process optimization has been achieved at pilot‐scale.

114 The Water Research Foundation

18. What media and media configuration should I use for the evaluation? Most utilities have dual‐media filters comprised of anthracite and sand or GAC and sand. There are various types of GAC available for evaluation. Most utilities use bituminous GAC. 19. Can I expedite the acclimation of the biofilters? Acclimation of biofilters has been reported to range from two weeks to six months. Currently, there are no reliable methods for expediting biofilter acclimation. 20. What water quality and operational data should I collect during the evaluation phase? At a minimum water quality and operational parameters should include typical parameters for filter evaluations (e.g., turbidity, headloss, temp, pH, alkalinity), target contaminants (e.g., TOC, manganese, iron, MIB, geosmin, micropollutants), nutrients (e.g., orthophosphate, ammonia) and biomass (e.g., adenosine triphosphate). Design Phase 21. Are biofiltration facilities predominantly retrofits or new construction? Based on the current Knowledge Base, over half of full‐scale facilities incorporated biofiltration as a retrofit of an existing treatment facility. 22. What are the typical requirements/standards for biofilter designs? There are no biofiltration‐specific design or monitoring requirements. See Chapters 3, 4, and 5 for recommended design considerations. 23. What type of media is most widely used in biofilters? Over half of respondents’ report using sand mono‐media or anthracite/sand dual media filters. Slightly less than half reported using GAC mono‐media or GAC/sand dual media filters. Recommendations will differ depending on the desire to leverage adsorptive capacity of GAC and variability of water quality. The performance of GAC media configurations appear to be less sensitive to changes in water quality. 24. What empty bed contact time should I use for my biofilters? EBCT is a key factor to determine duringe th evaluation phase of implementation, however common values range from 5 to 10 minutes. 25. What is the range of hydraulic loading rates associated with biofilters? Knowledge Base respondents have reported effective treatment under a wide range of loading rates. Values have been reported as low as 2.5 gpm/ft2 and as high as 10 gpm/ft2. Initial loading conditions will most likely be dictated by the primacy agency, while target loading conditions will need to be established during the evaluation phas 26. What type of water should I use for filter backwashing and what rate? Some studies have suggested that biomass in GAC‐based biofilters is relatively insensitive to a chlorine residual in the backwash supply, while biomass in non‐GAC biofilters is sensitive to chlorine. That said ‐ it is recommended that a backwash supply without a disinfectant residual be available for regular backwashing. Knowledge Base respondents have reported effective backwash rates between 11 and 22 gpm/ft2. 27. How and how often should I backwash my filters? Utilities have reported using conventional backwash protocols triggered by typical filter performance indicators (e.g., headloss, effluent turbidity). Consideration should be given to limiting or eliminating disinfectant residual in the backwash water Air scour is recommended. Reported biofilter run time ranges from 24 to greater than 72 hours.

Biofiltration Guidance Manual for Drinking Water Facilities 115

28. What type of media extraction should I use to replace old media? Frequent media change‐out is not generally required. Some utilities have reported media useful life spanning decades. Design considerations could include yard hydrants for vendor provided eductors, truck access adjacent to the filter boxes, and hose access to individual filter boxes. 29. What monitoring tools do I need and where should I locate them? Developing an effective monitoring approach is critical to successful biofilter operation. WRF 4231 provides a guide to developing a monitoring strategy based on treatment scheme and treatment objectives. 30. What are the key design provisions for biofilter underdrains? Underdrains should be designed to address the potential for headloss development due to excessive EPS production. EPS buildup in the underdrain, specifically in media retention caps, has led to underdrain failures during backwash. Utilities have various means to address this including use of steel underdrains anchored to the filter box floor, pressure relief piping that discharges to the filter box, use of oxidants (e.g., chlorine or H2O2) in the backwash water, and underdrain pressure sensors that can provide an early indication of underdrain fouling. 31. Can you design a filter for both conventional and biofiltration operation? Yes, provisions for biofiltration can be integrated into a conventional filter design. 32. What are the major cost factors for a biofilter design? The cost of designing and constructing a biofilter is essentially the same as that for a conventional filter, though testing may indicate the need for a deeper bed depth or GAC media. These factors can increase cost. Other cost factors may include additional piping for pressure relief on backwash lines or nutrient addition systems, if recommended during the evaluation phase. Operation Phase 33. What are the keys to planning a successful transition or start‐up of a biofiltration process? Utilities have reported acclimation periods of two weeks to six months. It is important to initiate biofiltration start‐up so that ample time is provided for acclimation prior to needing a functioning biofilter. 34. What water quality and performance data should I monitor during the start‐up and long‐term operation of the biofiltration process? Key parameters include turbidity, headloss, run time, unit filter run volume, temperature, nutrients, biomass (e.g., ATP), and total organic carbon removal across the filter. 35. What operational impacts can occur after converting to biofiltration? While biofilters often perform similarly to conventional filters, some utilities have reported operational impacts including shortened filter run times, monitoring equipment fouling, and increased post‐filter chlorine demand. 36. What mitigation strategies can I use to address these issues? Mitigation strategies largely focus on reducing unwanted biomass, specifically EPS, development. WRF 4215 and 4346 have shown that in some cases adding nutrients (e.g., phosphorus, nitrogen) to avoid a nutrient‐limited condition can reduce the production of EPS. Other mitigation strategies include the periodic use of an oxidant such as chlorine, hydrogen peroxide, or potassium permanganate to help address overproduction of EPS.

116 The Water Research Foundation

37. What levels of removal are achievable with long‐term biofiltration? Removal rates are specific to the facility and target contaminant, but utilities have reported TOC removal rates ranging from 10 to 20%, MIB/geosmin removal of greater than 99%, and removal of disinfection byproduct formation potential by 20 to 50%. 38.w Ho should I handle my filter‐to‐waste and backwash wastewater? Currently, utilities generally recycle filter‐to‐waste and backwash wastewater to the head of the plant prior to chemical addition. 39. What are common practices for filter shutdown, idling, and restart? Few utilities reported best practices for biofilter idling. Some reported the ability to bring biofilters online that have been idled for less than two weeks with repeated backwash and drain cycles. One utility reported using a caustic wash of the filter bed followed by a backwash prior to start‐up. For extended shutdown periods reaching months in duration, media change out and disinfection of the filter box may be required. 40. What impacts are experienced after start‐up? Start‐up dynamics are largely dependent on filter media selected. GAC biofilters leverage initial high‐ rate adsorption and will achieve excellent organics removal when brought online. Non‐GAC biofilters will require acclimation before significant contaminant removal can be expected. 41. How do biofilters perform when short‐term process upsets occur? No full‐scale operations reported impacts of short‐term process upsets. 42. How long does it take biofilters to reach acclimation/steady state? Utilities have reported acclimation periods of two weeks to six months. Treatment schemes using ozone generally acclimate more quickly than non‐ozonated systems due to the increased availability of biodegradable organic carbon. 43. Are there ways to accelerate biofilter acclimation in full‐scale operation? No full‐scale operations reported methods for accelerating acclimation. This could be an area of future research. 44. What parameters should I monitor for biological stability? Parameters such as assimilable organic carbon or biodegradable organic carbon can be used to directly quantify biological stability. Indirect measures of biological stability include disinfectant residual stability, and the occurrence of distribution system nitrification or corrosion. 45. What operations and maintenancee costs ar associated with biofiltration? Incidental biofiltration does not require additional investment to maintain beyond additional monitoring of biomass (e.g., ATP). In cases where nutrients or oxidants are used, ongoing relatively minor chemical costs will be incurred. 46. What are best practices for biofiltration training? One of the best ways to train operations staff is to engage them in the pilot testing phase of biofiltration. While training for sample collection and data analysis is important, it is also important that operators know how media in a healthy biological filter looks and what various headloss trends might indicate. Operators should also understande th basic biological processes including the role of nutrients and temperature in process performance. 47. Should I use supplemental chemicals to manage the biofiltration process? Supplemental chemicals may be helpful if your source water is nutrient limited or you are experiencing increase headloss due to EPS development.

Biofiltration Guidance Manual for Drinking Water Facilities 117

118 The Water Research Foundation APPENDIX C

Biofiltration Calculations

Media Equations 90% Passing Point . 𝑑 𝑑10 (Equation C‐1)

d90 = 90% passing point. The size at which 90% of media grains are smaller by weight d10 = 10% passing point UC = uniformity coefficient (d60/d10) d60 = 60% passing point

Uniformity Coefficient 𝑈𝐶, 𝑑 /𝑑 (Equation C‐2)

Filter Hydraulic and Design Equations Empty Bed Contact Time 𝐸𝐵𝐶𝑇 (Equation C‐3) V= volume occupied by the media, including porosity volume Q= flow rate L= media depth v= superficial flow velocity

Filter Loading Rate 𝐹𝑖𝑙𝑡𝑒𝑟 𝐿𝑜𝑎𝑑𝑖𝑛𝑔 𝑅𝑎𝑡𝑒 (Equation C‐4) Q = flow rate SA = filter surface area

Unit Filter Run Volume 𝑈𝐹𝑅𝑉 (Equation C‐5) SA = filter surface area

Filter Recovery 𝑅𝑒𝑐𝑜𝑣𝑒𝑟𝑦 ∗ 100 (Equation C‐6) UFRV = unit filter run volume (see above equation) UBWV = unit backwash volume = volume backwashed divided by surface area UFWV = unit filter‐to‐waste volume = volume wasted during filter‐to‐waste divided by surface area

Biofiltration Guidance Manual for Drinking Water Facilities 119

Headloss Accumulation Rate 𝐻𝑒𝑎𝑑𝑙𝑜𝑠𝑠 𝐴𝑐𝑐𝑢𝑚𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑅𝑎𝑡𝑒 (Equation C‐7) 𝐶𝑙𝑒𝑎𝑛 𝑏𝑒𝑑 ℎ𝑒𝑎𝑑𝑙𝑜𝑠𝑠 headloss at the end of the filter run 𝐶𝑙𝑒𝑎𝑛 𝑏𝑒𝑑 ℎ𝑒𝑎𝑑𝑙𝑜𝑠𝑠 headloss at the beginning of the filter run

L/d Ratio (Equation C‐8) Backwash Pressure 𝐵𝑎𝑐𝑘𝑤𝑎𝑠ℎ 𝑃𝑟𝑒𝑠𝑠𝑢𝑟𝑒 𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑙 𝐵𝑎𝑐𝑘𝑤𝑎𝑠ℎ 𝐼𝑛𝑙𝑒𝑡 𝑃𝑟𝑒𝑠𝑠𝑢𝑟𝑒 𝐵𝑎𝑐𝑘𝑤𝑎𝑠ℎ 𝑂𝑢𝑡𝑙𝑒𝑡 𝑃𝑟𝑒𝑠𝑠𝑢𝑟𝑒 (Equation C‐9)

Water Quality Equations Specific UV Absorbance 𝑆𝑈𝑉𝐴 𝑈𝑉 ∗ (Equation C‐10) ‐1 UV254 = absorbance at 254 nm, cm DOC = dissolved organic carbon, mg/L

Biological Equations Biofilm Formation Rate

, / 𝐵𝑖𝑜𝑓𝑖𝑙𝑚 𝐹𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛 𝑅𝑎𝑡𝑒 ∗ (Equation C‐11) ∗ ATP = adenosine triphosphate 𝑅𝐿𝑈 relative light units of the sample 𝑅𝐿𝑈 relative light units of the standard 𝑆𝐴 surface area

Adenosine Triphosphate 𝐴𝑇𝑃 𝑡𝐴𝑇𝑃 ∗ , (Equation C‐12) . , ATP= ATP normalized to the dry weight of filter media 𝑡𝐴𝑇𝑃 total ATP normalized to the wet weight of filter media 𝑚, mass of filter media sample on a wet basis 𝑚, mass of filter media sample on a dry basis

120 The Water Research Foundation

APPENDIX D

Biofilter Operations Checklist

Category Technique Parameter Location Purpose Daily/Weekly Turbidity Inf/Eff

DOC Inf/Eff

To identify obvious issues, assess UV (surrogate for TOC) Inf/Eff Online sensors 254 biofilter performance, and detect and grab changes in water quality that could Quality Oxidant residual (if pre‐oxidant samples Inf impact effluent quality or hydraulic is used) performance. Water pH Inf

Temperature Inf

DO consumption reflects biological Online sensor DO or LDO Inf/Eff activity and correlates well with TOC

Bio. removal.

Online sensor Headloss/ headloss To identify any obvious issues and and monitoring accumulation rate abnormal hydraulic changes that may

data Filter run time and/or UFRV require more attention

Backwash pressure/headloss Visual inspection Backwash bed expansion To identify any obvious issues and and online ensure that backwash process is Operational Backwash turbidity sensors effective Air scour rate/pressure Filter‐to‐waste time following Monthly

Temperature impacts the headloss (via Online sensor biological activity, water density, etc.). and monitoring Headloss/filter run time Filter run time should be adjusted data seasonally Operational

Grab samples ATP Media To assess health of biofilter Bio.

(Continued)

Biofiltration Guidance Manual for Drinking Water Facilities 121

Continued Annually Sample entire depth or filter for Determines particle size distribution if Filter coring media analysis media replacement is required

Assess filtering performance of solids

throughout the media bed. Before and Filter coring Floc/solids retention profile after backwash is ideal to evaluate Media backwash efficiency

Punch rods or Measurement of total media Evaluates if new media must be added Operational pipe depth to meet specified depth

Punch rods or Measurement of media Reveals if backwash is adequately pipe expansion during backwashing expanding filter bed (typically 25‐30%)

Analyze historical data and To assess the overall performance of the All online and evaluate long‐term filter run biofilter and implement and necessary grab sample data time trends changes for improvement Qual./Bio. Op./Water

5‐10 years

To evaluate the conditions and depth of Excavation of the filter bed Filter excavation each media layer. Also provides an through entire depth opportunity to inspect filter underdrains Operational

122 The Water Research Foundation

APPENDIX E

Biofilter Troubleshooting Guide

This appendix summarizes indicators, causes, and solutions to typical biofilter challenges.

Challenge #1: Variable Hydraulic Performance In general,

 With relatively stable temperature and flow rate, clean‐bed headloss should be consistent.  With relatively stable backwash flow, backwash pump discharge pressure (where applicable) should be consistent during each backwash event.  Monitoring pressure at backwash inlet at each filter (if possible) allows assessing backwash pressure effectively. Indicators: Flow rate, headloss, filter runtime, UFRV, filter effluent turbidity, particle counts (if applicable)

Methods to confirm:  Performance tracking tools, such as the one developed during WRF Project 4525 (Nyfennegger et al. 2016a), could be used to regularly monitor system performance through the use of trend plots and performance summaries.  Compare clean‐bed and terminal headloss, headloss accumulation rate, filter run time, and flow rates with historical trends.  Compare filter effluent turbidity and particle counts with historical data. Possible causes:

Cause #1: Malfunctioning biofilter control valves. Malfunctioning control valves can result in variable system performance since uniform division of the flow into the filters cannot be achieved and/or uniform flow conditions cannot be established in biofilters.

Mitigation strategies  Check the rate controllers for rapid cycling of valve position. Rapid movement last for several minutes.  Check if the cycling of rate control valves or “searching for valve set point position” is occurring. This may result in hydraulic surges through the filter, which can degrade filter performance.  Fix the control valve(s).

Cause #2: Faulty monitoring instruments/devices. Faulty readings can encourage operational changes, which can result in variable filter performance.

Biofiltration Guidance Manual for Drinking Water Facilities 123

Mitigation strategies  Evaluate if biofilter headloss pressure gauges and flow meters are functioning properly. Inadequate flow rate and pressure readings may result in variable filter run time, especially if not controlled by headloss. Excessive pressure build‐up can also damage mechanical parts and instruments.  Determine if the backwash pump and pressure gauges are functioning properly. While lower backwash flow rate may not achieve desired biomass removal, excess flow rate may result in media loss and/or excess removal of biomass. Inadequate backwash pressure reading may damage the underdrain or media bed.

Cause #3: Particle removal not achieved in upstream processes. Upstream coagulation/flocculation/clarification processes are designed to remove a certain number of particles. If particles are not adequately removed in these processes, biofilter may experience higher solids loading, resulting in rapid clogging and headloss buildup.

Mitigation strategies  Evaluate upstream treatment processes to ensure effective particle removal as designed.  Adjust coagulant dose and/or pH if needed.  If optimizing the upstream processes does not address the issue, backwashing protocol may need to be modified.

Cause #4: Variation in influent water quality. Changes in influent water characteristics, such as temperature, solid content, particle charge, BOM levels, and nutrient concentrations can result in variable hydraulic performance.

Mitigation strategies  Evaluate influent water characteristics.  If feed water solids content and turbidity levels have changed, adjust chemical dosing and system operation in upstream processes.  If feed water nutrient levels are low, consider supplementing the limiting nutrient (i.e., nitrogen or phosphorus) to enhance biological growth and activity in the biofilter. May consider assessing PHO:GLY ratio to determine phosphorus limitation.

Cause #5: Inadequate backwash. Backwash prevents irreversible biological fouling of the biofilter media, media support equipment, or filter underdrain through effective removal of excess biomass and deposited solids without impairing biological performance. Inadequate backwashing may result in fluctuating headloss due to excess biomass buildup, formation of mudballs, cracks, and preferential channels (i.e., hydraulic short circuiting). This could lead to irregular run times. Inadequate backwashing may also result in the generation of finer media particles and/or loss of media. Higher effluent turbidity and contaminant concentrations may also be observed.

124 The Water Research Foundation

Mitigation strategies  Check if backwash pumps are working correctly.  Evaluate backwash flow meters and pressure gauges.  Check if air blower is working properly (if applicable).  Check the functioning of backwash chemical feed (e.g., chlorine, peroxide, chloramines) feed system.  Check backwashing protocol to see if any unintentional changes occurred in the backwash settings.  If backwash settings are not changed, it is likely that the backwash protocol is not optimal, resulting in inadequate biomass removal (i.e., less/more biomass removal). Optimize backwashing steps.  If mudballs are observed, mudballs can be either: o removed using ‘mudball net’ during backwashing o broken using stiff garden rake o broken using high pressure washing o broken by adopting longer air scour cycles o broken by performing multiple backwashing If these steps do not eliminate the mudballs, take out the media, break the mudballs manually, and re‐pack (very expensive)

Challenge #2: Contaminant or Turbidity Breakthrough

Indicators: Effluent concentrations/levels of particles, turbidity, TOC, DOC, UV254, carboxylic acids, aldehydes, and contaminants of interest.

Methods to confirm:  Performance tracking tools, such as the one developed during WRF Project 4525, should be used to regularly monitor system performance through the use of trend plots.  Compare data on system specific biofilter effluent water quality parameters, such as turbidity, DOC, UV254, BOM, manganese, ammonia, T&O‐causing compounds, and trace organic contaminants, with historical data. Possible causes: Cause #1: Poor biological growth and activity. Poor biological growth and activity can lead to contaminant breakthrough.

Mitigation strategies Refer to evaluation and mitigation strategies suggested within Challenge #3.

Cause #2: Faulty feed pumps. Hydraulic loading rates/EBCT can vary if feed pumps are not working properly.

Biofiltration Guidance Manual for Drinking Water Facilities 125

Mitigation strategies  Compare real‐time flow rates with historical trend.  Check if the pump and flow meter are working correctly. Fix the issue. If needed, replace the pump. Cause #3: Inadequate backwash. Inadequate backwashing can lead to the formation of mudballs, cracks, and preferential channels. Worm holes cannot be corrected if backwashing is inadequate.

Mitigation strategies Refer to “Inadequate backwash” solution described under Challenge #1, possible cause #5.

Cause #4: Variation in biofilter influent water quality

Mitigation strategies  Compare biofilter influent water quality data (e.g., turbidity, contaminant concentrations, pH, and nutrient levels) with historical data trends.  Compare raw water quality with historical data. If the data show higher raw water TOC/DOC or contaminant concentrations, adjust/optimize upstream processes (coagulation/clarification processes).  If raw water concentrations are higher and biofilter influent concentrations cannot be controlled by upstream process optimization, operational modifications need to be made in biofilter operation.  Evaluate recent operational modifications to upstream processes. Any changes in coagulant dose, settled water pH, chlorine application, ozone dose, etc. may affect biofilter feed water characteristics, leading to contaminant breakthrough.  If biofilter influent nutrient levels are low, consider supplementing the limiting nutrient.  If manganese release is experienced, evaluate biofilter pH and pE (redox conditions) to determine manganese stability and the potential for additional release. Consider adjusting pH, supplementing nutrient, or adding non‐chlorine pre‐oxidant to improve manganese stability.  If these changes in biofilter operation do not help address the issue, biofilter loading rate may need to be lowered (i.e., increasing the EBCT). It is to be noted that this may lower water production at the facility.

Challenge #3: Poor Biological Growth and Activity Indicators: ATP, carboxylic acids, aldehydes, TOC, AOC, BDOC. Methods to confirm:  Evaluate if ATP in the biofilter is increasing over time. If not, biologicald growth an activity are not taking place.  Determine carboxylic acids, aldehydes, AOC, or BDOC removal by comparing influent and effluent concentrations. If there is little to no removal or if the removal is not increasing over time during the acclimation phase, biological growth is not occurring.

126 The Water Research Foundation

Possible causes: Cause #1: Low influent BOM. If the influent BOM levels are low, biological acclimation may not occur in a reasonable timeframe. Although a minimum BOM threshold (carboxylic acid, aldehydes, AOC, and BDOC) for supporting biological activity is not apparent in the literature. However, biofiltration typically removes 0.5 to 3/L mg TOC, suggesting that BDOC as low as 0.5 mg/L may support biological growth and activity in biofilters.

Mitigation strategies  Biofiltration may not be viable option.  Consider abiotic processes that may allow removing some of the contaminants (e.g., NH3, manganese).

Cause #2: Low influent nutrient concentrations. If influent N and/or P concentrations do not provide nutrient balanced conditions with respect to bioavailable carbon, microbial growth and activity can be slower. Typically, sustained biological growth can be achieved with a molar ratio of bioavailable C, N, and P of 100:10:1. This translates + into a mass ratio of 1 mg/L bioavailable carbon substrate: 0.117 mg/L NH4 –N: 3‐ 0.026 mg/L PO4 P. Evaluating PHO:GLY ratio may help accurately assess phosphorus limitation.

Mitigation strategies  Consider supplementing the limiting nutrient.

Cause #3: Influent temperature is not conducive to biological growth. Biofiltration has been successfully implemented in the temperature range of 5 to 30°C. Slower biological growth and activity can occur at temperatures outside of this range, requiring impractically long biological acclimation period.

Mitigation strategies  Nothing can be done. Biofiltration may not be a viable option.

Cause #4: Influent pH is not conducive to biological growth. Natural water sources typically have pH between 6 and 9 SUs, in which optimal biological growth is expected. However, changes in pH may occur due to the use of chemicals in upstream processes and biological activity, which may result in a very long acclimation period.

Mitigation strategies  Regularly monitor the biofilter influent and effluent pH.  If the pH falls beyond the normal range, consider adjusting the influent pH.

Biofiltration Guidance Manual for Drinking Water Facilities 127

128 The Water Research Foundation

APPENDIX F

Monitoring Tool Standard Operating Procedures

F.1 Standard Operating Procedure for ATP Analysis This Standard Operating Procedure is intended for use when analyzing samples using the LuminUltra Deposit and Surface Analysis (DSA) kit for ATP analysis. This SOP was adapted from Hooper et al. (2019). F.1.1 Materials The following materials are required for ATP testing:  Luminometer.  Laptop or mobile device with access to LuminUltra Cloud software.  LuminUltra DSA test kit and Instructions.  One, 100 μL micropipette.  One, 1 mL micropipette.  Drying oven.  Oven‐safe glass vials or metal weighing trays. F.1.2 Calibration 1. Ensure Luminase has been warmed to room temperature. 2. Perform ATP Standard Calibration by adding 100 µL of UltraCheck 1 and 100 µL of Luminase to a 12x55 mm test tube. 3. Swirl five times, then read using a luminometer. Record this result. F.1.3 Sample Analysis 1. Use a scale to measure out a media sample of approximately one gram and record sample weight. 2. Add the one gram of media to a 5 mL UltraLyse 7 tube. 3. Allow sample to incubate in the UltraLyse 7 tube for at least 5 minutes. 4. Transfer 1 mLd of liqui from the UltraLyse 7 tube to a 9 mL UltraLute Tube. Seal with the cap and invert 3 times. 5. Transfer 100 µL from UltraLute Tube to a new 12x55 mm test tube, add 100 µL of Luminase, and swirl gently 5 times. 6. Immediately insert the 12x55 mm test tube into the luminometer. Measure and record the result. 7. Record a temporary value of 1g in the amount box and adjust it later once the media has been dried overnight and weighed (see below). 8. Use LuminUltra software or conversion in procedure to calculate total ATP (tATP). 9. Dry the media sample overnight at 110 degrees Celsius and collect a final dry weight. 10. Run each sample in duplicate. 11. Enter results into the LuminUltra software to convert from RLU to picograms per gram ATP. 12. Normalize readings of ATP to the final dry weight of the media.

Biofiltration Guidance Manual for Drinking Water Facilities 129

F.2 Standard Operating Procedure for Biofilm Formation Rate This Standard Operating Procedure is intended for use for monitoring for Biofilm Formation Rate using ATP analysis. This is done by measuring biofilm formation on sample coupons. This SOP was adapted from Hooper et al. (2019). F.2.1 Installation and Materials Required for Testing The materials required for installation of the pipe loop and testing of the ATP sample coupons is included in the list below. Table F‐1 provides a detailed list of the supplies required to construct one pipe loop.  Piping and appurtenances for coupon installation (see Table F‐1)  Sample coupons (biofilm collectors) and coupon holders o Polycarbonate coupons: http://www.alspi.com/coupons‐flat.htm . Part # CO2649460004000 . Standard size of 1/2 X 3 X 1/8th of an inch. o PVC and Nylon coupon holders: http://www.alspi.com/holders‐flat.htm . Part # RC12EC100036 . Size: 3/4” plug, 3” nylon stem  50 mL sterile centrifuge tubes  Adhesive sample labels  Aluminum foil to cover clear PVC sections (if exposure to sunlight is possible)  Screwdriver For sample collection:  Sealable plastic bags  Kimwipes or paper towels  Coolers  Field Form  Chain of Custody Form  Nitrile gloves and safety glasses  Screwdriver  Ice for transfer and shipment of samples  Packing tape Table F‐1. Supplies Required for One Pipe Loop. Quantity Description Varies; at least 2 3/4” X 3” X 1/8” polycarbonate coupons 2 3/4” coupon holders 2 3/4” threaded female PVC couplings 7 3/4” x 10" threaded PVC pipe nipples 1 3/4" x 6" threaded clear PVC nipple 2 3/4” x 2" threaded PVC pipe nipples 5 3/4" threaded PVC elbows 2 3/4" threaded PVC tees 1 3/4" threaded PVC ball valve 1 3/8" flow control orifice (0.5 gpm) 2 3/4" x 3/8" hex bushings 2 3/8" x 2" PVC nipples

130 The Water Research Foundation

F.2.2 Experimental Setup and Maintenance General instructions for testing are included in the list below, and schematics of an example pipe loop are shown in Figure F‐1.  Coupons should be installed so that they are fully submerged and exposed to a constant flow of 0.5 gallons per minute (gpm). Flow should fall within ±0.05r gallons pe minute of 0.5 gpm.  Further consideration should be taken to ensure that the coupon assembly is easily accessible, and that flow can be visually verified through the effluent piping.  It is imperative that samples be stored properly upon collection, as most microbiological samples will begin to change upon sampling. Harvested coupons should be stored in the LuminUltra UltraLyseTM 7 test tube, placed in the refrigerator, and then shipped on ice as soon as possible. The sample is stable in the UltralyseTM7 tube up to one week in the refrigerator at 4 degrees Celsius, but if stored at room temperature then analyze within 24 hours, see Figure F‐2 for results form a holding time study.  Replace used Coupon B with new coupon and insert new Coupon A in sampling loopsg durin monthly sampling.  Orient coupon loops so the coupon holding sample loop is in the vertical position.  PVC section should use threaded fittings with Teflon tape so that PVC glue is not required.  Coupon holder for Coupon A should not be screwed in to allow air bubbles to exit the pipe loop.

Figure F‐1. Pipeloop Structure (a), Coupon Harvesting (b), and ATP Analysis (c). a) Coupons are installed into the flow stream in the PVC pipe at locations A and B. b) The coupon assembly is harvested after the incubation period. c) The harvested coupons are analyzed for ATP.

Biofiltration Guidance Manual for Drinking Water Facilities 131

UltraLyse7 Extract Tube Stability 1.0E+05 100%

90% 1.0E+04 80% Remaining

(pg/mL) 70%

ATP 1.0E+03

60% Percent

1.0E+02 50% 0 20406080100 Time (hr)

UltraLyse7 Extract ATP at Room Temperature UltraLyse7 Extract Stability at 4 Deg C Percent Remaining ‐ Room Temperature Percent Remaining ‐ 4 deg C Figure F‐2. ATP UltraLute7 Coupon Holding Time Study, Data Provided by Manufacturer. F.2.3 Sample Preparation and Data Analysis 1. Complete the label on each sample container, ensuring the following are included: a. Sample Location b. Sample Date c. Sample Time d. Initials of sampler 2. Complete field form and Chain of Custody form for each lab. 3. When a coupon is ready for extraction, remove the coupon from the sampling location. Coupons will be removed, and new coupons will be installed every two or four weeks. 4. Place aseptically in the LuminUltra UltraLyseTM 7 test tubes. To remove aseptically, wear new nitrile gloves to remove coupon from holder and place in tube. If coupon cannot be removed by hand, rinse a washed and rinsed screwdriver in isopropanol (rubbing alcohol), then used to unscrew coupon from holder. 5. Seal tube. 6. Clean coupon holder and screw using phosphate‐free soap and a clean brush. Rinse with DI water, then rinse with isopropanol to disinfect. 7. Aseptically attach an unused coupon to the holder by wearing new nitrile gloves and rinsing with isopropanol. 8. Reinstall new coupon, noting date and time coupon was installed. 9. Place collected samples in coolers and store in fridge for analysis. F.3 Standard Operating Procedure for Dissolved Oxygen Consumption This Standard Operating Procedure is intended for use for the installation, maintenance, and data analysis for dissolved oxygen (DO) consumption. This analysis is only recommended for utilities that do not use ozone as a pre‐oxidant. This SOP was adapted from Hooper et al. (2019). F.3.1 Materials The following materials are required for DO consumption testing.  2 Hach LDO probe (one at the filter influent and one at the filter effluent).  1‐2 flow‐through cells (filter effluent and/or filter influent).

132 The Water Research Foundation

 Tubing to connect flow‐through cell to filter influent/effluent.  0‐1 pole‐mounted stilling wells (only for filter influent if not using flow‐through cell).  Soft toothbrush.  Phosphate‐free detergent (e.g., Liquinox).  Sodium bisulfite solution (for removing iron deposits if dissolved iron concentration is greater than 0.1 mg/L). F.3.2 Calibration 1. The LDO probe comes factory calibrated but should be calibrated at the facility. 2. Enter the facility elevation or barometric pressure into the sensor. 3. Conduct one‐point calibration in air following manufacturer’s instructions. 4. Conduct, at a minimum, annual one‐point air calibrations of the sensor. Monthly calibrations are recommended. F.3.3 Installation 1. Set time stamp and logging interval on sensor. Logging interval should be between 5 and 60 minutes. Tie the data output to a SCADA system or separate memory device. 2. Install LDO probes at both the filter influent and filter effluent. a. The filter influent probe can be installed either in a flow‐through cell or pole‐mounted in a stilling well. b. The filter effluent probe should be installed in a flow‐through cell. 3. Mount LDO probe in flow‐through cell and connect flexible tubing from the filter effluent and/or filter influent to the cell. 4. The stilling well can be constructed of 3‐inch diameter polyvinyl chloride (PVC) pipe. a. Drill holes in bottom foot of the PVC stilling well. b. Mount probe within stilling well so that the bottom of the probe is flush with the bottom of the stilling well. c. Install probe and stilling well assembly within 6 inches of top of filter bed. F.3.4 Operations and Maintenance 1. LDO sensor inspections and cleaning a. Sensors should be inspected, cleaned, and calibrated monthly. b. Sensors can be cleaned with a soft toothbrush and a phosphate‐free detergent (e.g., Liquinox). c. Iron deposits will interfere with sensor readings. For water with dissolved iron concentrations greater than 0.1 mg/L, cleaning with sodium bisulfite will remove the deposits. d. Replace the sensor cap annually (as per Hach recommendations). 2. Flow‐through cells a. Flow rate through the flow‐through cells should be maintained between 0.5 and 50 gallons per minute (gpm) plus or minus percent. b. During startup, flow rates should be checked daily. c. Once flows are determined to be consistent, confirm flow rates weekly. d. If algae growth is suspected within the flow‐through cell, clean the apparatus between monthly inspections/calibrations. e. Note: Sensor drift is common in flow‐through cells and can be avoided by ensuring that adequate flow is passing through flow‐through cell at any given time. F.3.5 Data Analysis 1. Download timestamped filter influent and filter effluent DO data along with filter flow rate data from the SCADA system or other memory storage system.

Biofiltration Guidance Manual for Drinking Water Facilities 133

2. In a spreadsheet environment, include the date/time, filter flow rate, filter influent DO concentration, and filter effluent DO concentration in adjacent columns. 3. To calculate DO consumption across the biofilter, subtract the filter effluent DO concentration from the filter influent DO concentration under the following conditions: a. Filter must not be offline or backwashing; at a minimum, the filter flow rate must be greater than 0 gpm b. DO consumption must be greater than 0.2 mg/L (limit of detection for the method) 4. Note: Addition of new or regenerated granular activated carbon to the filter bed will sorb DO for approximately one to two weeks before the media becomes saturated. Data collected during this sorption time period should not be used for evaluating bioactivity until after the sensor readings have returned to prior DO concentrations. F.4 Standard Operating Procedure for EPS This Standard Operating Procedure is the protocol developed and tested in Keithley and Kirisits (2018) and is intended for use when analyzing samples using the Pierce™ BCA Protein Assay kit for EPS. This SOP was adapted from Hooper et al. (2019). F.4.1 Materials The following materials are required for EPS testing:  2 gram filter media collected  Spectrophotometer  Pierce™ BCA Protein Assay kit and Instructions  One, 10 mL pipette  One, 1 mL micropipette  Three, 25 mL graduated cylinders  One, 100 mL beaker  One, 100 mL Erlenmeyer flask  One, 15 mL conical vial  One, heating block safe and spectrophotometer compatible EPS sample vial  Sample vortexer  Temperature‐controlled centrifuge  Temperature‐controlled table shaker  Heating block  Laboratory scale F.4.2 Calibration 1. Ensure spectrophotometer has been calibrated using deionized water at the EPS absorbance wavelength (562 nm). F.4.3 EPS Extraction 1. Prepare extraction buffer by adding 10 millimolar (mM) tris(hydroxymethyl)aminomethane. (Tris), 2.5% sodium chloride, 10 mM ethylenediaminetetraacetic acid (EDTA) solutions to a 100 mL beaker. The pH of the solution should be 8 SU. 2. Weigh 2 gram filter media and add to a 15mL conical sample vial. 3. Add 15 mL of the extraction buffer to the 15 mL conical sample vial. 4. Vortex 15 mL conical sample vial at 2000 revolutions per minute (rpm) for 1 minute. 5. Incubate 15 mL conical sample vial on a temperature‐controlled shaker table at 200 rpm and 35 degrees Celsius (°C) for 4 hours.

134 The Water Research Foundation

6. Centrifuge 15 mL conical sample vial at 10,000 gram and 25°C for 10 minutes. F.4.4 Reagent Preparation and Sample Analysis 1. Prepare working reagent by adding 50 parts of BCA Reagent A and 1 parts of BCA Reagent B to a 100 mL Erlenmeyer flask. 2. Add 0.1 mL of the EPS extraction sample and 2.0 mL of the working reagent to a heating block safe and spectrophotometer compatible EPS sample vial 3. Incubate EPS sample vial at 37°C for 30 minutes. After incubation is complete allow to cool to room temperature (~25°C). 4. Add EPS sample vial to spectrophotometer and read at 562 nm. 5. Normalize readings of EPS (in mg bovine serum albumin) to the final dry weight of the media. F.5 Standard Operating Procedure for BDOC This Standard Operating Procedure is intended for use for analysis of biodegradable dissolved organic carbon (BDOC) or the fraction of dissolved organic carbon (DOC) that is readily biodegradable. This is done by adding a water sample to biologically active sand (BAS) and measuring DOC over time. F.5.1 Materials The following materials are required for BDOC testing:  Large (1‐Liter) Erlenmeyer flasks  MiliQ (ASTM type I) water  Hot plate and stir bar  Scale  TOC‐free flasks  TOC‐free aluminum boats  DI water  Aluminum foil  Incubator with shaker table  Mineral salt buffer (Difco 248510)  Acetate‐C  Autoclave • pH probe and meter

F.5.2 Inoculum Preparation and Maintenance F.5.2.1 Preparation of Chlorine‐Free Tap Water 1. Rinse a large Erlenmeyer flask with MilliQ (ASTM Type I) water. Fill the flask approximately 2/3 full with tap water. 2. Place the flask on a stirring hot plate and add a large stir bar to the flask. 3. Slowly heat the water to a boil, while stirring at medium speed. Allow water to boil for 10 minutes and cool to room temperature.

Biofiltration Guidance Manual for Drinking Water Facilities 135

F.5.2.2 BAS Collection and Shipment 1. BAS is collected from a sand filter prior to the chlorination step in a WTP. The BAS must be taken after a backwash, to remove algae buildup. BAS should be taken as a core, or otherwise excavated to account for vertical heterogeneity within the sand filter. 2. The BAS should be carefully placed into a wide mouth container until it is half full. The BAS should not be packed or shaken, as this may disrupt the biofilm. The large headspace is meant to prevent the BAS from going anaerobic during shipment. 3. Raw influent water collected immediately before the sand filter should be added until the BAS is saturated. Seal the container tightly. The BAS must not dry out during shipment. 4. BAS should be shipped overnight on ice. A temperature blank should be included with each shipment. F.5.2.3 BAS Storage 1. Rinse BAS provided three times with 3:1 (v:w) of freshly dechlorinated tap water with and then three times with 3:1 (v:w) DI water. Measure TOC of the last washing water and compare to DG. Sand can be used when the difference in DOC between washing water and DI is <0.1 mg/L. 2. BAS stock is stored at room temperature in raw water from the treatment plant or dechlorinated tap water. The ratio of water to sand is 1.3:1 (v:w). 3. If using raw water from the treatment plant, change out the water once every two to three weeks. 4. If using dechlorinated tap water, change out the water once a week, by decanting the water from the flask, and adding freshly dechlorinated tap water. 5. Aerate the stock at 4 L/hr or keep on a shaker table at 120 RPM. F.5.2.4 BAS Preparation 1. Weigh out the amount of BAS need for the experimental run from the BAS stock into a TOC‐free glass container (bake at 550 Deg C for 4 h). 2. Rinse the stock three times with 3:1 (v:w) of freshly dechlorinated tap water. 3. Decant as much water as possible before weighing sand. F.5.3 Sample Analysis 1. Rinse the appropriate number of 1L Erlenmeyer flasks with DI water. 2. Weigh out 200 gram of BAS sand onto a TOC‐free aluminum boat. 3. Add BAS to each TOC‐free flask. Gently rinse the sand five times with 100 mL of DI water. Decant water between rinses rather than using a pipette. 4. Add 600 mL of the water to be tested to each flask. The ratio of water to BAS should be 3:1 (v/w). 5. Swirl the flask to gently mix (mix gently enough to not stir up the BAS). 6. Pull a baseline sample for DOC analysis (prior to mixing with sand, D0) and a second sample for DOC after mixing with sand (D1). Initial DOC is the average of D0 and D1. The difference between D1 and D0 must be <0.1 mg/L. If the difference is > 0.1 mg/L, the BAS must be rinsed again prior to initiating test. Collect remaining baseline samples as described in 4.0. 7. Cover flasks with aluminum foil, and place in the incubator. 8. Set the incubator to 20C. Set the shaker to 120 RPM or aerate at 4L/hr. 9. Remove samples for analysis (see 4.0) per sampling plan frequency, at a minimum once per day. Continue sampling for seven days, or until a DOC minimum has been observed. 10. A QC standard will be included with each run. Use an autoclaved mineral salt buffer (Difco 248510) with pH adjusted to 7.2, and 2 mg of acetate‐C per 300 mL of buffer. To prepare mineral salt buffer: a. Dissolve 56.4 gram of the powder in 1 L of DI water b. Autoclave at 121°C for 15 minutes. c. Add 200 mL sterile M9 Minimal Salts, 5× (Difco 248510) to 750 mL sterile DI water cooled to 45‐ 50°C, adjusting the final volume to 1 liter. This creates the 1X dilution.

136 The Water Research Foundation

APPENDIX G

Tools Compendium

Table G‐1 details guidance tools for drinking water systems either operating or considering biofiltration. These tools provide utilities with case studies that identify peer facilities as well as provide optimization strategies and monitoring tools that allow utilities to more actively manage biofiltration to meet expanding treatment goals. Table G‐1. Biofiltration Tools. Tool Name Type Description Access BIOFILT Model Numerical  Published by Hozalski and Bouwer in 2001 www.waterrf.org/research Model  Evaluated in AWWARF Project 252 /projects/optimizing‐  Simulates removal of biodegradable organic matter under filtration‐biological‐filters varying operational conditions  Simulates the initial acclimation process and the effects of a sudden loss in biomass Biofiltration Online  Published in 2016 as part of WRF Project 4459 www.waterrf.org/research Knowledge Base Database  Database that hosts utility data from 45 full‐scale biofiltration /projects/north‐american‐ facilities on biofiltration planning, evaluation, design, and operation biofiltration‐knowledge‐  Includes 21 case studies, 10 data reports, a biofiltration base library, and frequently asked questions Biofiltration Microsoft  Published in 2016 as part of WRF Project 4525 www.waterrf.org/research Performance Excel  Facilitates compilation, analysis, and interpretation of raw /projects/full‐scale‐ Tracking Program Program biofiltra‐tion data files (e.g., data that might be extracted from engineered‐biofiltration‐ a SCADA) evaluation‐and‐ development‐performance Biofilter Conversion Manual  Published in 2017 as part of WRF Project 4496 www.waterrf.org/research Guidance Manual  Summarizes full‐scale case studies, best practices, and /projects/biofilter‐ common troubleshooting strategies for converting from conversion‐guidance‐ conventional to biological filtration manual Biofilter Conversion Microsoft  Published in 2017 as part of WRF Project 4496 www.waterrf.org/resource Assessment Tool Excel  Indicates the relative suitability of converting to /biofilter‐conversion‐ Program biofiltration and recommended mitigation factors based on assessment‐tool 24 multiple choice questions  Compares that facility to facilities in the Biofiltration Knowledge Base Recommended Standards  Published in 2018 www.health.state.mn.us/c Standards for Water  Includes an updated standard on aerobic biofiltration of ommunities/environment/ Works (Ten States surface water (Section 4.3.7) water/tenstates/standards. Standards)  Includes new policy statements on aerobic and anoxic html biological treatment of groundwaters The Guidance Manual  Published in 2019 as part of WRF Project 4620 www.waterrf.org/research Manual for  Recommends monitoring tools that provide rapid feedback /projects/guidance‐ Monitoring on biofilter performance manual‐monitoring‐ Biological Filtration  Provides instructions on how and when to implement these biological‐filtration‐ of Drinking Water tools, how to interpret the data, and how to translate the drinking‐water results to enhance biofilter performance Manual of Practice Manual  Under development, to be published in 2022. www.awwa.org/Publicatio (M80): Biological  Will provide operational and design guidance on high‐rate ns/Manuals‐of‐Practice Drinking Water aerobic biofiltration, anoxic biological treatment, biological Treatment treatment in reuse applications, monitoring and operational controls for biological treatment, and full‐scale case studies for each of these applications

Biofiltration Guidance Manual for Drinking Water Facilities 137

138 The Water Research Foundation

APPENDIX H

Select Biofilter Optimization Studies

Author / WRF Project Number Project Title Optimization Strategies Investigated Singh Sidhu et al. (2018)  Pre‐oxidation strategies for biofiltration  Ozone; hydrogen peroxide + performance enhancement ozone Beniwal et al. (2018)  Ozone/peroxide advanced oxidation in  Ozone; hydrogen peroxide + combination with biofiltration for taste and odor ozone control and organics removal McKie et al. (2015)  Engineered biofiltration for the removal of  With and without ozone: disinfection by‐product precursors and nitrogen; phosphorus; hydrogen genotoxicity peroxide McKie et al. (2016)  Conventional drinking water treatment and direct  Ozone biofiltration for the removal of pharmaceuticals and artificial sweeteners: A pilot‐scale approach Lee et al. (2012)  Ozone and biofiltration as an alternative to  Ozone reverse osmosis for removing PPCPs and micropollutants from treated wastewater Xing et al. (2018)  Effects of phosphate‐enhanced  Ozone + phosphorus ozone/biofiltration on formation of disinfection byproducts and occurrence of opportunistic pathogens in drinking water distribution systems Sun et al. (2018)  A pilot‐scale investigation of disinfection by‐  Ozone product precursors and trace organic removal mechanisms in ozone‐biologically activated carbon treatment for potable reuse Ross et al. (2019)  Effects of water quality changes on performance  Ozone + phosphorus of biological activated carbon (BAC) filtration de Vera et al. (2019)  Using upstream oxidants to minimize surface  Oxidants (chlorine, biofouling and improve hydraulic performance in monochloramine, hydrogen GAC biofilters peroxide); backwashing with oxidants Zhao et al. (2019)  Impact of carbon‐based nutrient enhancement  Carbon‐based nutrients (amino on biofiltration performance for drinking water acids, inulin, sucrose) treatment Black and Berube (2013)  Rate and extent NOM removal during oxidation  Ozone; UV/hydrogen peroxide and biofiltration Liu et al. (2001)  Factors Affecting Drinking Water Biofiltration  Backwashing with oxidants, backwashing strategies Lauderdale et al. (2012)  Engineered biofiltration: Enhanced biofilter  Nutrients (nitrogen and performance through nutrient and peroxide phosphorus); hydrogen peroxide addition Selbes et al.  Evaluation of Seasonal Performance of  Phosphorus (2016) Conventional and Phosphate‐Amended Biofilters Water Research Foundation Projects WRF 504  Ozone and Biological Treatment for DBP Control  Ozone and Biological Stability

Biofiltration Guidance Manual for Drinking Water Facilities 139

Author / WRF Project Number Project Title Optimization Strategies Investigated WRF 252  Optimizing Filtration in Biological Filters  Ozone; backwashing with (Huck et al. 2000) oxidants WRF 2775  Ozone‐Enhanced Biofiltration for Geosmin and  Ozone; backwashing with (Westerhoff et al. 2005) MIB Removal oxidants WRF 4215  Engineered Biofiltration for Enhanced Hydraulic  Nutrients (nitrogen and (Lauderdale et al. 2011) and Water Treatment Performance phosphorus); hydrogen peroxide, substrate augmentation WRF 4346  Optimizing Engineered Biofiltration  Nutrients (nitrogen and (Lauderdale et al. 2014) phosphorus); pH;hydrogen peroxide; ozone WRF 4448  Optimizing Filter Conditions for Improved  Nutrients (phosphorus), pH, (Lauderdale et al. 2016) Manganese Control During Conversion ‐ Report substrate augmentation, 4448 (WRF 2016) WRF 4525  Full‐Scale Engineered Biofiltration Evaluation and  Hydrogen peroxide, data (Nyfennegger et al. 2016) Development of a Performance Tracking Tool ‐ analytics WRF 4429  Chemically Enhanced Biological Filtration to  Enhanced coagulation; oxidants (Evans et al. 2016) Enhance Water Quality and Minimize Costs (permanganate, chlorine dioxide, ferrate) WRF 4555  Optimizing Biofiltration for Various Source Water  Nutrients (nitrogen and (Lauderdale et al. 2018) Quality Conditions phosphorus); pH adjustment; Oxidants (ozone, chloramine, permanganate, chlorine, peroxide); backwashing with oxidants WRF 4669  Biological Filtration: NDMA Control or Source of  Ozone; media type; pH (Evans et al. 2019) Precursors? adjustment; backwashing with oxidants; backwashing frequencies WRF 4749  Media type; pH adjustment;  Optimizing Biofiltration for Improved Manganese (Evans et al. 2020) hydrogen peroxide; EBCT, Control under Winter Conditions shutdown procedures

140 The Water Research Foundation

APPENDIX I

Biofilter Optimization Decision Trees

Figure I‐1. Reducing Algal Growth with Oxidant Addition.

Biofiltration Guidance Manual for Drinking Water Facilities 141

Figure I‐2. Optimizing Biofilter Hydraulics with Oxidants.

142 The Water Research Foundation

Figure I‐3. Optimizing Biofilter Hydraulics with Nutrient Addition.

Biofiltration Guidance Manual for Drinking Water Facilities 143

Figure I‐4. Alleviating Hydraulic Issues with Oxidant Addition and Modified Backwash Protocols.

144 The Water Research Foundation

Figure I‐5. Improving Organics Removal with Oxidant Addition.

Biofiltration Guidance Manual for Drinking Water Facilities 145

Figure I‐6. Improving Organics Removal with Nutrient Addition.

146 The Water Research Foundation

APPENDIX J

Select Biofilter Conversion Case Studies

Treatment Conversion Capacity Facility Location Water Source Year Conversion Driver (mgd) Oxidant/Media East River Davenport, Mississippi T&O, turbidity Station 2006 30 No oxidant/GAC IA River spikes, atrazine WTP Shark Jumping Monmouth River/Glendola 2006 Costs 14 Pre‐chlorine/GAC Brook WTP County, NJ Reservoir E. H. Aldrich Monongahela Station Elrama, PA 2006 Costs 50 Pre‐chlorine/GAC River WTP Richmond Kentucky River Lexington, River/Jacobson 2006 Manganese, T&O 25 Pre‐chlorine/GAC Station KY Reservoir WTP Swimming Particle removal, Swimming Colts Neck, River 2006 algae growth, 36 Ozone /GAC River WTP NJ Reservoir manganese Delaware River Delaware Delran, NJ 2006 Costs 30 Ozone /GAC Regional River WTP Washington Quail Creek County Reservoir/ DBP Precursors/ Hurricane Water Sand Hollow 2008 Chemical cost 60 Chlorine/Anthracite City, UT Conservanc Reservoir/ reduction y District Virgin River Dublin Columbus, 2015/ Road Water Scioto River DBP precursor 80 Ozone/GAC OH 2016 Plant Greenway Peoria, AZ Salt River 2012 ‐ 16 Ozone /GAC WTP Greenville Greenville, Tar River 2015 Biostability 22.5 Ozone /GAC WTP NC Greenville Greenville, Tar River 2015 Biostability 22.5 Ozone/anthracite WTP NC Tarrant BOM removal, County Euless, TX Lake Arlington 2014 87 Ozone/anthracite biostability WTP

Biofiltration Guidance Manual for Drinking Water Facilities 147

148 The Water Research Foundation

APPENDIX K

Full‐Scale Biofiltration Plants

Design Source Water Capacity Utility Type (mgd) Biofiltration Driver Pre‐Treatment Media Type Aurora Water Blend and 20 ‐ 33.3 TOC removal, AOC Coagulation GAC Lake stabilization, and taste/odor removal City of Fort Worth Lake 12 ‐ 200 TOC removal, DBP Coagulation, ozone Sand/anthracite precursor removal, taste/odor removal, filter performance and sustainable treatment City of Denton Lake 20 TOC Removal Coagulation, ozone Sand/anthracite Region of Peel, ON Lake 105 Taste/odor removal, Ozone, Biological GAC water stability Contactors, Membranes Gwinnett County Lake 75 ‐ 150 Filter performance Coagulation, ozone Sand/anthracite Greater Cincinnati River 240 DBP Precursor Removal, Coagulation Sand Water Works Taste and Odor Removal, TOC Removal Central Lake County Lake 52 Turbidity removal Coagulation, ozone GAC Joint Action Water Agency, IL Tacoma Water River – DBP precursor removal, Pre‐ozonation Anthracite/sand manganese oxidation followed by conventional filtration Lake Oswego Tigard River 38 DBP precursor removal, Actiflo followed by GAC/sand Water Treatment taste/odor removal intermediate Plant ozonation Tampa Bay Water River 120 TOC removal Actiflo, ozone GAC/sand Aurora Water Blend and 20 ‐ 33.3 TOC removal, AOC Coagulation GAC Lake stabilization, and taste/odor removal City of Fort Worth Lake 12 ‐ 200 TOC removal, DBP Coagulation, ozone Sand/anthracite precursor removal, taste/odor removal, filter performance and sustainable treatment City of Denton Lake 20 TOC Removal Coagulation, ozone Sand/anthracite Region of Peel, ON Lake 105 Taste/odor removal, Ozone, Biological GAC water stability Contactors, Membranes Gwinnett County Lake 75 ‐ 150 Filter performance Coagulation, ozone Sand/anthracite

Biofiltration Guidance Manual for Drinking Water Facilities 149

Design Source Water Capacity Utility Type (mgd) Biofiltration Driver Pre‐Treatment Media Type Greater Cincinnati River 240 DBP Precursor Removal, Coagulation Sand Water Works Taste and Odor Removal, TOC Removal Central Lake County Lake 52 Turbidity removal Coagulation, ozone GAC Joint Action Water Agency, IL Tacoma Water River – DBP precursor removal, Pre‐ozonation Anthracite/sand manganese oxidation followed by conventional filtration Lake Oswego Tigard River 38 DBP precursor removal, Actiflo followed by GAC/sand Water Treatment taste/odor removal intermediate Plant ozonation Tampa Bay Water River 120 TOC removal Actiflo, ozone GAC/sand Aurora Water Blend and 20 ‐ 33.3 TOC removal, AOC Coagulation GAC Lake stabilization, and taste/odor removal City of Fort Worth Lake 12 ‐ 200 TOC removal, DBP Coagulation, ozone Sand/anthracite precursor removal, taste/odor removal, filter performance and sustainable treatment City of Denton Lake 20 TOC Removal Coagulation, ozone Sand/anthracite

150 The Water Research Foundation APPENDIX L

Testing Plan Template

151

Testing Plan Template Every testing plan will be unique. The level of detail included should be tailored to each project’s goals and the team involved. The template provided here is an example that covers the key considerations discussed in Section 6.

Outline Instructions

Step 1: Add background on the INTRODUCTION project, utility (if applicable), and purpose of testing biofiltration (i.e., what business case was developed for Background conducting this study?).

If a desktop evaluation or testing at a different scale was conducted previously, summarize the results. Testing Objectives Step 2: Define and add site‐ specific and measurable testing Example operational objectives could be: objectives.  Less than 6 ft of headloss development over 72 hours [Refer to Section 7.1] during the summer  Less than 0.1 NTU filter effluent turbidity (>95% of readings) over 48 hours at the maximum filter loading rate of 4 gpm/sf during each testing season with a filter‐to‐ waste duration of less than 30 minutes  Removal of at least 10% dissolved organic carbon when water temperatures are below 10°C and at least 20% when water temperatures are above 20°C at a filter loading rate of 4 gpm/sf Full‐Scale Water Treatment Plant Processes Step 3: Add description of treatment processes, size (i.e., MGD) of WTP, key chemicals added for control and removal Figure L‐1. Full‐Scale WTP Process Flow Schematic. of various water quality parameters. Include a process flow schematic. [Refer to Section 7.2.3] Section continued on next page.

152 The Water Research Foundation

Outline Instructions

Table L‐1. Full‐Scale Filter Design and Operational Parameters. Step 4: Add full‐scale filter WTP – WTP – design and operational Parameter Units Design Typical/Average parameters that will be used as Flow MGD the basis of design for the Filter Loading Rate gpm/sf testing system. Parameters Empty Bed Contact Time min (EBCT) include design flow, media type Number of Filters and depth, empty bed contact Filter Surface Area sf time (EBCT), backwash Typical Backwash Flow gpm/sf strategies and triggers, etc. Rate Include both design and actual Target Bed Expansion % (e.g., media characterization Available Height for Bed ft Expansion from the most recent sieve Backwash Triggers Run time analysis) criteria as available. (hours) Filtered [Refer to Section 7.2.3 on Turbidity (NTU) Benchmarking for a list of Headloss (ft) design and operational factors Other that should be considered and Air Scour Rate scfm/sf documented.] Air Scour Duration min Surface Wash Duration min Also add or attach relevant Filter‐to‐Waste Rate gpm/sf SOPs, including the Filter‐to‐Waste Duration min backwashing protocol by step, Filter‐to‐Waste Flow gpm/sf for reference. Example tables are provided. Table L‐2. Full‐Scale Filter Media Characteristics.

Media Design or Depth, Effective Uniformity L/d10 Type Actual inches Size, mm Coefficient Ratio Anthracite 22 0.90 <1.8 700 Layer 2 Design Total – – Layer 1 Actual Layer 2 (DD/MM/YY) Total – –

Biofiltration Guidance Manual for Drinking Water Facilities 153

Outline Instructions Historical Water Quality and Performance Step 5: Summarize relevant historical water quality and performance data collected at Table L‐3. Summary of Historical Water Quality Sampling Points, different sample points in the Parameters, and Frequencies. treatment train, with specific Sampling Point Parameters Number of Samples focus on assessing the filter Raw Water pH 50 influent water quality and historical filter performance. Data over multiple years, Table L‐4. Water Quality and Performance Summary. capturing the range of water quality and operating Raw Coagulated / Filtered Tap Parameter Water Settled Water Water Water conditions anticipated in the Turbidity (NTU) future should be considered. th 5 Percentile 3.3 0.18 0.022 0.021 [Refer to Section 7.2.1 on Average 13 0.45 0.053 0.052 95th Percentile 66 2.7 0.074 0.075 Benchmarking for Parameter 2 (Units) recommended water quality 5th Percentile parameters and analysis.] Average 95th Percentile Describe any trends in the data over time and specifically in the filter influent and performance across the filters. These trends will help establish the testing approach. Example tables are provided.

Relevant Regulations and Permits Step 6: If applicable, add any relevant state or USEPA

regulations or regulatory guidance (e.g., if new media profiles are to be tested) that were considered when developing the testing plan. Also include any permits that may be required (e.g., for disposing of waste streams).

154 The Water Research Foundation

Outline Instructions

Step 7: Layout the approach TESTING APPROACH and experimental conditions to be tested. Add an overview of the testing system, including the testing scale, where testing will be conducted (e.g., at a WTP vs. at a University), where Figure L‐2. Overall Testing Diagram. the source water will come from, and the overall planned

duration of testing. Include an overall process flow schematic, if helpful, to show how the test filters will intersect with the WTP and/or any upstream testing equipment. [Refer to Section 7.3 on Testing Scales]

Biofiltration Guidance Manual for Drinking Water Facilities 155

Outline Instructions Test Filter Design Step 8: Describe how the test filters were designed and provide a process flow diagram. An example is provided. Summarize the key design and operational criteria for the test filters in a table. An example table is provided. [Refer to Section 7.5 on Testing Design]

Figure L‐3. Example Test Process Flow Diagram (Evans et al. 2020).

‐ Table L 5. Test Filter Design and Operational Parameters. Test – Test – Parameter Units Design Typical/Average Flow gpm Filter Loading Rate gpm/sf Empty Bed Contact min Time (EBCT) Number of Filters Filter Diameter inches Filter Surface Area sf Backwash Flow Rate gpm/sf Target Bed Expansion % Available Height for ft Bed Expansion Backwash Triggers Run time (hours) Filtered Turbidity (NTU) Headloss (ft) Other Air Scour Rate scfm/sf Air Scour Duration min Surface Wash min Duration Filter‐to‐Waste Rate gpm/sf Filter‐to‐Waste min

Duration Filter‐to‐Waste Flow gpm/sf

156 The Water Research Foundation

Outline Instructions

Equipment and Instrumentation Step 9: Add a detailed description of the testing Table L‐6. Testing Equipment Specifications. equipment and selected Parameter Units Specification instrumentation. Describe how Maximum Total Flow Rate gpm the testing equipment will be Filters Inches procured (e.g., rental vs Maximum Media Depth Inch Filtration Rate gpm/sf constructed on‐site), what Backwash Rate gpm/sf operational functions will be Backwash Tank Volume Gal automatic and manual, and Air Scour Rate scfm/sf what instruments will be Dimensions Inches available online. Describe the Skid Dimensions Inches overall system specifications; Skid Weight lbs Power Phase, A, VAC an example table is provided below. Consider including photos or images of the testing equipment. [Refer to Section 7.5.4] Section continued on next page.

Step 10: Add description of where the testing equipment will be located. Discuss provisions for power supply, HVAC, drains, waste stream handling, space for accessing all sides of the system for maintenance, secondary Figure L‐4. Example Photo of Testing Equipment (Evans et al. 2020). containment for chemicals, safety equipment (e.g., eye wash), etc. Also consider how the system will be setup (e.g., doorway opening requirements). Consider Figure L‐5. Photos and/or Schematics of the Testing Location. including photos or schematics of where the equipment will be located and where nearby power, drains, safety equipment, etc. are located.

Biofiltration Guidance Manual for Drinking Water Facilities 157

Outline Instructions

Media Design and Acclimation Step 11: Outline media characteristics for each of the Table L‐7. Filter Media Profiles. media profiles to be tested. Media Effective Include media type, media Filter Media Depth size EBCT depth, effective size, calculated No. Type (in) (mm) L/d10 (min) Comments 1 Anthracite 22 0.80 Control; L/d and EBCT. Consider Sand 10 0.45 matching including a control column. Total 32 ‐ 1260 7 full‐scale 2 Layer 1 Layer 1 Layer 1 Layer 1 Layer 1 An example table to document Layer 2 Layer 2 Layer 2 Layer 2 Layer 2 media characteristics is Total Total Total Total Total provided.

[Refer to Section 7.5.2 and Table 7‐6]

Step 12: Add and describe where media will be sourced from and how media will be acclimated. Describe the parameters that will be monitored and what values will be used to signify acclimation has been achieved.

Backwash and Filter‐to‐Waste Step 13: Add and describe the backwashing and filter‐to‐ Table L‐8. Backwashing and Filter‐To‐Waste Procedures. waste (if applicable) Step Rate Duration procedures. For backwashing, describe the backwash water

source, any chemicals added

(e.g., oxidant or coagulant), agitation method, rates and durations for each backwashing step and any automation. For filter‐to‐ waste, describe the flow, duration or trigger for ending filter‐to‐waste, and any automation. An example table is provided. [Refer to Section 7.5.2 and Table 7‐7]

158 The Water Research Foundation

Outline Instructions

Chemical Systems Step 14: Add and describe any chemicals that will be dosed Table L‐9. Chemicals. (e.g., filter aid, oxidants, Specific Feed nutrients, contaminant spiking) Chemical Source %w/w Gravity Point Purpose and the chemical systems (e.g., Sodium LabChem 50 1.5 Combined pH Hydroxide Filter adjustment pump types, level of Influent automation, etc.). Label chemical feed points on the testing process flow diagram.

[Refer to Section 7.5.2 and Table 7‐7 and 7‐8]

Testing Conditions Step 15: Add and describe the range of conditions that need Table L‐10. Testing Conditions and Ranges. to be tested to meet the Range to be defined testing objectives. Phase Conditions Tested Flow Include testing ranges that Water Temperature both reflect typical operations Pre‐Treatment and sufficiently challenge Target Contaminant(s) and Concentrations (Table 7‐9) the biofilters. Filter Aid Consider defining specific Filter Influent Chemical Enhancement testing phases (i.e., sets of Seasonal Conditions conditions) that test each Shutdowns variable independently. An example table is provided. A timeline is often used in place of a table. Compare the testing conditions to the filter design to ensure the design allows for testing the full range of conditions identified (e.g., that a sufficient number of chemical pumps are included and capacities match the flow ranges needed). [Refer to Section 7.5.3]

Biofiltration Guidance Manual for Drinking Water Facilities 159

Outline Instructions Testing Schedule Step 16: Add a testing schedule and deliverable dates. The overall testing duration should Table L‐11. Testing Schedule. allow for acclimation and Dates Chemical baseline testing prior to (From Flow EBCT Condition Dose 1 changing test conditions. Allow – To) Duration Phase (gpm/sf) (min) 1 (mg/L) for reaching steady‐state Acclimation Baseline performance before each Testing variable change. Include Condition contingency for unexpected Set A delays and to allow for Condition additional testing that may be Set B warranted following review of Contingency TBD preliminary results. An example table is provided. [Refer to Section 7.5.1]

Monitoring Plan Step 17: Identify and add water quality parameters to be

measured in the field and lab. TableL‐12. Monitoring Plan. Include sample location and Example frequency for the duration of Category Parameters Frequency Location Laboratory testing. Include any Flow operational parameters that Operations Headloss Online Filters On‐Site Run times will monitored on a daily, weekly, or monthly basis.

See Table 7‐12 for an example.

[Refer to Section 7.5.5]

Analytical Methods Step 18: Identify and add the methods that will be used for

each parameter. Include a subsection detailing the sample collection, preservation, shipping and analysis methods for any advanced analytical techniques, such as emerging contaminants, microbial community analysis, or media characterization.

160 The Water Research Foundation

Outline Instructions Data Management and Interpretation Step 19: Identify and add how data included field and lab, will

be recorded, maintained, and interpreted. Note any automatic data logging of key parameters, remote monitoring and control using a standard web browser, and email alarm notifications. Identify data management backup protocols. [Refer to Section 7.5.6]

Quality Control Step 20: Add and describe the quality control measures that

will be implemented and monitored through testing to ensure high quality data are obtained and the testing plan is routinely reviewed in parallel with preliminary testing results. Common quality control measures are summarized in Table 7‐14. [Refer to Section 7.5.8]

Biofiltration Guidance Manual for Drinking Water Facilities 161

Outline Instructions

Step 21: Identify and add roles ROLES AND RESPONSIBILITIES and responsibilities for each Example responsibilities might include: participating party (e.g., consultant staff, university During Procurement and Start‐Up: students, laboratories,  Lead/assist with pilot design and pilot components vendors, and utility staff). procurement Example responsibilities are  Accept delivery of pilot plant components provided.  Lead/assist with pilot setup and construction activities  Lead/assist startup and troubleshooting and resolve all [Refer to Section 7.5.7] startup issues  Train test equipment operators/participate in training

During Routine Operations:  Operate and maintain the testing equipment  Monitor testing equipment during weekends/holidays  Lead/assist with procuring chemicals and replenishing stock solutions  Perform field measurements and maintain field logs  Provide sample bottles, labels and chain‐of‐custody forms  Collect and ship samples to labs  Analyze samples for laboratory parameters  Document, analyze and interpret data

Step 22: Identify and add SAFETY appropriate personal protective equipment to be worn by staff on site and other safety protocols that should be implemented and followed. Include emergency contact information. Consider attaching project‐specific Health and Safety Plans and/or organizational policies regarding safety procedures.

[Refer to Section 7.5.9]

162 The Water Research Foundation

References

Ahmad, R., A. Amirtharajah, A. Al‐Shawwa, and P.M. Huck. 1998. Effects of backwashing on biological filters. Journal / American Water Works Association, 90(12), 62‐73. Alito, C. and A‐J Wangner. 2018. Challenges and Mitigation Strategies for Granular Media Filtration Operation.

Allen, M., S. Edberg, and D. Reasoner. 2004. Heterotrophic plate count bacteria – what is their significance in drinking water? Int. J. Food Microbiol. 92, 265‐274. https://doi.org/10.1016/j.ijfoodmicro.2003.08.017. Allgeier, S., R. Summers, J. Jacangelo, V. Hatcher, D. Moll, S. Hooper, J. Swertfeger, and R. Green. 1996. Simplified and Rapid Method for BDOC Measurement. Boston, MA: Proc. 1996 AWWA WQTC. American Water Works Association (AWWA). 1998. Water Treatment Plant Design. AWWA, Denver, Colorado. Amirsardari, Y., Q. Yu, and P. Williams. 1997. Effects of Ozonation and Coagulation on Turbidity and TOC Removal by Simulated Direct Filtration for Potable Water Treatment. Environmental Technology, 18(11), 1143‐1150. https://doi.org/10.1080/09593331808616633. Andrews, S. and P. Huck. 1994. Using Fractionated Natural Organic Matter to Quantitate Organic Byproducts of Ozonation. Ozone Sci. Eng. 16, 1‐12. https://doi.org/10.1080/01919519408552376. APHA, AWWA, WEF, 2012. Standard Methods for the Examination of Water and Wastewater, 22nd Edition, 22nd ed. Denver, Colorado. Archer, A. and P. Singer. 2006. Effect of SUVA and enhanced coagulation on removal of TOX precursors. J. ‐ Am. Water Works Assoc. 98, 97‐107. https://doi.org/10.1002/j.1551‐8833.2006.tb07737.x. Arnold, M., J. Batista, E. Dickenson, and D. Gerrity. 2018. Use of ozone‐biofiltration for bulk organic removal and disinfection byproduct mitigation in potable reuse applications. Chemosphere 202, 228‐ 237. https://doi.org/10.1016/j.chemosphere.2018.03.085. Becker, W.C., C. O'Melia, K.K. Au, and J.S. Young. 2006. Using Oxidants to Enhance Filter Performance. Final report to the American Water Works Association Research Foundation, Denver. Colorado. Benner, J., D.E. Helbling, H.‐P.E. Kohler, J. Wittebol, E. Kaiser, C. Prasse, T.A. Ternes, C.N. Albers, J. Aamand, B. Horemans, D. Springael, E. Walravens, and N. Boon. 2013. Is biological treatment a viable alternative for micropollutant removal in drinking water treatment processes? Water Res. 47, 5955‐ 5976. Berney, M., M. Vital, I. Hülshoff, H.‐U. Weilenmann, T. Egli, and F. Hammes. 2008. Rapid, cultivation‐ independent assessment of microbial viability in drinking water. Water Res. 42, 4010‐4018. https://doi.org/10.1016/j.watres.2008.07.017. Beverly, R.P. 2005. Filter Troubleshooting and Design Handbook, Denver, Colo.: American Water Works Association. Bouwer, E.J. and P.B. Crowe. 1988. Biological Processes in Drinking Water Treatment. J‐AWWA 80, 82‐ 93.

Biofiltration Guidance Manual for Drinking Water Facilities 163

Brown, J., G. Upadhyaya, J. Carter, T. Brown, and C. Lauderdale. 2016. North American Biofiltration Knowledge Base (No. 4459). Water Research Foundation, Denver, Colorado. Brown, J., R.S. Summers, E. Dickenson, and G. Upadhyaya. 2018. Biological Removal of Hazardous Pollutants in Drinking Water Treatment, in: Cecen, F., Tezal, U. (Eds.), Hazardous Pollutants in Biological Treatment Systems: Fundamentals and a Guide to Experimental Research. Volume 1. IWA Publishing, London, UK, pp. 313‐358. Camper, A., M.W. LeChevallier, S.C. Broadaway, and G.A. McFeters. 1985. Growth and persistence of pathogens on granular activated carbon filters. Appl. Environ. Microbiol. 50, 1378‐1382. Carlson, K.H. and G.L. Amy. 1998. BOM removal during biofiltration. Journal / American Water Works Association, 90(12), 42‐52. Cecen, F. and Ö.Aktas. 2011. Activated Carbon for Water and Wastewater Treatment: Integration of Adsorption and Biological Treatment. John Wiley & Sons, Weinheim, Germany. Cervi, S. 1996. Formation and Removal of Aldehydes as Ozonation By‐Products in a Pilot Scale Water Treatment Plant (Master’s Thesis). University of Windsor, Windsor, Ontario, Canada. Chaiket, T., P.C. Singer, A. Miles, M. Moran, and C. Pallotta. 2002. Effectiveness of coagulation, ozonation, and biofiltration in controlling DBPs. J. ‐ Am. Water Works Assoc. 94, 81‐95. https://doi.org/10.1002/j.1551‐8833.2002.tb10251.x. Chang, D.S. and P.C. Singer. 1991. The Impact of Ozonation on Particle Stability and the Removal of TOC and THM Precursors. Awwa, 83((3)), 71‐79. Chaudhary, D.S., S. Vigneswaran, H.‐H. Ngo, W.G. Shim, and H. Moon. 2003. Biofilter in Water and Wastewater Treatment. Korean J of Chem. Eng. 20, 1054. Chouari, R., W. Dardouri, F. Sallami, M.B. Rais, D. Le Paslier, and A. Sghir. 2014. Microbial Analysis and Efficiency of Biofiltration Packing Systems for Hydrogen Sulfide Removal from Wastewater Off Gas. Environ. Eng. Sci. 32, 121‐128. https://doi.org/10.1089/ees.2014.0290. Chowdhury, Z., A. Traviglia, J. Carter, T. Brown, R.S. Summers, C. Corwin, T. Zearley, M. Thurman, I. Ferrara, J. Olson, R. Thacker, and P. Barron. 2010. Cost Effective Regulatory Compliance with GAC Biofilters. Water Research Foundation, Denver, Colorado. Collins, L. M., J.J. Dziak, K.C. Kugler, and J.B. Trail. 2014. Factorial experiments: Efficient tools for evaluation of intervention components. American Journal of Preventive Medicine, 47, 498‐504. Crittenden, J.C., Montgomery Watson Harza (firm). 2012. MWH’s Water Treatment: Principles and Design, 3rd Edition, 3rd ed. John Wiley & Sons, Hoboken, New Jersey. Culp, R.L. and R.M. Clark. 1983. Granular Activated Carbon Installations. J‐AWWA 75, 398‐405. de Vera, G.A., C. Lauderdale, C.L. Alito, J. Hooper, and E.C. Wert. 2019. Using upstream oxidants to minimize surface biofouling and improve hydraulic performance in GAC biofilters. Water Res. 148, 526‐ 534. https://doi.org/10.1016/j.watres.2018.10.085. de Vera, G.A., D. Gerrity, M. Stoker, W. Frehner, and E.C. Wert. 2018. Impact of Upstream Chlorination on Filter Performance and Microbial Community Structure of GAC and Anthracite Biofilters. Environ. Sci.: Water Res. Technol. 4, 1133‐1144. Dickenson, E., E. Wert, E. Verdugo, K. Greenstein, J. Brown, G. Upadhyaya, P.G. Pope, O. Schneider, R. Marfil‐Vega, and R.S. Summers. 2018. Simultaneous Removal of Multiple Chemical Contaminants Using Biofiltration. Project #4559. Denver, Colorado: Water Research Foundation.

164 The Water Research Foundation

DuBois, M., K.A. Gilles, J.K. Hamilton, P.A. Rebers, and F. Smith. 1956. Colorimetric Method for Determination of Sugars and Related Substances. Anal. Chem. 28, 350‐356. https://doi.org/10.1021/ac60111a017. Edzwald, J.K. 1993. Coagulation in drinking water treatment: Particles, organics and coagulants. Water Science and Technology, 27(11), 21‐35. https://doi.org/10.2166/wst.1993.0261. Edzwald, J.K. 2011. Waterd Quality an Treatment: A Handbook on Drinking Water, 6th Edition. AWWA, Denver, Colorado. Eighmy, T.T., M.R. Collins, J.P. Malley Jr., J. Royce, and D. Morgan. 1997. Biologically Enhanced Slow Sand Filtration for Removal of Natural Organic Matter (No. 409). AWWA Research Foundation, Denver, Colorado. Elhadi, S., P.M. Huck, and R.M. Slawson. 2006. Factors affecting the removal of geosmin and MIB in drinking water biofilters. J‐AWWA 98, 108‐119. Emelko, M.B., P.M. Huck, B.M. Coffey, and E.F. Smith. 2006. Effects of Media, Backwash, and Temperature on Full‐Scale Biological Filtration. J. ‐ Am. Water Works Assoc. 98, 61‐73. https://doi.org/10.1002/j.1551‐8833.2006.tb07824.x Environmental Protection Agency (EPA). 1996. Method 8315A ‐ Determination of Carbonyl Compounds By High Performance Liquid Chromatography (HPLC). U.S. Environmental Protection Agency. Environmental Protection Agency (EPA). 2006. Long Term 2 Enhanced Surface Water Treatment Rule: A Quick Reference Guide, EPA 816‐F‐06‐008. Environmental Protection Agency (EPA). 2012. EPA Drinking Water Guidance on Disinfection By‐ Products, Advice Note No. 4. EPA Office of Environmental Enforcement. Escobar, I.C. and A.A. Randall. 2001. Assimilable Organic Carbon (AOC) and Biodegradable Organic Carbon (BDOC): Complementary Measurements. Water Research. 35(18): 3971‐3979. Evans, A., J. Carter, W. Mitch, R. Hozalski, P. Huck, and C. Russell. 2019. Biological Filtration: NDMA Control or Source of Precursors?. Project #4669. Denver, Colo.: Water Research Foundation. Evans, A., K. Onesios‐Barry, K. Casteloes, J. Carter, C. Ha, L. Hakes, A. Stoddart, M. Earle, L. Anderson, N. Allward, G. Gagnon, and W. Knocke. 2020 (unpublished). Optimizing Biofiltration for Improved Manganese Control under Cold‐ Water Conditions. Project #4749. Denver, Colorado: Water Research Foundation. Evans, P.J., D.H. Metz, R.D. Kashinkunti, Y. Hong, J. Springer, J.L. Hooper, and P.M. Richard. 2016. Chemically Enhanced Biological Filtration to Enhance Water Quality and Minimize Costs: Phase II Pilot‐ Scale Evaluation (No. 4429b). Water Research Foundation, Denver, Colorado. Evans, P.J., E.M. Opitz, P.A. Daniel, and C.R. Schulz. 2010. Biological Drinking Water Treatment Perceptions and Actual Experiences in North America (No. 4129). Water Research Foundation, Denver, Colorado. Evans, P.J., J.L. Smith, M.W. LeChevallier, O.D. Schneider, L.A. Weinrich, and P.K. Jjemba. 2013a. A Monitoring and Control Toolbox for Biological Filtration (No. 4231b). Water Research Foundation, Denver, Colorado. Evans, P.J., J.L. Smith, M.W. LeChevallier, O.D. Schneider, L.A. Weinrich, and P.K. Jjemba. 2013b. Biological Filtration Monitoring and Control Toolbox: Guidance Manual (No. 4321). Water Research Foundation, Denver, Colorado.

Biofiltration Guidance Manual for Drinking Water Facilities 165

Falkinham, J.O., A. Pruden, and M. Edwards. 2015. Opportunistic Premise Plumbing Pathogens: Increasingly Important Pathogens in Drinking Water. Pathogens 4, 373‐386. https://doi.org/10.3390/pathogens4020373. Farré, M.J., J. Reungoat, F.X. Argaud, M. Rattier, J. Keller, and W. Gernjak. 2011. Fate of N‐ nitrosodimethylamine, trihalomethane and haloacetic acid precursors in tertiary treatment including biofiltration. Water Res. 45, 5695‐5704. https://doi.org/10.1016/j.watres.2011.08.033 Findlay, R.H., G.M. King, and L. Watling. 1989. Efficacy of Phospholipid Analysis in Determining Microbial Biomass in Sediments. Appl. Environ. Microbiol. 55, 2888‐2893. Flemming, H.‐C. and J. Wingender. 2010. The biofilm matrix. Nat. Rev. Microbiol. 8, 623‐633. https://doi.org/10.1038/nrmicro2415 Fonseca, A.C., R.S. Summers, and M.T. Hernandez. 2001. Comparative measurements of microbial activity in drinking water biofilters. Water Res. 35, 3817‐3824. https://doi.org/10.1016/S0043‐ 1354(01)00104‐X. Frías, J., Ribas, F., and Lucena, F. 1992. A method for the measurement of biodegradable organic carbon in waters. Water Res. 26, 255. Funk, D., J. Hooper, M. Noibi, J. Goldman, R. Oliva, C. Schulz, K. Bell, D. Castaneda, C. Huang, and E. Machek. 2018. Ozone Biofiltration Direct Potable Reuse Testing at Gwinnett County. Water Research Foundation, Denver, Colorado. Gimbel, R., N. Graham, and M.R. Collins. 2006. Recent Progress in Slow Sand and Alternative Biofiltration Processes. IWA Publishing, London, UK. Glaze, W.H., M. Koga, and D. Cancilla. 1989a. Ozonation byproducts. 2. Improvement of an aqueous‐ phase derivatization method for the detection of formaldehyde and other carbonyl compounds formed by the ozonation of drinking water. Environ. Sci. Technol. 23, 838‐847. https://doi.org/10.1021/es00065a013. Glaze, W.H., M. Koga, E.C. Ruth, and D. Cancilla. 1989b. Application of Closed Loop Stripping and XAD Resin Adsorption for the Determination of Ozone By‐Products from Natural Water, in: Biohazards of Drinking Water Treatment. Lewis Publishers, Ine., pp. 201‐212. Greenstein, K.E., J. Lew, E.R.V. Dickenson, and E.C. Wert. 2018. Investigation of biotransformation, sorption, and desorption of multiple chemical contaminants in pilot‐scale drinking water biofilters. Chemosphere 200, 248‐256. Haddix, P.L., N.J. Shaw, and M.W. LeChevallier. 2004. Characterization of bioluminescent derivatives of assimilable organic carbon test bacteria. Appl. Environ. Microbiol. 70, 850‐854. Hammes, F., M. Berney, Y. Wang, M. Vital, O. Köster, and T. Egli. 2008. Flow‐cytometric total bacterial cell counts as a descriptive microbiological parameter for drinking water treatment processes. Water Res. 42, 269‐277. https://doi.org/10.1016/j.watres.2007.07.009 Hooper, J., K. Vickstrom, P. Evans, C. Alito, S. Black, and C. Lauderdale. 2019. Guidance Manual for Monitoring Biological Filtration of Drinking Water (No. 4620). Water Research Foundation, Denver, Colorado. Hozalski, R.M., E.J. Bouwer, and S. Goel. 1999. Removal of natural organic matter (NOM) from drinking water supplies by ozone‐biofiltration. Wat. Sci. & Technol. 40, 157‐163.

166 The Water Research Foundation

Hozalski, R.M., S. Goel, and E.J. Bower. 1995. TOC removal in biological filters. J. ‐ Am. Water Works Assoc. 87, 40‐54. Huck, P.M. and M.M. Sozanski. 2008. Biological Filtration for Membrane Pre‐Treatment and Other Applications: Towards the Development of a Practically Oriented Performance Parameter. Journal of Water Supply: Research and Technology ‐ Aqua. 57.4, 203‐224. Huck, P.M., B.M. Coffey, A. Amirtharajah, and E.J. Bower. 2000. Optimizing Filtration in Biological Filters (No. 252). AWWA Research Foundation, Denver, Colorado. Jekel, M.R. 1986. The stabilization of dispersed mineral particles by adsorption of humic substances. Water Research, 20(12), 1543‐1554. https://doi.org/10.1016/0043‐1354(86)90119‐3 Joret, J. C. and Y. Levi. 1986. Méthode Rapide d’Evaluation Du Carbone Eliminable Des Eaux Par Voie Biologique. Trib. Cebedeau. 510(39): 3‐9. Joret, J.C., Y. Levi, and C. Volk. 1991. Biodegradable Dissolved Organic Carbon (BDOC) Content of Drinking Water and Potential Regrowth of Bacteria. Water Science Technology 24(2):95‐101. Juhna, T. dan E. Melin. 2006. Ozonation and Biofiltration in Water Treatment, Operational Status and Optimization Issues. D.5.3.1.B. December 2006. The Netherlands: Techneau. Kawamura, S. 2000. Integrated Design and Operation of Water Treatment Facilities, 2nd ed. John Wiley & Sons, New York, New York. Keithley, S.E. and M.J. Kirisits. 2018. An improved protocol for extracting extracellular polymeric substances from granular filter media. Water Res. 129, 419‐427. https://doi.org/10.1016/j.watres.2017.11.020. Keithley, S.E. and M.J. Kirisits. 2019. Enzyme‐Identified Phosphorus Limitation Linked to More Rapid Headloss Accumulation in Drinking Water Biofilters. Environ. Sci. Technol. 53, 2027‐2035. https://doi.org/10.1021/acs.est.8b04573. Khammar, N., L. Malhautier, V. Degrange, R. Lensi, J.‐J. Godon, and J.‐L. Fanlo. 2005. Link between spatial structure of microbial communities and degradation of a complex mixture of volatile organic compounds in peat biofilters. J. Appl. Microbiol. 98, 476‐490. https://doi.org/10.1111/j.1365‐ 2672.2004.02474.x Kingsbury, R. 2010. Evaluation of MIEX Pre‐Treatment on Ozonation Performance and Disinfection By‐ Product Formation (Master’s Thesis). University of North Carolina, Chapel Hill. Kohl, P.M. and D. Dixon. 2012. Occurrence, impacts, and removal of manganese in biofiltration processes (No. 4021). Water Research Foundation, Denver, Colorado. Kohl, P.M. and S.J. Medlar. 2006. Occurrence of Manganese in Drinking Water and Manganese Control (No. 91147). AWWA Research Foundation, Denver, Colorado. Kotlarz, N., N. Rockey, T.M. Olson, S.‐J. Haig, L. Sanford, J. LiPuma, and L. Raskin. 2018. Biofilms in Full‐ Scale Drinking Water Ozone Contactors Contribute Viable Bacteria to Ozonated Water. Environ. Sci. Technol. 52, 2618‐2628. https://doi.org/10.1021/acs.est.7b04212. Krasner, S.W., M.J. Sclimenti, and B.M. Coffey. 1993. Testing biologically active filters for removing aldehydes formed during ozonation. Journal / American Water Works Association, 85(5), 62‐71. Laspidou, C. and B. Rittmann. 2002. A unified theory for extracellular polymeric substances, soluble microbial products, and active and inert biomass. Water Res. 36, 2711‐2720. https://doi.org/10.1016/S0043‐1354(01)00413‐4.

Biofiltration Guidance Manual for Drinking Water Facilities 167

Lauderdale, C., C. Alito, J. Hooper, K. Dowdell, and E. Wert. 2018. Optimizing Biofiltration for Various Source Water Quality Conditions (No. 4555). Water Research Foundation, Denver, Colorado. Lauderdale, C., G. Pope, K. Scheitlin, J. Zwerneman, M.J. Kirisits, and S.W. Bae. 2016. Optimizing Filter Conditions for Improved Manganese Control During Conversion to Biofiltration (No. 4448). Water Research Foundation, Denver, Colorado. Lauderdale, C., J. Brown, P. Chadik, and M. Kirisits. 2011. Engineered Biofiltration for Enhanced Hydraulic and Water Treatment Performance (No. 4215). Water Research Foundation, Denver, Colorado. Lauderdale, C., P. Chadik, M. Kirisits, and J. Brown. 2012. Engineered biofiltration: Enhanced biofilter performance through dnutrient an peroxide addition. J‐AWWA 104, E298‐E309. Lauderdale, C.V., K. Scheitlin, J. Nyfennegger, G. Upadhyaya, J. Brown, L. Raskin, T.‐H. Chiao, and A. Pinto. 2014. Optimizing Engineered Biofiltration (No. 4346). Water Research Foundation, Denver, Colorado. Lavrakas, P. 2008. Encyclopedia of survey research methods Thousand Oaks, CA: Sage Publications, Inc. doi: 10.4135/9781412963947. Le Bihan, Y. and P. Lessard. 2000. Monitoring biofilter clogging: biochemical characteristics of the biomass. Water Res. 34, 4284‐4294. https://doi.org/10.1016/S0043‐1354(00)00202‐5 LeChevallier, M.W., O.D. Schneider, L.A. Weinrich, P.K. Jjemba, P.J. Evans, and J.L. Hooper. 2015. An Operational Definition of Biostability for Drinking Water (No. 4312). Water Research Foundation, Denver, Colorado. LeChevallier, M.W., W. Schulz, and R.G. Lee. 1991. Bacterial nutrients in drinking water. Appl. Environ. Microbiol. 57, 857‐862. Li, M., C. Feng, Z. Zhang, S. Yang, and N. Sugiura. 2010. Treatment of nitrate contaminated water using an electrochemical method. Bioresour. Technol. 101, 6553‐6557. https://doi.org/10.1016/j.biortech.2010.03.076 Li, Q., S. Yu, L. Li, G. Liu, Z. Gu, M. Liu, Z. Liu, Y. Ye, Q. Xia, and L. Ren. 2017. Microbial Communities Shaped by Treatment Processes in a Drinking Water Treatment Plant and Their Contribution and Threat to Drinking Water Safety. Front. Microbiol. 8. https://doi.org/10.3389/fmicb.2017.02465 Lim, M.H., S.A. Snyder, and D.L. Sedlak. 2008. Use of Biodegradable Dissolved Organic Carbon (BDOC) to assess the potential for transformation of wastewater‐derived contaminants in surface waters. Water Res. 42, 2943. Linder, K. and B. Martin. 2015. Self‐Assessment for Water Treatment Plant Optimization: Partnership for Safe Water. American Water Works Association, Denver, CO. Liu, X., P.M. Huck, and R.M Slawson. 2001. Factors affecting Drinking Water Biofiltration. American Water Works Association, 93(12), 90‐101. https://doi.org/10.1016/j.cis.2010.06.007 Logsdon, G.S., A.F. Hess, M.J. Chipps, and A.J. Rachwal. 2002. Filter Maintenance and Operations Guidance Manual (No. 2511). AWWA Research Foundation, Denver, Colorado. Lowry, O.H., N.J. Rosebrough, A.L. Farr, and R.J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193, 265‐275. Ma, J., G.B. Li, Z.L. Chen, G.R. Xu, and G.Q. Cai. 2001. Enhanced coagulation of surface waters with high organic content by permanganate preoxidation. Water Science and Technology: Water Supply, 1(1), 51‐ 61.

168 The Water Research Foundation

Magic‐Knezev, A. and D. van der Kooij. 2004. Optimisation and significance of ATP analysis for measuring active biomass in granular activated carbon filters used in water treatment. Water Res. 38, 3971‐3979. https://doi.org/10.1016/j.watres.2004.06.017. Marda, S, D.I. Kim, J. Myung, J. Gordy, S. Pierpont, J. Gianatasio, A. Amirtharajah, and J.H., Kim. 2008. Plant conversion experience: Ozone BAC process installation and disinfectant residual control. J. Am. Water Works Assoc. 100(4), 117‐128. Mauclaire, L., A. Schürmann, M. Thullner, J. Zeyer, and S. Gammeter. 2004. Sand filtration in a water treatment plant: biological parameters responsible for clogging. J. Water Supply Res. Technol.‐Aqua 53, 93‐108. https://doi.org/10.2166/aqua.2004.0009. McKie, M.J., C. Bertoia, L. Taylor‐Edmonds, S.A. Andrews, and R.C. Andrews. 2019. Pilot‐scale comparison of cyclically and continuously operated drinking water biofilters: Evaluation of biomass, biological activity and treated water quality. Water Res. 149, 488‐495. https://doi.org/10.1016/j.watres.2018.11.033. McKie, M.J., L. Taylor‐Edmonds, S.A. Andrews, and R.C. Andrews. 2015. Engineered biofiltration for the removal of disinfection by‐product precursors and genotoxicity. Water Res. 81, 196‐207. https://doi.org/10.1016/j.watres.2015.05.034. McKie, M.J., M.C. Ziv‐El, L. Taylor‐Edmonds, R.C. Andrews, and M.J. Kirisits. 2019. Biofilter scaling procedures for organics removal: A potential alternative to piloting. Water Res. 151, 87‐97. https://doi.org/10.1016/j.watres.2018.12.006. Metz, D.H., R.D. Kashinkunti, Y. Hong, J. Springer, P.J. Evans, J. Hooper, and P. Richard. 2016. Chemically Enhanced Biological Filtration to Enhance Water Quality and Minimize Costs: Phase II Pilot‐Scale Evaluation (No. 4429b). Water Research Foundation, Denver, Colorado. Mogren, E.M., P. Scarpino, and R.S. Summers. 1990. Measurement of biodegradable dissolved organic carbon in drinking water. Cincinnati, OH: Proc. 1990 AWWA ACE. Moll, D.M., R.S. Summers, and A. Breen. 1998. Microbial Characterization of Biological Filters Used for Drinking Water Treatment. Appl. Environ. Microbiol. 64, 2755‐2759. Najm, I.N. and S.W. Krasner. 1995. “Effects of bromide and NOM on by‐product formation.” J. ‐ Am. Water Works Assoc. 87, 106‐115. https://doi.org/10.1002/j.1551‐8833.1995.tb06305.x Nieminski, E. and S. Perry. 2015. State and provincial perspectives on the regulation of biological filtration. J‐AWWA. 107 (12). 31‐38. Nix, D. and J. Taylor. 2018. Filter Evaluation Procedures for Granular Media, 2nd ed. AWWA, Denver, Colorado. Nugroho, W.A and Reungoat, J. 2010. The Performance of BAC in Removing Pharmaceuticals in Drinking Water Treatment. Journal of Applied Sciences in Environmental Sanitation, 5(2), 131‐141. Nybroe, O., P.E. Jørgensen, and M. Henze. 1992. Enzyme activities in wastewater and activated sludge. Water Res. 26, 579‐584. https://doi.org/10.1016/0043‐1354(92)90230‐2. Nyfennegger, J., S. Burns, N. Ross, G. Upadhyaya, J. Brown, C.V. Lauderdale, M.J. Kirisits, S.E. Keithley, and S. Bae. 2016. Full‐Scale Engineered Biofiltration Evaluation and Development of a Performance Tracking Tool (No. 4525). Water Research Foundation, Denver, Colorado.

Biofiltration Guidance Manual for Drinking Water Facilities 169

Oh, S., F. Hammes, and W.‐T. Liu. 2018. Metagenomic characterization of biofilter microbial communities in a full‐scale drinking water treatment plant. Water Res. 128, 278‐285. https://doi.org/10.1016/j.watres.2017.10.054 Peldszus, S., P.M. Huck, and S.A. Andrews. 1996a. Determination of short‐chain aliphatic, oxo‐ and hydroxy‐acids in drinking water at low microgram per liter concentrations. J. Chromatogr. A 723, 27‐34. Peldszus, S., P.M. Huck, and S.A. Andrews. 1996b. Determination of Carboxylic Acids in Drinking Water at Low µg/L Concentrations: Method Development and Application. Presented at the 24th Annual AWWA Water Quality Technology Conference, AWWA, Denver, Colorado. Persson, F., G. Heinicke, T. Hedberg, M. Hermansson, and W. Uhl. 2007. Removal of Geosmin and MIB by Biofiltration ‐ an Investigation Discriminating Between Adsorption and Biodegradation. Environ. Technol. 28, 95‐104. Pharand, L., M.I. Van Dyke, W.B. Anderson, and P.M. Huck. 2014. Assessment of biomass in drinking water biofilters by adenosine triphosphate. J. Am. Water Works Assoc. 106, E433‐E444. Pinto, A., C. Xi, and L. Raskin. 2012. Bacterial Community Structure in the Drinking Water Microbiome Is Governed by Filtration Processes. Environ. Sci. Technol. 46, 8851‐8859. https://doi.org/10.1021/es302042t Rittman B.E., Tang, Y., Meyer, K., Nerenberg, R., and Brown, J. 2020. Biological Processes: Chapter 17 of American Water Works Association's/American Society of Civil Engineer's Book Water Treatment Plant Design, 6th Edition. In press. Rittmann, B. and D. Stillwell. 2002b. Modeling biological treatment: the integrated biofiltration model. Journal of Water Supply: Research and Technology ‐ Aqua. 51.1, 1‐14. Rittmann, B. and P. McCarty. 2001. Environmental Biotechnology: Principles and Applications. McGraw‐ Hill, Boston, MA. Rittmann, B., D. Stillwell, and A. Ohashi. 2002a. The transient‐state, multiple‐species, biofilm model for biofiltration processes. Water Research. 36, 2342‐2356. Schulz, C.R. 2014. Designing Integrated Ozone‐Biofiltration Treatment Systems. Ozone Sci. Eng. 36, 276‐ 286. https://doi.org/10.1080/01919512.2014.904741 Selbes, M., J. Amburgey, C. Peeler, A. Alansari, and T. Karanfil. 2016. Evaluation of Seasonal Performance of Conventional and Phosphate‐Amended Biofilters. J. ‐ Am. Water Works Assoc. 108, E525‐E532. Seredyńska‐Sobecka, B., M. Tomaszewska, M. Janus, and A.W. Morawski. 2006. Biological activation of carbon filters. Water Res. 40, 355‐363. https://doi.org/10.1016/j.watres.2005.11.014 Servais, P., Billen, G., and Hascoet, M.‐C. 1987. Determination of the biodegradable fraction of dissolved organic matter in waters. Water Res. 21, 445. Servais, P., G. Billen, and P. Bouillot. 1994. Biological Colonization of Granular Activated Carbon Filters in Drinking‐Water Treatment. J. Environ. Eng. 120, 888‐899. https://doi.org/10.1061/(ASCE)0733‐ 9372(1994)120:4(888) Sharma, D., L. Taylor‐Edmonds, and R.C. Andrews. 2018. Comparative assessment of ceramic media for drinking water biofiltration. Water Res. 128, 1‐9. https://doi.org/10.1016/j.watres.2017.10.019 Simpson, D.R., 2008. Biofilm processes in biologically active carbon water purification. Water Res. 42, 2839‐2848.

170 The Water Research Foundation

Snyder, S.A., E.C. Wert, H. Lei, P. Westerhoff, and Y. Yoon. 2007. Removal of EDCs and Pharmaceuticals in Drinking and Reuse Treatment Processes. AWWA Research Foundation, Denver, Colorado. Sontheimer, H., J.C. Crittenden, and R.S. Summers. 1988. Activated carbon for water treatment, 2nd ed. DVGW‐Forschungsstelle, Karlsruhe, FRG. Stanford, B., E.R.V. Dickenson, E. Wert, and M. Inyang. 2017. Controlling Trace Organic Contaminants Using Alternative, Non‐FAT Technology for Indirect Potable Water Reuse | WateReuse (No. 13‐10). Water Environment & Reuse Foundation, Alexandria, Virginia. Steele, J.A., F. Ozis, J.A. Fuhrman, and J.S. Devinny. 2005. Structure of microbial communities in ethanol biofilters. Chem. Eng. J., Selected papers from the USC‐CSC‐TRG Biofiltration Conference, October 2004 113, 135‐143. https://doi.org/10.1016/j.cej.2005.04.011 Terry, L.G. and R.S. Summers. 2018. Biodegradable organic matter and rapid‐rate biofilter performance: A review. Water Research, 128, 234‐245. https://doi.org/10.1016/j.watres.2017.09.048 Terry, L.G., P. Pruisner, E. Peterson, E. Dickenson, J. Wang, and R.S. Summers. 2019. Scale‐Up methodology for biological filtration removal of dissolved organic matter. Environmental Engineering Science, 36(4), 405‐412. https://doi.org/10.1089/ees.2018.0316 Terry, L.G., P. Pruisner, E. Peterson, E. Dickenson, J. Wang, and R.S. Summers. 2018. Scale‐Up Methodology for Biological Filtration Removal of Dissolved Organic Matter. Environmental Engineering Science, 36, 4, 405‐412. Texas A&M Engineering Extension Service (TEEX), 2013. Surface Water Production I. J.C. Brown, A. Evans, J. Carter, C. Lauderdale, and O. Schneider. 2017. Biofilter Conversion Guidance Manual. Project #4496. Denver, CO.: Water Research Foundation. Tobiason, J.E., W.R. Knocke, J. Goodwill, P. Hargette, R. Bouchard, and L. Zuravnsky. 2008. Characterization and Performance of Filter Media for Manganese Control. Tremblay, C.V., A. Beaubien, P. Charles, and J.A. Nicell. 1998. Control of biological iron removal from drinking water using oxidation‐reduction potential. Water Sci. Technol. 38, 121‐128. https://doi.org/10.2166/wst.1998.0244 Uhl, W. 2008. Biofiltration Processes for Organic Matter Removal, in: Rehm, H.J., Reed, G., Puhler, A., Stadler, P. (Eds.), Biotechnology: Environmental Processes. John Wiley & Sons, Weinheim, Germany, pp. 457‐478. Urfer, D. and P.M. Huck. 2001. Measurement of biomass activity in drinking water biofilters using a respirometric method. Water Res. 35, 1469‐1477. https://doi.org/10.1016/S0043‐1354(00)00405‐X Vaccari, D.A., P.F. Strom, and J.E. Alleman. 2006. Quantifying Microorganisms and their Activity, in: Environmental Biology for Engineers and Scientists. John Wiley & Sons, New York, New York, pp. 290‐ 341. https://doi.org/10.1002/0471741795.fmatter Velten, S., F. Hammes, M. Boller, and T. Egli. 2007. Rapid and direct estimation of active biomass on granular activated carbon through adenosine tri‐phosphate (ATP) determination. Water Res. 41, 1973‐ 1983. https://doi.org/10.1016/j.watres.2007.01.021 Velten, S., M. Boller, O. Koster, J. Helbing, H. Weilenmann, and F. Hammes. 2011. Development of Biomass in a Drinking Water Granular Active Carbon (GAC) Filter. Water Research. 45(19): 6347‐6354. Volk, C., C. Renner, C. Robert, and J.C. Joret. 1994. Comparison of two techniques for measuring biodegradable dissolved organic carbon in water. Environ. Technol. 15, 545‐556.

Biofiltration Guidance Manual for Drinking Water Facilities 171

Wang, H., L. Ho, D.M. Lewis, J.D. Brookes, and G. Newcombe. 2007. Discriminating and assessing adsorption and biodegradation removal mechanisms during granular activated carbon filtration of microcystin toxins. Water Research, 41(18), 4262‐4270. https://doi.org/10.1016/j.watres.2007.05.057. Water Research Foundation, 2019. Biofiltration: Defining Benefits and Developing Utility Guidance [WWW Document]. URL http://www.waterrf.org/the‐foundation/research‐programs/focus‐area‐ program/Pages/biofiltration.aspx (accessed 4.30.19). Weinberg, H.S., W.H. Glaze, S.W. Krasner, and M.J. Sclimenti. 1993. Formation and Removal of Aldehydes in Plants That Use Ozonation. J. ‐ Am. Water Works Assoc. 85, 72‐85. https://doi.org/10.1002/j.1551‐8833.1993.tb05988.x Weinrich, L.A., E. Giraldo, and M.W. LeChevallier. 2009. Development and Application of a Bioluminescence‐Based Test for Assimilable Organic Carbon in Reclaimed Waters. Appl Env. Microbiol 75, 7385‐7390. https://doi.org/10.1128/AEM.01728‐09 Wert, E.C., J.J. Neemann, D.J. Rexing, and R.E. Zegers. 2008. Biofiltration for removal of BOM and residual ammonia following control of bromate formation. Water Res. 42, 372‐378. Westerhoff, P., R.S. Summers, Z. Chowdhury, and S. Kommineni. 2005. Ozone‐Enhanced Biofiltration for Geosmin and MIB Removal (No. 2775). Water Research Foundation, Denver, Colorado. Wingender, J., T.R. Neu, and H.‐C. Flemming. 1999. What are Bacterial Extracellular Polymeric Substances?, in: Wingender, J., Neu, T.R., Flemming, H.‐C. (Eds.), Microbial Extracellular Polymeric Substances: Characterization, Structure and Function. Springer Berlin Heidelberg, Berlin, Heidelberg, pp. 1‐19. https://doi.org/10.1007/978‐3‐642‐60147‐7_1. Wolfe, T.A. and N.G. Pizzi. 1999. Optimizing Filter Performance. J. N. Engl. Water Works Assoc. 133, 6‐21. Xiang, H., X. Lu, L. Yin, F. Yang, G. Zhu, and W. Liu. 2013. Microbial community characterization, activity analysis and purifying efficiency in a biofilter process. J. Environ. Sci. China 25, 677‐687. Zanacic, E., D.W. McMartin, and J. Stavrinides. 2017. From source to filter: changes in bacterial community composition during potable water treatment. Can. J. Microbiol. 63, 546‐558. https://doi.org/10.1139/cjm‐2017‐0077 Zearley, T.L. and R.S. Summers. 2012. Removal of Trace Organic Micropollutants by Drinking Water Biological Filters. Environ. Sci. Technol. 46, 9412‐9419. Zhang, S., S.W. Gitungo, L. Axe, R.F. Raczko, and J.E. Dyksen. 2017. Biologically active filters – An advanced water treatment process for contaminants of emerging concern. Water Res. 114, 31‐41. Zhang, X. and P.L. Bishop. 2003. Biodegradability of biofilm extracellular polymeric substances. Chemosphere 50, 63‐69. Zhao, Y., L. Tayler‐Edmonds, and R.C. Andrews. 2019. Impact of carbon‐based nutrient enhancement on biofiltration performance for drinking water treatment ‐ ScienceDirect. J. Environ. Sci. 82, 124‐131. https://doi.org/10.1016/j.jes.2019.03.002. Zhu, I.X., T. Getting, and D. Bruce. 2010. Review of Biologically Active Filters in Drinking Water Applications. J‐AWWA 102, 67‐77.

172 The Water Research Foundation

advancing the science of water®

1199 North Fairfax Street, Suite 900 6666 West Quincy Avenue Alexandria, VA 22314-1445 Denver, CO 80235-3098

www.waterrf.org | [email protected]