                                                Pilot and Bench Study Costs for Drinking Water Rules 
                                                
                                                
                                                
                                                BPA EP-BPA-15-C-0002, Task Order 17
                                                
                                                
                                                
                                                
                                                
                                                
                                                
                                                
                                                DRAFT 
                                                December 2020
                                                
                                                                               
                                                
                                                
                                                Prepared for:
                                                U.S. Environmental Protection Agency
                                                Office of Groundwater and Drinking Water
                                                
                                                
                                                
                                                Submitted by:
                                                Abt Associates 
                                                6130 Executive Blvd.
                                                Rockville, MD 20852
                                       
Introduction
The Safe Drinking Water Act Amendments of 1996 and numerous other statutes and executive orders require the U.S. Environmental Protection Agency (EPA, or the Agency) to estimate regulatory compliance costs as part of its rulemaking process. EPA uses these estimates in social benefit-cost analyses, as well as various distributional analyses (such as those required by the Regulatory Flexibility Act and the Small Business Regulatory Enforcement Fairness Act). These types of analyses help the Agency make decisions regarding regulatory options for drinking water systems. Thus, improvements in the accuracy of compliance cost estimates enhance the Agency's policymaking capabilities.
--------------------------------------------------------------------------------
Components in EPA's WBS Cost Models
--------------------------------------------------------------------------------

--------------------------------------------------------------------------------
The WBS-based models include detailed analyses of the following types of factors influencing costs:
--------------------------------------------------------------------------------
- raw water quality
--------------------------------------------------------------------------------
- treated water quality standard
--------------------------------------------------------------------------------
- treatment volume
--------------------------------------------------------------------------------
- engineering parameters
--------------------------------------------------------------------------------
- treatment equipment requirements
--------------------------------------------------------------------------------
- treatment structural requirements
--------------------------------------------------------------------------------
- add-on costs (i.e., permits, land, and pilot studies)
--------------------------------------------------------------------------------
- indirect costs, such as
--------------------------------------------------------------------------------
	- geotechnical
--------------------------------------------------------------------------------
	- standby power
--------------------------------------------------------------------------------
- annual O&M requirements.
EPA has developed new models to estimate the cost of drinking water rules. These models are based on a work breakdown structure (WBS), or a structuring of work that defines and displays all of the major elements needed to accomplish the project objectives. EPA pursued the WBS approach as part of an effort to address recommendations regarding its methods for estimating drinking water compliance costs. 
One of the types of costs in the WBS-based cost models is the cost of conducting bench tests or pilot studies. These are small-scale studies of one or more aspects of a technology to determine whether it can remove contaminants to levels that meet drinking water standards. In particular, a demonstration of treatment effectiveness is required for technologies that are not designated best available technologies (BAT) in 40 CFR 141 or EPA guidance, or small system compliance technologies (SSCT) in EPA guidance or regulations. Even when a technology is a BAT or SSCT, drinking water facility operators may conduct bench and pilot studies to ensure that it is suitable for site-specific water quality conditions.
Bench and pilot studies also guide process design and operation. These studies can assist with determining appropriate loading rates, chemical feed rates, waste handling requirements, and other process parameters. Operating a pilot system can ultimately reduce capital and operating costs through optimal design and operation of a treatment process. Pilot system operation, however, can be costly. Therefore, it is important to consider cost trade-offs when deciding whether and how to operate a pilot system.
This paper provides a discussion of pilot study requirements and protocols for various technologies as well as estimated costs for the WBS design and cost models. Note that the cost estimates represent an average scenario and could vary based on regulatory or site-specific factors. Section 2 contains background information on types of studies and implementation options. Section 3 presents the method EPA used to incorporate the costs of bench and pilot studies in the suite of WBS technology models. Section 4 provides pilot study requirements and cost estimates by technology type. Section 5 provides a summary of cost estimates across technologies.

Background
Bench tests and pilot studies are conducted to achieve objectives including one or more of the following: 
Verify that the technology is suitable for site-specific conditions, including source water quality and variability
Reduce costs by gaining information to optimize facility and process design
Gain experience in operating a treatment technology before full-scale implementation
Ensure compliance with local, state, and federal regulations.
Different objectives can be met using different study approaches. For example, a pilot-scale study uses a scaled-down version of a full-scale treatment train that is operated for as little as a few weeks or as long as 12 months or more, with changing process parameters to determine the ranges of workable and optimal operating conditions. A bench-scale study is less resource-intensive, and may last just a few days to a few weeks. Bench studies for water treatment technologies include batch tests (e.g., jar tests) or continuous-flow tests. For instance, one type of continuous-flow test is the rapid small-scale column test (RSSCT). An RSSCT, used to test adsorptive media, uses a small laboratory column to simulate the performance of a full-scale adsorber (Ford et al., 2001). Although pilot testing is performed onsite because of the volume of source water required, RSSCTs and jar tests may be performed either off-site or onsite. Finally, in some situations, treatment plants may not need to perform bench or pilot studies; a "desktop" or "flat sheet" analysis (i.e., a review of data from other treatment facilities, universities, vendors, or other sources, and/or engineering computations) may be sufficient.
Before constructing a new treatment train or adding to an existing treatment train, a water treatment plant that must comply with a new regulation would perform one or more types of analyses: desktop analysis, bench tests, and/or pilot tests. The type of pre-construction study or studies depends on the specific new regulation, as well as the existing treatment technology, source water characteristics, source water variability, financial resources available to the water system, and local, state, and federal regulations. For example, a groundwater system that has source water with a low dissolved mineral content and that is considering adding packed tower aeration in order to remove VOCs usually would not need to construct a pilot aeration tower. This is because the performance of packed tower aeration systems can be predicted with design equations (NAS, 1997; Fitz, 2004). For adsorptive media systems, although a pilot-scale study provides more information to optimize operating parameters, bench-scale RSSCTs can be adequate for evaluating the performance of various media under site-specific conditions (Kini et al., 2011; Schers et al., 2011; Kohlmeier, 2011; Cleveland and Mount, 2011; Kwan et al., 2010; Chiu et al., 2009). Thus, systems with limited resources might perform only an RSSCT before constructing the full-scale treatment plant, whereas systems with more resources might perform a pilot study as well (NAS, 1997; Westerhoff, 2003). Bench-scale studies are also used to refine options that are then explored at greater depth through pilot studies. Bench tests also can be conducted to fine-tune the design of the pilot study (e.g., by testing coagulant doses and flocculation times), and can lower the cost of pilot studies. Section 4 addresses bench and pilot study requirements for different types of technologies.
Variations in treatment technology, source water, target contaminant, and other aspects affect not only the necessity of bench and pilot studies, but also study costs. The components of costs for a bench-scale or pilot-scale study include:
Equipment rental or purchase, and installation
Chemicals, filtration media, and other materials (e.g., membranes) used for treatment
Contaminant, if it is needed to "spike" the source water (i.e., artificially increase contaminant loading)
Electricity
Labor to design and implement the study and report the results
Sampling and analysis for target contaminant(s) and certain nontarget constituents.
The duration of a pilot study will affect operating costs such as labor and analysis expenses. Analytical costs will also vary depending on the target contaminant, source water characteristics, and the technology being tested.
A variety of resources are available to help water systems choose appropriate technologies to comply with new federal drinking water regulations. These resources help to reduce the cost of pre-construction studies, especially for small systems. EPA issues a list of BATs that meet certain effectiveness and feasibility requirements, especially for medium and large systems. EPA also issues a list of SSCTs that meet similar requirements for small systems. For rules that affect small systems (those serving up to 10,000 people), EPA publishes guidance documents to aid systems in implementing technologies. For example, the Arsenic Treatment Evaluation Handbook for Small Water Systems (U.S. EPA, 2003d) provides decision trees to help small systems select between various response options that include process optimization, source water blending, adsorptive media, ion exchange, and reverse osmosis. EPA's Environmental Technology Verification (ETV) program also helps identify appropriate treatment technologies, by establishing testing protocols and treatment goals and certifying that they have been met. Finally, the Technology Assistance Center Network, or TACNET, is an EPA-funded network of research and resource centers at eight universities around the United States that provides assistance to small systems, including technical assistance for pilot testing.
The estimated costs associated with pilot testing are difficult to standardize for a model or average system. Section 3 describes EPA's method for estimating the pilot study costs used in the WBS models. These costs will vary across rules as information requirements and the state of technological knowledge and experience advance. For example, a relatively new process that requires pilot testing to verify treatment effectiveness in most applications could eventually require less frequent or intensive studies, as experience with the technology in a range of operating conditions establishes greater understanding of performance among operators and regulators. Furthermore, costs will vary across rules because test protocols will depend on the contaminant(s) of concern. The pilot study costs reported in this paper reflect estimates that are not specific to any particular rule; cost estimates will vary for specific proposed regulations.
Method
The spreadsheet EPA developed to estimate bench test and pilot study costs divides the costs into three main categories: 
Materials (treatment equipment, chemicals, and media)
Lab analysis of water and waste samples
Labor time to implement the study, monitor performance, and analyze results. 
There will also be incremental waste disposal costs, but these are likely to be minor given the small flows involved in bench and pilot tests. Therefore, the spreadsheet does not explicitly include incremental waste disposal costs. If incremental disposal costs are expected to account for a large proportion of overall pilot study costs, they can be incorporated in material costs (e.g., media residuals that are classified as hazardous may require costly treatment or disposal). 
Rather than bundle all of these costs into a single estimate, EPA adapted the work breakdown structure that it is using to estimate treatment process capital and O&M costs to pilot studies. This approach improves the transparency of EPA's cost analysis. It also provides insights into pilot study cost drivers, which could inform EPA's efforts to improve the cost effectiveness of its regulatory measures. The following sections describe EPA's method for estimating each of the three cost elements.
Material Costs
Pilot systems can be custom-built or mobile, prefabricated units. They can be obtained from original equipment manufacturers (OEMs) such as membrane suppliers, or from companies that build and/or operate pilot systems. In either case, the systems are a small-scale version of a large-scale plant and, therefore, comprise a variety of equipment.
In the WBS technology models, equipment costs are developed at the component level. For example, the adsorptive media model contains engineering specifications and costs for over 40 components including pressure vessels, pumps, pipes, valves, instruments, structures, chemicals, and media. Although the same approach could be used to derive costs for pilot study equipment, this level of detail is generally not applicable for pilot studies because vendors provide prices for complete pilot systems that include all equipment and materials. Therefore, EPA's pilot study module incorporates equipment costs as lump sum estimates. The estimates vary by technology and by type of study. In particular, equipment costs for bench tests are substantially lower than pilot studies.
Equipment estimates can also vary by system size when scalability needs dictate different equipment requirements for large and small systems. More often, however, the pilot study cost module incorporates variations across system size in study objectives and circumstances. For example, pilot study objectives for large systems may include extensive study of system operation parameters to optimize performance and cost. Extensive operational studies would increase pilot study costs, but the incremental costs may be small compared to potential cost savings. In contrast, small systems may be more concerned with ensuring treatment effectiveness for their particular source water; further investigations into treatment optimization may not provide cost savings large enough to justify the additional expense. The method for estimating bench and pilot costs for adsorptive technologies demonstrates this approach. Costs for large-flow systems are based on bench tests (to identify which adsorptive media works best given site-specific water quality) and pilot studies (to collect operational data such as identifying the breakthrough curve). Costs for small-flow systems (treating less than 1 million gallons per day, or mgd) include only a bench test, because operating a pilot system long enough to collect data for the breakthrough curve might be too costly. 
In addition, small systems often have later compliance dates than large systems. As a result, EPA assumes that more information on technology performance is available by the time small systems make compliance decisions, which should reduce pilot study requirements and costs. As noted in Section 2, there are additional information sources such as the ETV program and EPA guidance documents that help small systems select effective compliance methods. 
For pilot study equipment costs, EPA obtained quotes from firms and organizations that provide engineering services or pilot systems for a variety of technologies as well as from OEM vendor marketing materials that contain pilot study cost estimates. Both data sources were necessary because many OEMs do not have a price list for standard pilot systems. Instead, they customize the pilot system based on characteristics such as design flow, influent water quality, and effluent standards. EPA updated many of the quotes during development of this draft of the report. Where updated cost data were not available, EPA escalated the original prices to 2019 dollars using the Consumer Price Index.  
Analytical Costs
During a bench or pilot study, samples of influent and effluent water are analyzed for target contaminant concentrations and other ("non-target") water parameters. In addition, waste streams can also be monitored. These analytical results provide information to evaluate treatment effectiveness or identify optimal operation procedures and parameters. Comparing the concentration of the target parameter(s) in influent and effluent demonstrates treatment performance. Analysis of non-target parameters can provide insights for optimizing treatment, and can also raise "red flags" of potential negative impacts on important water quality parameters, such as alkalinity levels. Surface water systems, in particular, may need to determine how treatment practices should respond to seasonal changes in source water quality. 
Analysis costs include the time spent obtaining and preparing the samples as well as the costs of laboratory analysis. This section addresses lab analysis; Section 3.3 discusses costs for obtaining and preparing samples. Analytical costs in the pilot study module depend on the number of samples and per-sample analysis costs. The costs used are average per-sample analysis costs obtained from several certified labs. These costs reflect variations in analytical methods across parameters. EPA updated most of these costs during development of this draft of the report. Where updated cost data were not available, EPA escalated the original prices to 2019 dollars using the Bureau of Labor Statistics Employment Cost Index, based on the assumption that labor is the most significant factor in laboratory analysis costs. Exhibit 1 identifies commonly tested water parameters. 
Exhibit 1. Commonly Tested Non-Target Parameters 
 Algae (number and speciation)
 Alkalinity
 Aluminum (total)
 Aluminum (dissolved)
 Assimilable Organic Carbon (AOC)
 Barium
 Biochemical Oxygen Demand (BOD)
 Bromate
 Bromide
 Calcium
 Carbonate
 Chloride
 Color
 Copper
 Dissolved Organic Carbon (DOC)
 Dissolved Oxygen (DO)[1]
 Fluoride
 Haloacetic acids (HAAs)
 Total Hardness
 Heterotrophic Plate Count (HPC)
 Iron (total)
 Iron (dissolved)
 Lead
 Magnesium
 Manganese (total)
 Manganese (dissolved)
 Nitrate
 Particle Count
 pH[1]
 Phosphate
 Silica (Silicon Dioxide)
 Sulfate
 Total Sulfides
 Taste and Odor Compounds
 Toxicity Characteristics Leaching Procedure (TCLP)
 Total Dissolved Solids (TDS) [1]
 Temperature[1]
 Total Organic Carbon (TOC)
 Total Coliform
 Total Suspended Solids (TSS)
 Total Trihalomethanes (TTHMs)
 Turbidity
 UV-254 absorbance
 Volatile Suspended Solids (VSS)
1. Can be monitored onsite using hand-held or in-line monitoring equipment. Cost of monitoring equipment is excluded from pilot study costs when this equipment is included in the relevant WBS engineering model to prevent double counting capital expenditures.

Excluded from this list of parameters are target contaminants. Tests for these contaminants typically use specific EPA analytical methods. EPA methods are procedures used to analyze samples to determine the identity and concentration of specific sample components (U.S. EPA 2003a). Many groups develop analytical methods, and those that are approved by EPA's Office of Groundwater and Drinking Water can be used to monitor drinking water compliance (U.S. EPA 2003a). Exhibit 2 shows EPA methods that might be used for select target contaminants. 
Exhibit 2. EPA Methods for Target Parameters
                                    Method
                                  Parameters
                                 100.1, 100.2
Asbestos
                                     200.8
Antimony, Arsenic, Barium, Beryllium, Cadmium, Chromium, Mercury, Nickel, Selenium, Thallium
                                     200.9
Antimony, Arsenic, Beryllium, Cadmium, Chromium, Nickel, Selenium, Thallium
                                     335.4
Cyanide
                                     353.2
Nitrate, Nitrite
                                     502.1
1,1,1-Trichloroethane, 1,1,2-Trichloroethane, 1,2-Dichlropropane, Carbon tetrachloride, o-Dichlorobenzene
                                     502.2
1,1-Dichlroethylene, 1,1,1-Trichlroethane, 1,1,2-Trichloroethane, 1,2-Dichlroethane, 1,2,4-Trichlorobenzene, 1,2-Dichloropropane, Benzene, Carbon tetrachloride, cis-1,2-Dichloroethylene, Dichloromethane, Ethylbenzene, o-Dichlorobenzene, Styrene, Toluene, Total Xylene, trans-1,2-Dichloroethylene, Trichloroethylene, Vinyl chloride
                                     503.1
1,2,4-Trichlrobenzene, Benzene, Ethylbenzene, o-Dichlorobenzene, Styrene, Toluene, Total Xylene
                                      505
Alachlor, Atrazine, Chlordane, Endrin, Heptachlor, Heptachlor epoxide, Hexachlorobenzene, Hexachlorocyclopentadiene, Lindane, Methoxychlor, Simazine, Toxaphene
                                      507
Atrazine, Simazine
                                      508
Alachlor, Chlordane, Endrin, Heptachlor, Heptachlor epoxide, Hexachlorobenzene, Lindane, Methoxychlor, Toxaphene
                                     508A
Alachlor, Polychlorinated biphenyls (PCBs) general screen
                                     515.1
2,3,4-TP (Silvex), 2,4-D (2,4-Dichlorophenoxyacetic acid), Dalapon, Dinoseb, Pentachlorophenol, Picloram
                                     515.4
2,3,4-TP (Silvex)
                                     524.1
1,1,1-Trichloroethane, 1,1,2-Trichloroethane, 1,2-Dichloropropane, Benzene, Carbon tetrachloride, Ethylbenzene, o-Dichlorobenzene, Styrene, Toluene, Total Xylene
                                 524.2, 524.3
1,1-Dichloroethylene, 1,1,1-Trichloroethane, 1,1,2-Trichloroethane, 1,2-Dichlrorethane, 1,2,4-Trichlrobenzene, 1,2-Dichloropropane, Benzene, Carbon tetrachloride, cis-1,2-Dichloroethylene, Dichloromethane, Ethylbenzene, o-Dichlorobenzene, Styrene, Toluene, Total xylenes, trans-1,2-Dichloroethylene, Trichloroethylene, Vinyl chloride
                                     525.1
Alachlor (general screen), Atrazine, Benzo(a)pyrene, Chlordane (alpha and gamma), Heptachlor, Heptachlor epoxide, Hexachlorobenzene, Hexachlorocyclopentadiene, Pentachlorophenol, Simazine, Toxaphene
                                     525.2
Di(2-ethylhexyl) adipate, Di(2-ethylhexyl) phthalate, Endrin, Lindane, Methoxychlor
                                     531.1
Aldicarb, Aldicarb sulfone, Aldicarb sulfoxide, Baygon, Carbaryl, Carbofuran, 3-Hydroxycarbofuran, Methiocarb, Methomyl, Oxamyl
                                     531.2
Aldicarb, Aldicarb sulfone, Aldicarb sulfoxide, Carbofuran, 3-Hydroxycarbofuran, Methiocarb, Methomyl, 1-Napthol, Oxamyl, Propoxur
                                      533
11-Chloroeicosafluoro-3-oxaundecane-1-sulfonic acid, 9-Chlorohexadecafluoro-3-oxanonane-1-sulfonic acid, 4,8-Dioxa-3H-perfluorononanoic acid, Hexafluoropropylene oxide dimer acid, Nonafluoro-3,6-dioxaheptanoic acid, Perfluorobutanoic acid, Perfluorobutanesulfonic acid, 1H,1H, 2H, 2H-Perfluorodecane sulfonic acid, Perfluorodecanoic acid, Perfluorododecanoic acid, Perfluoro(2-ethoxyethane)sulfonic acid, Perfluoroheptanesulfonic acid, Perfluoroheptanoic acid, 1H,1H, 2H, 2H-Perfluorohexane sulfonic acid, Perfluorohexanesulfonic acid, Perfluorohexanoic acid, Perfluoro-3-methoxypropanoic acid, Perfluoro-4-methoxybutanoic acid, Perfluorononanoic acid, 1H,1H, 2H, 2H-Perfluorooctane sulfonic acid, Perfluorooctanesulfonic acid, Perfluorooctanoic acid, Perfluoropentanoic acid, Perfluoropentanesulfonic acid, Perfluoroundecanoic acid
                                     537.1
Hexafluoropropylene oxide dimer acid, N-ethyl perfluorooctanesulfonamidoacetic acid, N-methyl perfluorooctanesulfonamidoacetic acid, Perfluorobutanesulfonic acid, Perfluorodecanoic acid, Perfluorododecanoic acid, Perfluoroheptanoic acid, Perfluorohexanesulfonic acid, Perfluorohexanoic acid, Perfluorononanoic acid, Perfluorooctanesulfonic acid, Perfluorooctanoic acid, Perfluorotetradecanoic acid, Perfluorotridecanoic acid, Perfluoroundecanoic acid, 11-chloroeicosafluoro-3-oxaundecane-1-sulfonic
acid, 9-chlorohexadecafluoro-3-oxanonane-1-sulfonic acid, 4,8-dioxa-3H-perfluorononanoic acid
                                      547
Glyphosate
                                     548.1
Endothall
                                     549.2
Diquat
                                     551.1
Dibromochloropropane, Endrin, Ethylene Dibromide, Lindane, Methoxychlor, Trichloroethylene
                                      900
Radioactivity (gross alpha and gross beta particles)
                                      903
Radium-226 and Radium 228 combined
                                     908.1
Uranium
                                     9215B
Heterotrophic plate count
                                     1613
2,3,7,8-TCDD (Dioxin)
                                     00.02
Radon

The pilot study spreadsheet contains a lookup table with average per-sample analysis costs for each of these methods obtained from several certified labs. It is possible, however, that EPA will identify new analytical methods for a target contaminant when it proposes or promulgates a new standard. New analytical methods could be particularly likely when the target contaminant was previous unregulated. Therefore, to generate the costs reported in Section 4, EPA used a generic target contaminant cost of $200/sample. This analysis cost is slightly higher than the average cost ($187/sample) or the median cost ($183/sample) across the EPA methods shown in Exhibit 2. The cost is more comparable with the average cost for testing for organic chemicals than for inorganic contaminants. The pilot study spreadsheet allows two target contaminants, but the pilot study costs developed in this paper represent a typical case of one target contaminant.
Analytical costs depend on the number of samples analyzed, which is a function of the study length, the number of sampling points, and the number of parameters evaluated. Although these factors generally vary by technology type and study objective, there is no standard sampling approach in terms of sample frequency and analytes. The National Academy of Sciences National Research Council (NAS, 1997; p. 136) notes that "water treatment system designers generally conduct bench and pilot studies using their own individual methods" for documenting water quality. Therefore, to estimate the pilot study costs presented here, EPA incorporated sampling plans reported in published field studies and guidance documents. 
Sampling frequency data in the pilot study spreadsheet indicate the average number of samples per month, and the spreadsheet allows variation in sampling frequencies across analytes, sampling location, and source water types. Sampling location options include raw or influent water, treated water, and a waste stream that can reflect samples for recycle streams, sludge, or membrane concentrate flows. The source water options are groundwater and surface water. Surface water variability can require greater sampling frequency (and longer pilot study lengths); groundwater quality tends to be less variable over time. Total pilot study analytical costs presented in this document and used in the WBS reflect the source water that is most typical for each technology (i.e., groundwater for adsorptive media, surface water for conventional and direct filtration).
Although the pilot study spreadsheet does not have explicit initial and long-term sampling phases, the average number of samples per month can reflect variations in sampling frequency. For example, if a sampling plan includes weekly samples during the initial study month and monthly sampling for the next two months, the average monthly sampling frequency is 2 samples (i.e., a total of 6 samples divided by 3 months).
Labor Costs
Labor requirements for a pilot study include time spent planning and implementing the study and evaluating the results. The planning phase includes developing the study objectives and planning operational changes. Implementation activities include monitoring the pilot system, collecting operational data such as flow rate and pressure, chemical doses, and contact time, and collecting water quality samples for lab analysis and parameters evaluated onsite such as temperature, pH, turbidity, and hardness. Evaluation costs account for the time spent analyzing water quality and operational data to evaluate technology effectiveness under different conditions, and documenting the results.
EPA developed pilot study labor costs based on incremental labor hour requirements for two reasons. First, pilot study equipment or service suppliers rarely provided prices for pilot study labor services, but a couple provided estimates of how much time might be spent in various activities. Second, price data from equipment vendors excludes the costs of tasks undertaken by utility employees.
In principle, labor requirements will vary by type, length, and complexity of study. Labor requirements tend to be lowest for bench-scale studies, which can last less than two weeks and have minimal planning and reporting requirements. This is particularly true if the results are intended to direct pilot study efforts toward a more promising treatment alternative (e.g., a bench study to identify the adsorptive media that has superior removal efficiencies among several candidates, followed by a pilot study to collect additional data on optimal operating parameters). Labor estimates for bench tests, where applicable, are discussed in Section 4 on a technology-specific basis.
Most vendors provided no information on labor requirements for pilot-scale studies. Consequently, EPA used generic assumptions of 120 hours per study to design the pilot test protocols, contact vendors, install and test equipment, and document results (Neemann, 2004), and 8 hours per week to monitor and adjust the pilot system and collect samples (Kommineni, 2004). 
EPA valued labor time using an engineering wage rate for civil engineers in the water supply industry (NAICS 221300) from the Bureau of Labor Statistics (U.S. BLS, 2020a). EPA incorporated a load factor of 1.5 to account for benefits. This load factor is based on the ratio of total compensation including benefits to the wage rate for civilian employees in production occupations (U.S. BLS, 2020b). This valuation approach generates a conservative cost estimate (i.e., tending to overstate costs), because activities such as sampling can be conducted by staff that have lower wage rates.

Data
This section provides the technology-specific data EPA collected to estimate bench and pilot study costs, including:
Necessity of bench and pilot studies
Study duration
Equipment rental and purchase costs
Monitoring and testing (analytical) costs
Test design, operating, and report writing costs, where they differ from the general assumptions noted in Section 3.3.
This section is not intended to provide exhaustive information about how to conduct pilot or bench studies for individual technologies, or to summarize existing EPA guidance, although it does draw upon EPA guidance where applicable. The issues addressed here focus on costs and variables that drive costs (e.g., test duration), and how these variables are incorporated in EPA's bench and pilot cost estimates.
This section groups the WBS technologies into the following major categories:
Adsorptive media and ion exchange
Lime softening
Membrane filtration
Conventional and direct filtration
Ozone pre-oxidation and disinfection
Ultraviolet disinfection
Ultraviolet advanced oxidation
Aeration
Biological treatment
Biologically active filtration.
The remainder of this section addresses pilot study costs for each technology group.
Adsorptive Media and Ion Exchange Systems
Several WBS models are for technologies that use adsorption to remove contaminants from water. Adsorption is a physical and chemical process in which ions collect on the surface of an oxidized media as water flows through it. The models cover media including activated alumina (AA), granular ferric oxide (specifically SORB 33, manufactured by Severn Trent Services), granulated ferric hydroxide (GFH), granular activated carbon (GAC), and granular calcined diatomite (Media G2[(R)]). They also cover greensand filtration and ion (anion or cation) exchange. Most of these technologies remove anions or negatively charged contaminants such as nitrates, arsenic, and selenium. Cation exchange treatment systems, however, remove contaminant cations such as barium, radium, and strontium; they are generally used in "water softening" processes that remove hardness ions such as Ca[2+] and Mg[2+], replacing them with sodium or potassium.
Testing adsorptive or ion exchange systems can be accomplished through bench- or pilot-scale studies or, in some cases, both. The decision of whether to bench test, pilot test, or both depends on a facility's objectives, financial resources, and other site-specific characteristics. The objectives and uncertainties that drive this decision usually fall into the following categories:
Technical - to determine adsorbent performance characteristics, pollutant removal efficiencies, and design features to ensure that treatment objectives are met and that the proposed design parameters are correct
Economic - to minimize capital and operating costs.
Bench-scale systems for adsorptive media usually involve using microcolumn equipment such as RSSCTs. RSSCTs are a time- and cost-effective method for evaluating adsorptive media contaminant removal potential. RSSCTs are less expensive and require less time than pilot-scale tests (Kini et al., 2011). Although this type of testing is not applicable for all media (Gurian et al., 2010), RSSCTs have been used to evaluate and compare the performance of various types and brands of adsorptive media for the removal of arsenic and the reduction of disinfection byproduct precursors (Kini et al., 2011; Schers et al., 2011; Kohlmeier, 2011; Cleveland and Mount, 2011; Kwan et al., 2010; Chiu et al., 2009).
Microcolumns mimic liquid chromatography equipment, and use a narrow-size fraction of pulverized adsorbent (e.g., activated carbon) in high-pressure tubing. Other media can also be tested in this equipment as long as pulverization does not change the media's adsorption or ion exchange properties. Bench-scale systems can be used to compare various types of media, evaluate media effectiveness or pollutant removal efficiencies, and determine certain operating parameters such as breakthrough (i.e., determine the point in time where the contaminant ion no longer efficiently bonds with the resin). Conducting bench tests reduces the cost of pilot studies by reducing the time needed for piloting, and can even eliminate the need to pilot, depending on treatment objectives. The volume of water required for testing is also reduced with microcolumn testing because the tests are performed on a much smaller scale.
Pilot tests are similar to bench tests, except they are performed on a larger scale. If the media of choice cannot be pulverized in order to work in a microcolumn without affecting its adsorptive or ion exchange properties, a pilot-scale system could be necessary to test the effectiveness of the media, as well as develop breakthrough curves. Pilot testing could also be necessary to optimize design and operating parameters (e.g., determine degree of pretreatment necessary, optimal residence time, or dimension of column) to reduce capital and O&M costs, and to characterize process residuals.
Adsorption and ion exchange pilot studies most often use small columns that are designed to simulate a cylindrical core of a full-scale column, thereby reproducing the physical conditions of the full-scale application [e.g., quantity of media used, superficial velocity (rate of linear motion) through the column, and column height that is representative of a full-scale system]. Pilot columns for field applications should be sturdy, and be assembled using commercially available units (i.e., off-the-shelf units). To effectively simulate full-scale conditions, the pilot columns for granular adsorbents are usually at least one inch in diameter and four to six feet high. However, the objectives of the specific pilot study usually determine the system's actual size. The most critical factor in size determination is the amount of untreated water available with known chemical composition, because it is necessary to calculate the influent and effluent contaminant concentrations as a function of volume treated before breakthrough. The pilot column must maintain a certain size ratio to mimic the final system; therefore, the amount of water necessary for piloting is predetermined. 
Commercial and university laboratories offer bench-scale testing services for treatment facilities. For example, Arizona State University provides RSSCT testing services, including the costs of a selection of standard media (activated alumina, granular ferric hydroxide, and SORB-33) and the costs of taking samples (but not analyzing them). The test consists of running water from the facility, typically from a 50-gallon sample, through a column containing the medium, and taking the samples necessary to plot the breakthrough curve. Twenty to thirty samples are typically required to establish a breakthrough curve (Westerhoff, 2003). A typical bench study might consist of testing three media (Kommineni, 2003).
EPA estimated the per medium cost of bench-scale testing for adsorptive or ion exchange media based the cost per medium of RSSCT testing services and analysis of 25 target contaminant samples per medium (using the generic target contaminant analysis cost discussed in Section 3.2). This estimate might understate study costs because it does not include the cost of selecting the media to bench test or transporting water to the facility performing the tests. Conversely, it will overstate costs for target contaminants that cost substantially less per sample to analyze than the target contaminant analysis cost used here.
For pilot-scale testing for adsorptive and ion exchange media, EPA obtained estimates from companies that lease pilot columns and assumed a testing period of six months. Pilot testing for adsorptive media systems typically lasts about six months (Kommineni, 2004). The length of a pilot study is dictated by the time it takes for breakthrough to occur. Breakthrough is dictated by the concentration of the pollutant in the influent, adsorption capacity of the media, concentration of interfering ions, and pH (Kommineni, 2004).
A typical sampling protocol for a pilot-scale adsorptive media test would include sampling the influent and effluent for the contaminant of concern, for any competing ions, and for additional parameters to that may affect treatment effectiveness. The specific analytes would depend on the study objectives, contaminant of concern, treatment medium, and other variables. Exhibit 3 shows the sampling protocol that EPA used as a basis for estimating costs of analyzing water samples. 
Exhibit 3. Sampling Assumptions for Adsorptive Media Pilot Tests 
                                   Parameter
                                  Feed Water
                                   Effluent
                                Backwash/ Brine
                                   Residual
                                 (Spent Media)
Target Contaminant
                           20-30 times per study[1]
                           20-30 times per study[1]
                                   monthly 
                                     once
Alkalinity
                                    weekly
                                    weekly
                                      --
                                      --
Aluminum (dissolved)
                                    monthly
                                    monthly
                                      --
                                      --
Aluminum (total)
                                    weekly
                                    weekly
                                    monthly
                                      --
Total Hardness
                                    monthly
                                    monthly
                                      --
                                      --
Iron (dissolved)
                                    monthly
                                    monthly
                                      --
                                      --
Iron (total)
                                    weekly
                                    weekly
                                    monthly
                                      --
Manganese (dissolved)
                                    monthly
                                    monthly
                                      --
                                      --
Manganese (total)
                                    weekly
                                    weekly
                                    monthly
                                      --
Nitrate
                                    monthly
                                    monthly
                                      --
                                      --
pH
                                    weekly
                                    weekly
                                    monthly
                                      --
Sulfate
                                    weekly
                                    weekly
                                      --
                                      --
TOC
                                     once
                                     once
                                      --
                                      --
TSS
                                     once
                                     once
                                    monthly
                                      --
Turbidity
                                    monthly
                                    monthly
                                      --
                                      --
Percent Moisture (in residuals)
                                      --
                                      --
                                      --
                                     once
TCLP Metals (in residuals)
                                      --
                                      --
                                      --
                                     once
Source: Based on U.S. EPA (2000), Tables 3-3 and 3-4.
-- = None.
TOC = Total Organic Carbon, TSS = Total Suspended Solids, TCLP = Toxicity Characteristic Leaching Procedure.
1. EPA assumed that the target contaminant would be monitored at different intervals (i.e., less frequently in the beginning and then more often as the media nears breakthrough) throughout the testing in order to develop a breakthrough curve. Although more than 20-30 samples may be taken over the study period, not all of the samples taken would be analyzed. The estimate of labor for sampling and other tasks includes the time needed to take these additional samples.

The decisions of small, medium, and large systems to perform bench and pilot studies, the number of media to be tested, and the durations of studies vary depending on treatment objectives, the frequency of regeneration of the medium, and other factors. For the purpose of estimating average costs, EPA calculated costs of bench and pilot scale tests based on the following assumptions:
Small systems (with design flows less than 1 mgd) would bench test 2 media
Medium systems (with design flows of 1 to 10 mgd) would bench test 3 media and run a 6-month pilot test on one medium
Large systems (with design flows of over 10 mgd) would bench test 3 media and run a 6-month pilot test on one medium.
Additional costs include labor to design pilot tests, contact vendors, install pilot equipment, collect water quality samples, monitor the operation of the system, analyze the sampling results and adjust the system accordingly, and develop a final report containing recommendations for the design and operation of a full-scale system. For labor, EPA used the assumptions noted in Section 3.3. Exhibit 4 provides a summary of the cost components for bench and pilot studies by system size.
Exhibit 4. Bench and Pilot Study Costs for Adsorptive Media Systems (2019 dollars)
                                   Component
                          Small Systems 
(<1 mgd)
                           Medium Systems
(1-10 mgd)
                          Large Systems
(>10 mgd)
Bench Study
                                    $15,000
                                    $22,500
                                    $22,500
Pilot study  -  Equipment
                                      NA
                                    $7,000
                                    $7,000
Pilot study - Lab analysis (target contaminant)
                                      NA
                                    $4,800
                                    $4,800
Pilot study - Lab analysis (non-target contaminant)
                                      NA
                                    $9,800
                                    $9,800
Pilot study  -  Labor
                                      NA
                                    $22,800
                                    $22,800
Total
                                    $15,000
                                    $67,000
                                    $67,000
Lab analysis costs reflect groundwater sampling frequency. 
Detail may not add to totals because total amounts are rounded to report costs to the nearest $100.
NA = Not applicable.

Lime Softening
The lime softening process involves raising the pH of the water to precipitate calcium carbonate and magnesium hydroxide; the process can also remove radium, arsenic, and other inorganic chemicals. After mixing, flocculation, sedimentation, and pH readjustment, the softened water is filtered (U.S. EPA, 1998).
The chemistry of lime softening can be complex; for example, a failure to maintain the proper pH in softened water prior to filtration could result in precipitation of excess lime in filter beds and the formation of calcium carbonate deposits within the filters (U.S. EPA, 1998). Thus, the technology is best suited to relatively large surface water systems, which are more likely to have a trained operator onsite at all times, or groundwater systems with stable water quality (NAS, 1997; U.S. EPA, 1998). 
Pilot testing for lime softening, however, is rare. For example, a consulting engineer with a firm that has installed many lime softening plants reported that he is aware of only one system (other than those using lime softening as pretreatment) that installed a pilot test before installing the full-scale system (Hulsey, 2004). Most systems perform a desktop evaluation, following source water characterization, to determine the appropriate dose and determine other system parameters (e.g., one-stage or two-stage). 
Typically, systems using lime softening would test the feed water for alkalinity, pH, calcium and magnesium hardness, iron, and manganese, prior to the desktop analysis (Hulsey, 2004). For cost estimating purposes, EPA assumed that existing public water systems already monitor for these parameters and, therefore, have this information. The desktop analysis itself would take about eight hours, including time to evaluate the water quality data, determine the appropriate lime dose, and determine the appropriate design and operating parameters (Hulsey, 2004). Therefore, EPA used the cost of eight hours of engineering labor to estimate the cost of desktop analysis for small, medium, and large systems. Exhibit 5 provides a summary of the cost components for bench and pilot studies by system size.
Exhibit 5. Bench and Pilot Study Costs for Lime Softening Systems (2019 dollars)
                                   Component
                          Small Systems 
(<1 mgd)
                           Medium Systems
(1-10 mgd)
                          Large Systems
(>10 mgd)
Desktop Analysis
                                     $600
                                     $600
                                     $600
Total
                                     $600
                                     $600
                                     $600
Amounts are rounded to report costs to the nearest $100.

Membranes
Membrane processes utilize semi-permeable membranes to separate suspended and/or dissolved particles from water. Membrane systems in the suite of WBS drinking water models include reverse osmosis (RO), nanofiltration (NF), microfiltration (MF), ultrafiltration (UF), and electrodialysis reversal (EDR). RO and NF are considered high pressure processes, MF and UF are grouped together as low pressure membrane processes, and EDR is a membrane process in which an electric field is used to remove ionic components from aqueous solution by passing through ion exchange membranes. The WBS models further categorize low pressure processes as either pressurized or submerged. Because utilities often pilot test pressurized and submerged membranes alongside one another, the pilot cost estimates ignore this distinction.
The different membrane processes vary in their usual applications, their associated pretreatment needs, and the questions that must be evaluated in their implementation. Therefore, this analysis considers their pilot study costs separately.
Low-Pressure Membranes (MF and UF)
Pilot testing of low-pressure membrane systems provides important operational information. It is common to pilot test two or more manufacturers' systems side by side, to determine whether they can meet the treatment goals without excessive fouling, and to estimate operational parameters. The results are then used to select a vendor, by computing an annualized cost for each system based on vendors' quotes and the operational parameters for each system.
The detailed goals of bench- and pilot-scale testing for low-pressure membrane systems can include the following (NAS, 1997; U.S. EPA, 2005; AWWA, 2005):
Measuring feed water characteristics that influence membrane system design
Comparing the performance of different membrane systems for the water being treated
Obtaining data to optimize operational parameters such as membrane flux
Obtaining data on membrane fouling and optimize cleaning cycles
Determining the design criteria for pretreatment, if needed to reduce membrane fouling or enhance contaminant removal
Verifying that finished water meets the project's treatment goals
Training operations staff on a new technology
Complying with state or local regulatory requirements.
Systems typically use bench-scale testing to characterize their membrane feed water; those that use coagulants may also perform jar testing (NAS, 1997; Wachinski, 2003). EPA's Membrane Filtration Guidance Manual for the Long Term 2 Enhanced Surface Water Treatment Rule (LT2ESWTR) lists water quality parameters that should be measured prior to pilot testing to provide membrane vendors the data they need to recommend operational parameters (U.S. EPA, 2005). Exhibit 6 provides a summary. EPA assumed that systems would take one sample for each parameter recommended.
Exhibit 6. List of Measurements Required Prior to Membrane Pilot Testing
                                   Parameter
                      Microfiltration or Ultrafiltration
                       Nanofiltration or Reverse Osmosis
Algae (surface water only)
                                      Yes
                                       
Alkalinity
                                       
                                      Yes
Aluminum
                                       
                                      Yes
Ammonia
                                       
                                      Yes
Barium
                                       
                                      Yes
Calcium
                                      Yes
                                      Yes
Chloride
                                       
                                      Yes
Fluoride
                                       
                                      Yes
Iron (total)
                                      Yes
                                      Yes
Magnesium
                                      Yes
                                      Yes
Manganese (total)
                                      Yes
                                      Yes
Nitrate
                                       
                                      Yes
pH
                                      Yes
                                      Yes
Potassium
                                       
                                      Yes
Silica (Silicon Dioxide)
                                      Yes
                                      Yes
Silt Density Index
                                       
                                      Yes
Sodium
                                       
                                      Yes
Strontium
                                       
                                      Yes
Sulfate
                                       
                                      Yes
TDS
                                       
                                      Yes
TOC
                                      Yes
                                      Yes
TSS
                                      Yes
                                       
Turbidity
                                      Yes
                                      Yes
UV-254 absorbance
                                      Yes
                                      Yes
Source: based on U.S. EPA (2005), Table 6.1.
Blank cells indicate that the parameter will not be measured.
TDS = Total Dissolved Solids, TOC = Total Organic Carbon, TSS = Total Suspended Solids.

EPA received varied information about pilot testing for small systems. A manager at a leading membrane manufacturing company stated that it is unlikely that systems treating less than 1 mgd would perform a pilot study; those systems would more likely consult with membrane manufacturers and then buy a membrane filtration module that is third-party certified to the ANSI/NSF 61 standard (Wachinski, 2003). Conversely, a survey of MF and UF installations (Adham et al., 2005) reports that four out of six installations with design flows less than 1 mgd conducted pilot testing; the two that did not had design flows less than 0.1 mgd. The survey also shows that pilot testing is common for medium and large systems. Because pilot testing is often closely tied to the selection of a vendor, EPA included pilot testing costs for systems of all sizes.
U.S. EPA (2005) states that pilot tests for MF/UF systems are typically three to seven months, which would represent about 2,200 to 7,000 hours of continuous operation per system being tested. Studies are conducted over this duration for two major reasons: first, to ensure the membrane system works for different source water characteristics that can vary seasonally and affect performance (e.g., turbidity, algae content, and temperature); and second, to operate the membrane through several (ideally three) cleaning cycles (U.S. EPA, 2005). The duration of membrane pilot tests will also vary depending on specific treatment objectives and other site-specific characteristics. For example, a manager at a membrane manufacturing company reported that he is aware of pilot tests that have lasted for as little as one month or as long as 12 months (Wachinski, 2003).
EPA collected equipment rental cost estimates for low-pressure membrane filtration pilot study equipment from manufacturers and vendors. EPA also added an estimated cost of freight both ways. For small and medium systems, EPA estimated costs based on a 3-month pilot test. For large systems, EPA estimated costs based on a 7-month pilot test. Because it is common to pilot more than one system, EPA assumed that medium and large systems would test units from two vendors, doubling effluent analysis, equipment rental, and labor costs. EPA assumed small systems would pilot only one vendor's equipment.
The test protocol for membrane pilot systems depends on the nature of a specific rule, as described in Section 3. To estimate potential analytical costs, EPA used sample protocols for membrane systems from the suggested water quality sampling schedule for membrane piloting from the Membrane Filtration Guidance Manual for LT2ESWTR (U.S. EPA, 2005), shown in Exhibit 7. The manual recommends continuous (online) measurement for temperature, turbidity, and particle counts; EPA assumed that the pilot unit rental would include the instruments needed for this monitoring. Many membrane systems will operate in dead-end mode, that is, with backwash but no concentrate flow. To produce a conservative cost estimate, EPA included the cost of sampling the concentrate water as recommended in Exhibit 7.
Additional costs for the pilot test include labor to design the test, contact vendors, install equipment, collect water quality samples, monitor the operation of the system, analyze the sampling results and adjust the system accordingly, and develop a final report containing recommendations for the design and operation of a full-scale system. For pilot testing labor, EPA used the assumptions noted in Section 3.3.
Exhibit 7. Analysis Sampling Schedule for Membrane Pilot Testing
                                   Parameter
                      Microfiltration or Ultrafiltration
                       Nanofiltration or Reverse Osmosis

                                  Feed Water
                          Concentrate[1] or Backwash
                                   Filtrate
                                  Feed Water
                                  Concentrate
                                   Filtrate
Algae (surface water only)
                                    weekly
                                      --
                                    weekly
                                    weekly
                                      --
                                      --
Alkalinity
                                    weekly
                                    monthly
                                    weekly
                                    weekly
                                    weekly
                                    weekly
Barium
                                      --
                                      --
                                      --
                                    weekly
                                    weekly
                                    weekly
Color
                                    weekly
                                    monthly
                                    weekly
                                    weekly
                                      --
                                    weekly
Total Hardness
                                      --
                                      --
                                    monthly
                                    weekly
                                    weekly
                                    weekly
HPC
                                    monthly
                                      --
                                    monthly
                                    monthly
                                      --
                                    monthly
Iron (total)
                                    weekly
                                      --
                                      --
                                    weekly
                                    weekly
                                    weekly
Manganese (total)
                                    weekly
                                      --
                                      --
                                    weekly
                                    weekly
                                    weekly
Particle Count
                                  continuous
                                      --
                                  continuous
                                      --
                                      --
                                      --
pH
                                     daily
                                    weekly
                                    weekly
                                  continuous
                                    weekly
                                    weekly
Silica (Silicon Dioxide)
                                      --
                                      --
                                      --
                                    weekly
                                    weekly
                                    weekly
Silt Density Index
                                      --
                                      --
                                      --
                                    weekly
                                      --
                                      --
Sulfate
                                    weekly
                                      --
                                      --
                                    weekly
                                    weekly
                                    weekly
Taste and Odor Compounds
                                    monthly
                                    monthly
                                    monthly
                                    monthly
                                    monthly
                                    monthly
Temperature
                                  continuous
                                      --
                                      --
                                  continuous
                                      --
                                      --
Total Coliform
                                    weekly
                                      --
                                    weekly
                                    weekly
                                      --
                                      --
TDS
                                      --
                                      --
                                    monthly
                                  continuous
                                    weekly
                                  continuous
TOC
                                    weekly
                                    monthly
                                    weekly
                                    weekly
                                    monthly
                                    weekly
TSS
                                    monthly
                                   weekly[2]
                                    monthly
                                      --
                                      --
                                      --
Turbidity
                                  continuous
                                   weekly[2]
                                  continuous
                                  continuous
                                      --
                                    weekly
UV-254 absorbance
                                    weekly
                                    monthly
                                    weekly
                                    weekly
                                      --
                                    weekly
Source: based on U.S. EPA (2005), Table 6.2.
-- = No sample included in generic schedule, HPC = Heterotrophic Plate Count, TDS = Total Dissolved Solids, TOC = Total Organic Carbon, TSS = Total Suspended Solids.
1. Where applicable, based on particular hydraulic configuration used.
2. The reference recommends weekly testing of TSS and turbidity in the backwash water. 

High-Pressure Membrane Systems (RO and NF)
The detailed goals of bench- and pilot-scale testing for high-pressure membrane systems can include the following (NAS, 1997; U.S. EPA, 2005; AWWA, 2005):
Measuring feed water characteristics that influence membrane system design
Obtaining data to optimize operational parameters such as membrane flux
Obtaining data on membrane fouling and optimize cleaning cycles
Determining requirements for acid or scale inhibitor pretreatment
Verifying that membrane permeate water meets the project's treatment goals
Training operations staff on a new technology
Complying with state or local regulatory requirements.
A feed water quality analysis is required for the design of a high-pressure membrane system. The analysis is needed to identify cations and anions that may cause scaling of the membranes when they are concentrated in the membrane system, the potential for fouling by silica precipitation, and the potential for biological fouling of the membrane elements (AWWA, 2007). EPA assumed that all systems would measure the parameters shown in Exhibit 6 for NF and RO.
In contrast to low-pressure membrane systems, many aspects of RO or NF can be effectively predicted with a desktop simulation (U.S. EPA, 2005). For the desktop analysis costs for NF/RO systems, EPA assumed eight hours of labor time for a civil engineer. This includes the time taken by a system engineer to send data on water quality parameters to a process design company (or send a water sample), the costs of the computer study, and time to report the results of the study. EPA applied this cost to NF/RO systems of all sizes.
NAS (1997) suggests that for small groundwater systems using membranes, the only site-specific analysis that may be necessary is evaluation of source water characteristics to determine the potential for chemical scaling of the membranes, though it also states that researchers recommend a pilot study for small surface water systems. EPA notes that data for removal effectiveness is increasingly available from manufacturers (U.S. EPA, 2005) as well as the scientific literature (U.S. EPA, 2020). When membrane effectiveness for removing contaminants is established via challenge testing by manufacturers themselves, the costs of effectiveness tests will be included in the costs of membranes. EPA concluded that small water systems are unlikely to conduct pilot tests to assess the effectiveness of membranes for contaminant removal because three factors reduce the need for pilot testing: the increased use of membrane technology for water systems since the publication of the NAS report; the relatively small number of membrane element manufacturers and equipment options (i.e., available package plants); and the verification and certification efforts through ETV and NSF 61. Therefore, NF/RO processes for small systems have costs for a desktop analysis.
For medium and large NF/RO systems, EPA used equipment rental cost estimates from manufacturers and vendors. EPA also added an estimated cost of freight both ways. The Membrane Filtration Guidance Manual for LT2ESWTR states that pilot tests for NF and RO systems are typically two to seven months, which represents 1,500 to 5,000 hours of continuous operation (U.S. EPA, 2005). As with low-pressure membranes, EPA assumed that medium systems would operate a pilot unit for three months, and large systems would operate it for seven months. EPA also added the cost to purchase a silt density index kit for use in monitoring this parameter during pilot testing.
As with MF/UF, EPA estimated potential analytical costs based on the suggested water quality sampling schedule for membrane piloting from the Membrane Filtration Guidance Manual for LT2ESWTR (U.S. EPA, 2005), as shown in Exhibit 7. The manual recommends continuous (online) measurement for pH, TDS, temperature, and turbidity. EPA assumed that the pilot unit rental would include the instruments needed for this monitoring. For the purposes of estimating costs, EPA assumed that systems would also sample for one target contaminant, weekly, in the feed water, the concentrate, and the filtrate.
Additional costs for the pilot test include labor to design the test, contact vendors, install equipment, collect water quality samples, monitor the operation of the system, analyze the sampling results and adjust the system accordingly, and develop a final report containing recommendations for the design and operation of a full-scale system. For pilot testing labor, EPA used the generic assumptions noted in Section 3.3.
Electrodialysis Reversal Systems (EDR)
The detailed goals of bench- and pilot-scale testing for EDR systems can include the following (AWWA, 1995; Boulos, 2013; Reed, 2002):
Measuring feed water characteristics that influence membrane system design
Determining optimum reactor configuration and size (depth of anode ribbons, optimum gap between anodes)
Establishing operational hydraulic residence time
Determining optimum electric current range and density
Determining optimum power for contaminant removal
Verifying that membrane permeate water meets the project's treatment goals
Training operations staff on a new technology
Complying with state or local regulatory requirements.
EPA reviewed available literature and determined that utilities use pilot scale analysis to evaluate the effectiveness of EDR treatment and to develop design parameters. Typically, a single membrane vendor provides a pilot system. Small systems, however, can rely on packaged vendor designs selected based on bench-scale information and/or desktop simulation, which can effectively predict EDR performance (AWWA, 1995). Therefore, similar to RO/NF, EPA limited pilot study costs for small systems to feed water quality analysis and desktop simulation.
For the desktop analysis costs for EDR systems, EPA assumed eight hours of labor time for a civil engineer. This labor includes the time taken by a system engineer to send data on water quality parameters to a process design company (or send a water sample), the costs of the computer study, and time to report the results of the study. EPA included this cost for EDR systems of all sizes.
The design of an EDR membrane system requires an initial feed water quality analysis to identify cations and anions that could potentially pass through the membrane barriers when an electrical charge is applied. EPA included the cost of initial feed water quality analysis for EDR systems of all sizes. For large and medium systems, EPA included ongoing monitoring during the pilot test as shown in Exhibit 8. EPA assumed that the rental pilot unit would include the instruments needed for monitoring temperature and pH continuously on the influent and effluent lines. For the purposes of estimating costs, EPA assumed that systems would also sample for one target contaminant, weekly, in the feed water, the concentrate, and the effluent.
Exhibit 8. Non-Target Contaminant Sampling Parameters and Frequencies for EDR
                                   Parameter
                                  Feed Water
                                   Effluent
                                  Concentrate
Temperature
                                 Continuously
                                 Continuously
                                    Weekly
pH
                                 Continuously
                                 Continuously
                                    Weekly
Sodium
                          Once initially, then weekly
                                    Weekly
                                    Weekly
Calcium
                          Once initially, then weekly
                                    Weekly
                                    Weekly
Magnesium
                          Once initially, then weekly
                                    Weekly
                                    Weekly
Chloride
                          Once initially, then weekly
                                    Weekly
                                    Weekly
Fluoride
                          Once initially, then weekly
                                    Weekly
                                    Weekly
Potassium
                          Once initially, then weekly
                                    Weekly
                                    Weekly
Bicarbonate
                          Once initially, then weekly
                                    Weekly
                                    Weekly
Sulfate
                          Once initially, then weekly
                                    Weekly
                                    Weekly
Nitrate
                          Once initially, then weekly
                                    Weekly
                                    Weekly
TDS
                          Once initially, then weekly
                                    Weekly
                                    Weekly
Alkalinity
                          Once initially, then weekly
                                    Weekly
                                    Weekly
Hardness
                          Once initially, then weekly
                                    Weekly
                                    Weekly
Turbidity
                          Once initially, then weekly
                                    Weekly
                                    Weekly
Sources: AWWA (1995), Reed (2002)
TDS = Total Dissolved Solids.

EPA used equipment rental cost estimates from membrane manufacturers and vendors. EPA also added an estimated cost of freight both ways. As with other membrane technologies, EPA assumed that medium systems would operate a pilot unit for three months, and large systems would operate it for seven months.  
Additional costs for the pilot test include labor to design the test, contact vendors, install equipment, collect water quality samples, monitor the operation of the system, analyze the sampling results and adjust the system accordingly, and develop a final report containing recommendations for the design and operation of a full-scale system. For EDR pilot testing labor, EPA used the generic assumptions noted in Section 3.3.
Summary
Exhibit 9 is a summary of the cost components for bench and pilot studies by system size for the different membrane technologies. As stated previously, these represent approximate costs for a typical analysis; durations of actual pilot tests and actual equipment costs will vary depending on site-specific factors.
Exhibit 9. Bench and Pilot Study Costs for Membrane Systems (2019 dollars)
                                   Component
                           Small Systems
(<1 mgd)
                           Medium Systems
(1-10 mgd)
                          Large Systems
(>10 mgd)
MF and UF Systems
Bench study
                                      NA
                                      NA
                                      NA
Pilot study - Equipment
                                    $16,200
                                    $32,400
                                    $72,400
Pilot study - Lab analysis
                                    $23,900
                                    $36,600
                                    $84,700
Pilot study - Labor
                                    $15,800
                                    $31,600
                                    $50,400
Total
                                    $55,900
                                   $100,600
                                   $207,400
NF and RO Systems
Bench Study
                                     $600
                                     $600
                                     $600
Pilot study - Equipment
                                      NA
                                    $17,100
                                    $37,100
Pilot study - Lab analysis
                                     $600
                                    $29,800
                                    $68,800
Pilot study - Labor
                                      NA
                                    $15,800
                                    $25,200
Total
                                    $1,200
                                    $63,300
                                   $131,600
EDR Systems
Bench Study
                                     $600
                                     $600
                                     $600
Pilot study - Equipment
                                      NA
                                    $31,500
                                    $72,000
Pilot study - Lab analysis
                                     $200
                                    $16,000
                                    $36,900
Pilot study - Labor
                                      NA
                                    $15,800
                                    $25,200
Total
                                     $800
                                    $63,900
                                   $134,700
Lab analysis costs reflect surface water sampling frequency. 
Detail may not add to totals because total amounts are rounded to report costs to the nearest $100.
NA = Not applicable.

Conventional and Direct Filtration
Conventional filtration or coagulation/filtration systems incorporate coagulant addition prior to media filtration using rapid sand, dual media, multi media, and slow sand systems (although EPA's WBS model for conventional filtration does not accommodate slow sand). Conventional filtration systems also include a sedimentation basin for particle removal prior to filtration; direct filtration systems do not. Direct filtration is employed when raw water quality is fairly stable, with low turbidity, low turbidity-causing agents, and low color (Viessman and Hammer, 1998).
For conventional filtration systems that use well-known technologies, site-specific pilot testing might not be needed. Bench-scale jar tests can be sufficient to determine coagulant doses. Engineers can often use established principles to design systems to obtain adequate control of certain parameters, such as turbidity, because conventional filtration has been used for many years and for many different source waters. However, pilot tests could be necessary for conventional filtration systems if additional parameters, such as algae, are also of concern, to determine their effect on performance characteristics (e.g., fouling, pollutant removal efficiencies, and residuals management) (NAS, 1997). In addition, pilot tests can be useful to optimize operational parameters (e.g., coagulant dose, backwash and cleaning frequencies) (NAS, 1997). Pilot tests generally are conducted for direct filtration systems, unless there is already a direct filtration system treating the same source water successfully, because these systems are highly sensitive to water quality (NAS, 1997).
Some states recognize that cost may be an issue for smaller systems, and provide for exceptions to pilot testing. For example, the Washington State Department of Health allows, in cases where the cost of pilot testing approaches the cost of the full-scale technology, that the pilot testing phase may be included as part of the start-up process. A certified engineer can use data from the full-scale operation as pilot data, and make adjustments to the operational parameters as needed (WADOH, 2003). 
Thus, EPA estimated costs for small (less than 1 mgd) conventional filtration systems based on jar testing only (to determine the optimal coagulant dose and flocculation parameters) and not pilot tests. EPA estimated costs for medium and large systems based on both jar testing (to determine initial coagulant dose and flocculation parameters) and pilot testing. Note that these assumptions were used for cost estimating purposes only; the applicability, duration, sampling protocol, and equipment requirements of jar tests and pilot tests will vary depending on site-specific factors.
Jar test costs include equipment and labor time to design and operate the test and evaluate the results. To estimate costs for jar test equipment, EPA collected cost data from companies that manufacture jar testing equipment. Labor time requirements are minimal because test design and implementation is straightforward. EPA assumed a total of two days (or 16 hours) would be required to design the test protocols, contact vendors, conduct the tests and document results.
For small systems (less than 1 mgd), EPA used equipment costs based on the least expensive model identified. Since the price range and specifications for the remaining models does not vary greatly, EPA assumed that both medium and large systems would use a jar testing model with the average cost of the remaining models.
Pilot studies typically last from three to six months, including at least one cleaning cycle per month during the study period (Kommineni, 2003). EPA used equipment rental cost estimates from a vendor that provides pilot testing services. For the purposes of estimating costs, EPA assumed a three month pilot test for medium sized plants and a six month test for large systems. 
For cost estimating purposes, sampling parameters and frequencies are based on the sampling protocol, as reported in NAS (1997), for a 9-month pilot study at a conventional filtration system for a source water with moderate turbidity, algae problems, color, and periodic iron and manganese. Exhibit 10 provides a summary of the parameters and frequencies indicated in that document. EPA also assumed weekly testing for the target contaminant in both the feed water and the effluent. Online analyzers for continuous monitoring of temperature, pH, and turbidity are not included in pilot sampling costs because they are already accounted for in the WBS model cost for the full-scale technology. 
Additional costs include labor to design pilot tests, contact vendors, install pilot equipment, collect water quality samples, monitor the operation of the system, analyze the sampling results and adjust the system accordingly, and develop a final report containing recommendations for the design and operation of a full-scale system. EPA used the assumptions described in Section 3.3 to estimate labor time needed for these tasks. Exhibit 11 provides a summary of the cost components for bench and pilot studies by system size.
Exhibit 10. Non-Target Contaminant Sampling Parameters and Frequencies for Conventional or Direct Filtration Pilot Study
                                   Parameter
                                  Feed Water
                                   Effluent
Temperature
                                     Daily
                                     Daily
Alkalinity
                                    Weekly
                                    Weekly
Aluminum (total)
                                    Weekly
                                    Weekly
Iron (total)
                                    Weekly
                                    Weekly
Manganese (total)
                                    Weekly
                                    Weekly
Color
                                    Weekly
                                    Weekly
Algae (applicable to surface water)
                                    Weekly
                                    Weekly
TOC
                                    Weekly
                                    Weekly
TTHMs (in simulated distribution system)
                                    Monthly
                                    Monthly
HAAs (in simulated distribution system)
                                    Monthly
                                    Monthly
pH
                                 Continuously
                                 Continuously
Turbidity
                                 Continuously
                                 Continuously
Source: NAS (1997)
TOC = Total Organic Carbon, TTHMs = Total Trihalomethanes, HAAs = Haloacetic Acids.

Exhibit 11. Bench and Pilot Study Costs for Conventional and Direct Filtration Systems (2019 dollars)
                                   Component
                           Small Systems
(<1 mgd)
                           Medium Systems
(1-10 mgd)
                          Large Systems
(>10 mgd)
Bench study
                                    $2,600
                                    $4,500
                                    $4,500
Pilot study - Equipment
                                      NA
                                    $44,400
                                    $88,800
Pilot study - Lab analysis
                                      NA
                                    $18,000
                                    $36,100
Pilot study - Labor
                                      NA
                                    $15,800
                                    $22,800
Total
                                    $2,600
                                    $82,700
                                   $152,100
Lab analysis costs reflect surface water sampling frequency. 
Detail may not add to totals because total amounts are rounded to report costs to the nearest $100.
NA = Not applicable.

Ozone
Ozone is a powerful oxidant that can be used as a disinfectant or a preoxidant to improve coagulation and aid in the removal of taste and odor compounds, color, metals, and pesticides. The use of ozone in potable water treatment has increased substantially in the last 20 years because it provides superior inactivation of microorganisms (e.g., cysts) compared to chlorine, chloramine, and chlorine dioxide. Ozone is applied to drinking water as a gas, which is generated onsite. The ozone gas is transferred into water either by injecting under pressure or bubbling. Ozone can be applied to untreated, coagulated/settled, or filtered water (NSF, 2003a). Thus, it can be used in conjunction with other technologies (e.g., filtration) and applied at the beginning, the middle, or the end of the treatment train. 
Pilot or bench testing can be used to:
establish design parameters for a full-scale system (e.g., size of reaction vessel, demand and decay information, dose and contact time)
assess the effects of ozone on downstream components of the treatment train
establish the optimum ozone dose for the desired purpose
satisfy regulatory requirements
gain experience operating the technology.
When used as a preoxidant, the benefits of ozone include improving coagulation and reducing turbidity. However, it is generally undesirable to have residual ozone in the downstream process because it can volatilize easily and pose health and safety problems for plant operators. For this reason, plants that are considering using ozone as a pretreatment (preoxidation) step may use pilot testing to determine the optimum ozone dosage, and/or the need to "quench" ozone reactions by adding other chemicals to the water (Neemann, 2004). However, systems may also be able to set these parameters by experimentation after the full-scale system has been installed (Neemann, 2004). Ozone can also increase the production of assimilable organic carbon (AOC) as well as bromate, although bromate formation is less likely when ozone is used as a preoxidant because of the lower dose and contact time. Nevertheless, these are other reasons that pilot or bench studies might be used prior to building full-scale ozone preoxidation systems.
When used as a disinfectant, ozone is more likely to result in bromate formation. Bromate formation occurs in waters with high concentrations of bromide ion; even bromide concentrations on the order of 35 ug/L can result in downstream concentrations of bromate of 10 ug/L or more (i.e., in excess of the MCL for bromate), depending on other source water characteristics and the dose and contact time used for ozone (Overbeck, 2004). Determining the optimum dose and contact time for ozone to produce a residual for disinfection, along with determining any pretreatment steps necessary to inhibit bromate formation, can be sufficiently complex as to require pilot testing. In this situation, pilot testing would be beneficial to determine equipment design parameters such as the size of the contactor vessel, type of injection system (e.g., Venturi injectors or another system), and upstream or downstream treatment components. However, the optimum ozone dosage and contact time could also be determined through bench testing. In a bench test, ozone would be bubbled through the water and then the water would be tested for the ozone residual, bromate concentration, and other key parameters. Bench testing could be preformed onsite or by laboratories, depending on the number of different dose and contact time combinations tested (Neemann, 2004). 
Additional factors that can influence the decision to pilot or bench test ozone disinfection systems include the regulatory environment and the need to gain operator experience. States have widely different amounts of experience with ozone disinfection. Because ozone is not a technology in widespread use (e.g., compared to chlorination), regulators in some states may require pilot testing to verify the capacity to achieve the required level of disinfection and preclude bromate formation. In states where ozone is used relatively widely, regulators may exempt small systems from the requirement to pilot test, for example, if the consultant installing the system is well known, or if the system can identify another system with the same or similar source water already using ozone (Overbeck, 2004). 
Equipment necessary for a pilot test includes:
ozone generation equipment (gas supply system and ozone generator)
a transfer mechanism (e.g., Venturi injector)
a contactor or reaction vessel
an offgas destruction unit (Neemann, 2004).
The same equipment is generally required for a bench test, although the transfer mechanism may be simpler because bench tests use a batch reaction instead of a continuous flow (Neemann, 2004). For example, a bench test can involve bubbling ozone through distilled water under a fume hood to achieve a high concentration, then extracting a quantity of the water (e.g., 100 mL) and injecting it into a known quantity of source water (e.g., 900 mL). After allowing the reactions to take place, the water could be tested for parameters such as pH, dissolved residual ozone, bromate, AOC, THMs, and HAAs (Neemann, 2004).
Most bench-scale studies test a number of various ozone doses and contact times in order to determine optimal operating parameters. It is usually most cost-effective to send water samples to a laboratory for this bench testing (Neemann, 2004). EPA used estimates from laboratories of testing costs per condition (e.g., per dose and contact time), and assumed systems would test 10 to 20 different conditions. Analysis of additional parameters such as bromate and THMs would increase the per-test cost through additional analytical costs (Neemann, 2004).
For pilot-scale studies, EPA used equipment rental cost estimates from manufacturers and vendors. EPA also added the cost to purchase a dissolved ozone monitor, because some of the equipment available for rental did not include this item. EPA also added an estimated cost of freight both ways. For systems running multiple pilot tests at different times of the year, EPA counted freight costs multiple times.
The duration of pilot tests varies depending on the objectives of the study, the complexity of the treatment train, and other site-specific characteristics. It typically takes one to two weeks to establish the optimum ozone dose for disinfection or preoxidation and to verify the applicability of the technology (Hulsey, 2004), although it can take more time to optimize additional components of an existing or planned treatment train (e.g., upstream chemical addition or downstream filtration). Systems with consistent source water quality, such as small groundwater systems, are generally able to obtain the necessary data in a couple of weeks. Larger groundwater systems that use water from multiple wells might run pilot tests on different mixes of water; for example, a system using multiple wells with varying water quality might run separate pilot tests on the most problematic water and the average blend (Hulsey, 2004). Surface water systems with seasonally variable water quality might need to run a pilot test more than once (e.g., two to four times in a year to capture the range of seasonal variability) (Hulsey, 2004). For the purposes of estimating costs, EPA assumed that, on average:
small groundwater systems using ozone for disinfection would run a 2-week pilot test
medium and large groundwater systems using ozone for disinfection would run two separate 2-week pilot tests
small, medium, and large surface water systems using ozone for disinfection would run three separate 2-week pilot tests, in different periods of the year
all systems using ozone for pre-oxidation (regardless of size or source water type) would perform a bench test.
However, actual choices about whether to run a bench test, pilot test, both, or neither, as well as durations for those tests, will depend on treatment objectives, source water quality, availability of funds, regulatory environment, and other site-specific factors. 
For pre-oxidation bench costs, EPA assumed that small systems would test 10 conditions, medium systems would test 15 conditions, and large systems would test 20 conditions. Although some systems might use pilot tests before installing ozone pre-oxidation systems, EPA assumed for cost estimating purposes that pre-oxidation systems would perform bench testing only.
A sampling protocol for an ozone disinfection pilot test would depend on site-specific variables. For the purposes of estimating average costs, EPA used a modified version of the ETV protocol for inactivating microbial agents (NSF, 2003a), for ozone disinfection. EPA modified the protocols used to estimate sampling analysis costs for disinfection by excluding analysis of feed and treated water for bacteria, viruses, and protozoa. In addition, EPA excluded analytical tests suggested in both ETV protocols that are applicable only to using ozone in conjunction with ultraviolet light or hydrogen peroxide, along with tests of the concentration of any quenching agents used (e.g., to stop reactions that result from upstream chemical addition). Finally, because the purpose of ETV testing is different than the purpose of pilot testing, EPA used different (generally lower) testing frequencies.
Exhibit 12 provides a summary of the testing protocols for water quality tests used for cost estimating. EPA did not incorporate costs for tests for the target contaminant, because disinfection effectiveness would likely be measured using CT calculations, as discussed above. In addition to water quality parameters, the ozone system would also be monitored for operational parameters, including ozone production concentration, off-gas ozone concentration, ozone operating voltage and power consumption, oxygen feed rate, and other parameters (NSF, 2003a). EPA assumed the time needed to measure these operational parameters would not add significantly to the administrative and monitoring labor estimates described below.
Additional costs include labor to design pilot tests, contact vendors, install pilot equipment, collect water quality samples, monitor the operation of the system, analyze the sampling results and adjust the system accordingly, and develop a final report containing recommendations for the design and operation of a full-scale system. EPA used the assumptions noted in Section 3.3 to estimate these labor costs. Exhibit 13 provides a summary of the cost components for bench and pilot studies by system size, for pre-oxidation systems. Exhibit 14 provides a summary of costs for ozone disinfection systems.
Exhibit 12. Sampling Protocol for Ozone Disinfection Pilot Study
                                   Parameter
                                  Feed Water
                                 Treated Water
Temperature
                                 Continuously
                                 Continuously
Dissolved Ozone Residual
                                      --
                             15 times per study[1]
pH
                                     Daily
                                      --
Alkalinity
                                    Weekly
                                      --
TOC
                                    Weekly
                                      --
DOC
                                     Daily
                                      --
Color
                                     Daily
                                     Daily
Turbidity
                                 Continuously
                                 Continuously
Bromide
                                     Once
                                     Once
Bromate
                             15 times per study[1]
                             15 times per study[1]
AOC
                                      --
                                    Once[2]
THMs Formation Potential
                                     Once
                                    Once[2]
HAAs Formation Potential
                                     Once
                                    Once[2]
Iron (total)
                                     Once
                                      --
Manganese (total)
                                     Once
                                      --
Manganese (dissolved) (manganese concentration passing through 0.2um filter)
                                     Once
                                      --
Total Sulfides (groundwater only)
                                     Daily
                                     Daily
DO
                                     Once
                             15 times per study[1]
Calcium
                                    Weekly
                                      --
Total Hardness
                                    Weekly
                                      --
Source: NSF (2003a and 2003b) and Hulsey (2004).
-- = No sample included in generic schedule, TOC = Total Organic Carbon, DOC = Dissolved Organic Carbon, AOC = Assimilable Organic Carbon, DO = Dissolved Oxygen.
1. The dissolved ozone residual and certain other parameters would be measured once for each ozone dosage level. The dosage levels would vary over a wide range in the beginning of the pilot test, then span a narrower range as engineers get closer to determining the optimum dose. EPA assumed that on average, about 15 different doses would be tested in a pilot study.
2. EPA assumed that engineers would test for assimilable organic carbon and THM and HAA formation potential once per study, after determining the optimum dose.

Exhibit 13. Bench and Pilot Study Costs for Ozone Pre-Oxidation Systems (2019 dollars)
                                   Component
                           Small Systems
(<1 mgd)
                           Medium Systems
(1-10 mgd)
                          Large Systems
(>10 mgd)
Bench study
                                    $2,400
                                    $3,600
                                    $4,800
Pilot study - Equipment
                                      NA
                                      NA
                                      NA
Pilot study - Lab analysis
                                      NA
                                      NA
                                      NA
Pilot study - Labor
                                      NA
                                      NA
                                      NA
Total
                                    $2,400
                                    $3,600
                                    $4,800
Lab analysis costs reflect surface water sampling frequency. 
Detail may not add to totals because total amounts are rounded to report costs to the nearest $100.
NA = Not applicable.

Exhibit 14. Bench and Pilot Study Costs for Ozone Disinfection Systems (2019 dollars)
                                   Component
                           Small Systems
(<1 mgd)
                           Medium Systems
(1-10 mgd)
                          Large Systems
(>10 mgd)
Bench study
                                      NA
                                      NA
                                      NA
Pilot study - Equipment
                                    $4,400
                                    $8,300
                                    $8,300
Pilot study - Lab analysis
                                    $16,200
                                    $26,700
                                    $26,700
Pilot study - Labor
                                    $10,000
                                    $11,100
                                    $11,100
Total
                                    $30,500
                                    $46,100
                                    $46,100
Lab analysis costs reflect surface water sampling frequency. 
Detail may not add to totals because total amounts are rounded to report costs to the nearest $100.
NA = Not applicable.

Ultraviolet Disinfection
Ultraviolet (UV) disinfection systems use ultraviolet lamps to inactivate pathogens. UV lamps are typically encased in a lamp sleeve, which maintains optimal operating temperature and protects the lamp from breaking. Lamp sleeves can foul, decreasing the effectiveness of the disinfection process. Fouling can occur on the internal surface of the sleeve as a result of material deposition from temperature and UV light exposure, and on the external surface of the sleeve as a result of compounds in the source water reacting with the sleeve surface (U.S. EPA, 2003c). 
Generally, pilot testing is used to:
Establish or confirm system design factors such as lamp fouling/aging factor, especially when high iron or calcium concentrations or uncommon upstream chemicals are present
Test system reliability (e.g., improve estimation of safety factors for UV doses to ensure log inactivation for large water systems in which such an investment could reduce capital costs)
Evaluate O&M needs through first-hand operating experience (U.S. EPA, 2003c).
As more performance and fouling information is developed, the need for pilot studies could decrease. For example, extensive pilot testing data is available for waters with moderate hardness and low iron content. Therefore, standard cleaning protocols and wiper frequencies (e.g., one sweep every 15 minutes to 1 hour) are adequate, and a pilot study would not be needed (other than to gain first-hand experience or improve estimation of safety factors). However, for water with higher hardness and iron content, a pilot study could be advantageous to evaluate how best to keep sleeves clean to prevent under dosing, which would cause the system to go into alarm frequently. Pilot studies can also be used to determine the effect of uncommon upstream chemicals such as hypochlorite, ferric iron, permanganate, and ozone on the efficiency of the disinfection process (U.S. EPA, 2003c).
Pilot testing may be implemented at larger systems to improve UV dose safety factors. The safety factors are a function of the estimated lamp fouling/aging factors. A lamp aging factor of 0.7 is commonly used; however, for larger systems, it could be economical to more accurately determine this factor to balance operational costs with capital costs (e.g., a lower lamp aging factor will result in less frequent lamp replacement, but might require a larger system to ensure system performance) (U.S. EPA, 2003c).
Pilot testing can also be useful for staff to obtain operational experience with the UV disinfection system prior to implementation, determine staffing/training needs, and to develop a maintenance plan applicable to the system (e.g., lamp changing frequency, timing and procedures for UV intensity sensor calibration, and manual lamp cleaning protocols). If operational experience is the only desired outcome for a pilot study, staff can also gain operational experience through site visits and partnerships with systems already using UV, personal communications with manufacturers, attendance of seminars on UV disinfection operation, and various training programs (U.S. EPA 2003c).
Therefore, smaller systems would most likely use data from other pilot studies and manufacturers to design their UV system and train staff, without a costly pilot study. Operational parameters can always be adjusted once the system is online. For medium and large systems, pilot studies can reduce initial capital costs, as well as O&M costs, by improving design criteria estimation and providing operating experience. Under most circumstances, a single lamp UV system would be sufficient (Neofotistos, 2004; Cotton, 2004). Pilot system sizes vary based on the budget and purpose of the pilot study. For example, a 100 mgd facility installed a single lamp system because they were only concerned with lamp fouling. However, a 20 mgd facility implemented a 6 mgd pilot system because they wanted to test the difference in parameters such as HPC and nitrification between the non-UV effluent and the UV effluent (Neofotistos, 2004). In many cases, facilities do not implement pilot systems due to the increase in available data from systems already in operation (Cotton, 2004). However, EPA conservatively (i.e., erring on the side of higher costs) assumed that larger systems (i.e., greater than 1 mgd) would implement a single lamp UV pilot system for two to six months. EPA obtained a pilot system rental cost estimate from an equipment vendor. EPA also added an estimated cost of freight both ways.
Sampling for a pilot study should include those parameters that are most likely to affect the performance of the UV systems, as well as parameters that indicate the system's level of performance or efficiency. The main considerations for assessing water quality and effectiveness are:
UV absorbance (A254)
fouling potential
lamp fouling and aging factor
upstream chemical impacts.
A254 is a commonly used water quality parameter that characterizes the decrease in the amount of incident light as it passes through a water sample over a specified distance or path length (U.S. EPA, 2003c). A254 should be monitored continuously for about one to two months in order to determine the lamp maintenance schedule and development of fouling over time (U.S. EPA, 2003c). A254 is monitored onsite by a light sensor mounted onto the UV unit, so there are no sampling costs associated with monitoring A254; however, analysis of the results is necessary.
Although fouling is not expected to be a significant problem for most utilities because of the availability of operational information from other treatment plants, the following water quality parameters should be monitored: calcium, iron, alkalinity, hardness, pH, manganese, aluminum, chloride, carbonate, sulfide, and phosphate. Monitoring these parameters will assist in a qualitative assessment of the fouling potential for the UV reactors and assist in determining what cleaning system should be employed. U.S. EPA (2003c) indicates that for normal operation these parameters should be monitored weekly. However, if fouling is not prevalent, the frequency can be reduced over time. For the purposes of estimating costs, EPA assumed feed water would be tested weekly for the duration of the study. 
Lamp fouling/aging factor includes the effects of both sleeve fouling and lamp aging. The lamp fouling/aging factor will be site-specific and based on the assessment of fouling potential and lamp aging information (obtained from the manufacturer). Monitoring for target parameters (e.g., total coliform, fecal coliform, heterotrophic plate count, or challenge microorganisms) is not necessary for a pilot study because performance efficiency does not scale up to full-scale systems, and these parameters should be measured during validation of the UV reactor. Exhibit 15 summarizes frequencies for testing non-target parameters. 
Exhibit 15. Sampling Protocol for UV Fouling Pilot Study
                                   Parameter
                                  Feed Water
                                Particle Count
                                 continuously
                                      pH
                                    weekly
                                  Temperature
                                 continuously
                                   Turbidity
                                 continuously
                               UV-254 absorbance
                                 continuously
                                  Alkalinity
                                    weekly
                                    Calcium
                                    weekly
                                Total Hardness
                                    weekly
                                 Iron (total)
                                    weekly
                             Manganese (total)[1]
                                    weekly
                              Aluminum (total)[1]
                                    weekly
                               Total Sulfides[1]
                                    weekly
                                  Chloride[1]
                                    weekly
                                 Carbonate[1]
                                    weekly
                                 Phosphate[1]
                                    weekly
Source: U.S. EPA (2003c).
1. U.S. EPA (2003c) indicates that these parameters should be monitored to assess fouling; however, frequencies are not specified. The analytical costs are estimated based on monitoring of these parameters at the same frequency as alkalinity, calcium, hardness, iron, and pH.

Note that most common chemicals added prior to disinfection or found naturally in source water such as alum, ammonia, and calcium will not have significant effects on fouling. However, if chemicals such as hypochlorite, ferric iron, permanganate, and ozone are used, additional influent sampling may be necessary to determine their effects on the UV system. 
Additional costs include labor to design pilot tests, contact vendors, install pilot equipment, collect water quality samples, monitor the operation of the system, analyze the sampling results and adjust the system accordingly, and develop a final report containing recommendations for the design and operation of a full-scale system. EPA used the assumptions noted in Section 3.3 to develop the labor estimates for these activities. Exhibit 16 provides a summary of the cost components for bench and pilot studies by system size.
Exhibit 16. Bench and Pilot Study Costs for Ultraviolet Disinfection Systems (2019 dollars)
                                   Component
                           Small Systems
(<1 mgd)
                           Medium Systems
(1-10 mgd)
                          Large Systems
(>10 mgd)
Bench Study
                                      NA
                                      NA
                                      NA
Pilot study - Equipment
                                      NA
                                    $3,800
                                    $9,000
Pilot study - Lab analysis
                                      NA
                                    $1,900
                                    $5,600
Pilot study - Labor
                                      NA
                                    $13,500
                                    $22,800
Total
                                      $0
                                    $19,100
                                    $37,400
Lab analysis costs reflect surface water sampling frequency. 
Detail may not add to totals because total amounts are rounded to report costs to the nearest $100.
NA = Not applicable.
Ultraviolet Advanced Oxidation
UV disinfection systems, when combined with an oxidant, such as chlorine, ozone, or hydrogen peroxide, can function as an advanced oxidation process (AOP). The oxidant is typically added to the water to be treated prior to UV lamps exposure, resulting in the production of hydroxyl radicals, which become the primary oxidizing species. UVAOPs are utilized to remove many different types of organic contaminants, including taste and odor compounds, pharmaceuticals and personal care products (PCPPs).
Generally, pilot testing is used to:
determine oxidant to be used and proper dosage
satisfy regulatory requirements
determine UV intensity needed
determine if residual oxidant quenching is needed.
In some cases, desktop analysis using computer models, followed by bench scale tests, can be sufficient to assess UVAOP performance and guide full-scale design. However, an onsite pilot may be needed if, for example, water quality is variable (Calgon Carbon, 2020). Furthermore, at least some states (e.g., New York) require pilot testing for UVAOP designs (Pech, 2020). Therefore, to be conservative, EPA assumed that all systems would conduct onsite pilot testing.
UVAOP pilot testing commonly involves examining the effect of some or all of the following variables on treatment efficiency:
specific oxidant used (e.g., hydrogen peroxide, hypochlorite)
dosage of the oxidant
pH
UV dosage (controlled by altering the flow rate through the system).
The number of conditions to be tested (i.e., combinations of these variables) affects both the duration of the pilot study and the analytical requirements. Pilot studies described in the literature had a duration of two to 12 months and tested from four to 30 different conditions (Chuang et al., 2019; Glover et al., 2018; Lekkerkerker-Teunissen et al. 2013; Robinson, 2016; Rosenfeldt et al., 2013; Sarathy et al., 2011; Wang et al., 2015a and 2015b; Zhang et al., 2019). The longer studies, however, were associated with potable reuse projects. UVAOP vendors confirmed that reuse pilot studies can last up to 12 months; studies for drinking water applications typically can test the desired number of conditions in two weeks to four months (Diefenthal, 2020; Pech, 2020). Therefore, EPA assumed pilot studies would last two weeks for small systems, two months for medium systems, and four months for large systems. EPA assumed the number of treatment conditions tested would be four for small systems, eight for medium systems, and 16 for large systems.
Analytes tested during a UVAOP pilot will vary depending on influent water quality and the specific oxidant(s) being evaluated. It is advisable to monitor disinfection by-products and their precursors, such as bromate or nitrate, and bromide or nitrite, respectively (Linden, 2015). It is also typical to monitor parameters that indicate or affect the ability of light to pass through the water, including turbidity and UV-254 absorbance. Pilot studies designed to evaluate the need for post-treatment quenching will need to monitor the level of residual oxidant in treated water (Wang et al., 2015; Roccaro, 2016). Although not all pilot tests would evaluate multiple oxidants, EPA included monitoring for both free chlorine (as an indicator of hypochlorite as a residual oxidant) and pH (as an indicator of oxidation-reduction potential for monitoring residual hydrogen peroxide) in the sampling protocol. EPA assumed two samples each of feed water and treated water per treatment condition studied. Exhibit 17 shows the sampling protocol that EPA used as a basis for estimating costs of analyzing water samples.
Exhibit 17. Sampling Assumptions for Ultraviolet Advanced Oxidation
                                   Parameter
                                  Feed Water
                                 Treated Water
Alkalinity
                             2 times per condition
                             2 times per condition
Bromide
                             2 times per condition
                             2 times per condition
Bromate
                             2 times per condition
                             2 times per condition
Free Chlorine (online)
                             Continuous monitoring
                             Continuous monitoring
Nitrate
                             2 times per condition
                             2 times per condition
Nitrite
                             2 times per condition
                             2 times per condition
pH
                             2 times per condition
                             2 times per condition
TOC
                             2 times per condition
                             2 times per condition
Temperature (online)
                             Continuous monitoring
                             Continuous monitoring
Turbidity
                             2 times per condition
                             2 times per condition
UV-254 Absorbance
                             2 times per condition
                             2 times per condition
Target Contaminant
                             2 times per condition
                             2 times per condition
TOC = Total Organic Carbon

UVAOP pilot equipment is available for purchase or rent from companies that manufacture full-scale UV equipment. EPA collected rental cost estimates from these vendors. The rental costs of the pilot systems vary depending on their capabilities, specifically the number of different oxidants and other conditions they can accommodate. Because they would test fewer conditions, EPA assumed small systems would rent a simpler, lower cost pilot unit. EPA assumed medium and large systems would rent a more flexible, higher cost unit.
Additional costs include labor to design pilot tests, contact vendors, install pilot equipment, collect water quality samples, monitor the operation of the system, analyze the sampling results and adjust the system accordingly, and develop a final report containing recommendations for the design and operation of a full-scale system. EPA used the assumptions noted in Section 3.3 to develop the labor estimates for these activities. Exhibit 18 provides a summary of the cost components for pilot studies by system size.
Exhibit 18. Bench and Pilot Study Costs for Ultraviolet Advanced Oxidation Systems (2019 dollars)
                                   Component
                           Small Systems
(<1 mgd)
                           Medium Systems
(1-10 mgd)
                          Large Systems
(>10 mgd)
Bench study
                                      NA
                                      NA
                                      NA
Pilot study - Equipment
                                    $3,900
                                    $27,200
                                    $53,200
Pilot study - Lab analysis
                                    $9,300
                                    $18,600
                                    $37,300
Pilot study - Labor
                                    $10,000
                                    $13,500
                                    $18,200
Total
                                    $23,200
                                    $59,300
                                   $108,600
Lab analysis costs reflect groundwater sampling frequency. 
Detail may not add to totals because total amounts are rounded to report costs to the nearest $100.
NA = Not applicable.

Aeration
WBS aeration models include packed tower aeration (PTA), multi-stage bubble aeration (MSBA), tray aeration, diffuse aeration, and spray aeration. These technologies are used to remove volatile organic chemicals and gases such as radon, which are generally only found in groundwater. All of the aeration technologies rely on contact between air and water to remove these contaminants. However, as discussed below, pilot studies typically are not required for PTA. They can be necessary for the other aeration technologies. Therefore, this analysis considers pilot costs separately for the following categories of aeration:
PTA
low-profile aeration, which covers MSBA, tray aeration, and diffuse aeration
spray aeration.
Packed Tower Aeration (PTA)
According to NAS (1997), the performance of PTA can be predicted with design equations that are based on hundreds of installations using a variety of source waters. Therefore, site-specific pilot testing for the purpose of designing the aeration system should not be necessary (NAS, 1997). However, NAS (1997) notes that designers must pay careful attention to possible foulants, such as reduced iron and manganese. Similarly, Mead and Leibbert (1998) warn that PTA columns can become fouled from iron, calcium, manganese, or biological growth. However, the authors do not prescribe a pilot test to diagnose how to address potential fouling; instead, they note that strippers must be cleaned periodically, either by washing with an acid solution or by removing the packing and cleaning it separately. Similarly, a manufacturer of packing material and associated products for packed tower aeration advises using a cleaning solution or adding an antiscaling agent to the feed water, and suggests that such chemical treatments are easily integrated after the towers have been installed (Jaeger Products, Inc., 2003). 
Aeration can also affect downstream processes or distribution systems. For example, aeration increases the content of dissolved oxygen, which usually (in the pH ranges of most source waters) promotes oxidation of iron, which could affect some downstream processes (Montana University System Water Center, 2001). Aeration also results in the removal of natural carbon dioxide, which can raise the pH and reduce the corrosivity of water. However, for PTA, impacts like the increase in dissolved oxygen and removal of carbon dioxide are well understood and can be predicted with design equations. Therefore, pilot testing to evaluate downstream effects is unlikely except in very site-specific cases (e.g., where there are significant downstream treatment processes or distribution system concerns).
As a result of these factors, even though some systems that install PTA in response to a regulatory requirement might use a pilot test (e.g., as an investment risk reduction tool, to convince consumers the treatment is effective, to assess cleaning requirements from potential foulants, or to assess downstream impacts), EPA believes that, in general, systems would not pilot test PTA. Instead, EPA assumed that systems would require only a desktop evaluation based on the established engineering design equations plus engineering or vendor professional judgement regarding fouling and downstream impacts.
For the desktop analysis costs for PTA systems, EPA assumed eight hours of labor time for a civil engineer. This labor includes the time taken by a system engineer to send data on water quality parameters to a process design company or vendor, the costs of the computer study, and time to report the results of the study. EPA included this cost for PTA systems of all sizes. Exhibit 19 provides a summary of the cost components for bench and pilot studies by system size.
Exhibit 19. Bench and Pilot Study Costs for Packed Tower Aeration Systems (2019 dollars)
                                   Component
                          Small Systems 
(<1 mgd)
                           Medium Systems
(1-10 mgd)
                          Large Systems
(>10 mgd)
Desktop Analysis
                                     $600
                                     $600
                                     $600
Total
                                     $600
                                     $600
                                     $600
Amounts are rounded to report costs to the nearest $100.

Low-Profile Aeration
Manufacturers of low-profile aeration equipment typically have models or design equations for predicting system performance (Ahmed et al., 2014; USACE, 2001; Pico, 1997). However, these design methods usually are proprietary and/or manufacturer-specific, due to differences among manufacturers in factors such as diffuser characteristics (in diffuse aeration and MSBA), baffle arrangement (in MSBA), and tray design and configuration (in tray aeration). While generic design equations exist for low-profile technologies (see, for example, El-Behlil and Adma, 2012; Crittenden et al., 2012; AWWA, 2011; Hokanson et al., 1999), these design equations are not as well-established as those for PTA and performance in the field can deviate from their predictions because of manufacturer-specific factors such as those identified above. Predictions of downstream impacts (e.g., iron oxidation, increase in treated water pH) suffer from the same limitations.
Like PTA, low-profile aeration systems can be subject to scaling and fouling. However, these systems are somewhat less susceptible to scaling and fouling and are physically easier to clean (Ahmed et al., 2014; USACE, 2001; NWRI, 2000). In addition, like PTA, chemical treatment should be easy to integrate after installation of the system. Thus, pilot tests rarely are needed to address these issues.
Review of the literature shows that, while not always necessary for low-profile aeration technologies, pilot or bench tests are sometimes conducted (USACE, 2001; Lowry and Lowry, 2000; Pico, 1997; Kinner et al., 1988; Sherant, 2008; Brooke and Collins, 2011; Billelo and Singley, 1986; Alabdula'aly and Maghrawy, 1999; Ahmed et al., 2014; U.S. EPA, 1983). The purposes of these tests can include the following:
Verifying manufacturer predictions of performance
Optimizing design and operating parameters such as air-to-water ratio (for all low-profile technologies), number of stages (for MSBA) or trays (for tray aeration), and/or contact time and water depth (for diffuse aeration)
Evaluating downstream impacts.
For pilot-scale studies for tray aeration and MSBA, EPA used equipment rental cost estimates for very small units of the size typically used for pilot testing from manufacturers and vendors. EPA also added an estimated cost of freight both ways. Because diffuse aeration systems typically are custom-constructed onsite, rental units are not available. Diffuse aeration pilot tests can be accomplished by temporarily installing diffusers in existing tanks (e.g., Sherant, 2008) or using bench-scale equipment (e.g., Sherant 2008; Brooke and Collins, 2011; U.S. EPA, 1983). For diffuse aeration equipment costs, EPA conservatively assumed an existing tank would not be available and bench-scale equipment would need to be purchased. Accordingly, EPA calculated the cost for a one-time purchase of two diffusers, a small vessel, and instrumentation to measure temperature. EPA assumed the purchased diffusers and vessel would be smaller than that used in the final process and, thus, not reusable. Final process costs also do not include temperature monitoring equipment. Therefore, it is appropriate to include these costs in the pilot study costs. For tray aeration and MSBA, EPA also added the cost to purchase temperature instrumentation, because it was not clear that the equipment available for rental included this item (and it is not included in the final process costs).
Because aeration systems reach steady-state operation quickly, individual test runs of these systems are short, with durations ranging from less than 30 minutes (Billelo and Singley, 1986) to one to two hours (Brooke and Collins, 2011; Kinner et al., 1988) to six to nine hours (Lowry and Lowry, 2002; Alabdula'aly and Maghrawy, 1999). As discussed below, pilot tests usually include multiple test runs. Even with multiple runs, however, the total testing time remains short (days to weeks). There are a few examples in the pilot testing literature studying ongoing operation of low-profile aeration systems, but even these lasted less than 30 days (Sherant, 2008; Kinner et al., 1988). Therefore, for small systems, EPA assumed pilot tests would last two weeks. For medium and large systems, EPA assumed pilot test duration of one month.
Pilot tests of low-profile aeration often include multiple test runs using different design parameters and operating conditions (e.g., combinations of air-to-water ratio and contact time). When multiple runs are used, the number of conditions (e.g., sets of design/operating parameters) tested can range from approximately a dozen (Pico, 1997; Kinner et al., 1988; Ahmed et al., 2014) to 24 to 32 (Billelo and Singley, 1986; Alabdula'aly and Maghrawy, 1999). For small systems, EPA assumed the purpose of the pilot test would be primarily to verify the vendor's performance claims and evaluate potential downstream impacts, and encompass only one set of conditions (the vendor's recommended conditions). EPA assumed medium and large systems would test multiple sets of conditions with the additional goal of optimizing design and operating parameters. Medium systems would test 12 sets of conditions and large systems would test 32 sets of conditions. 
The assumptions about the number of test runs affect the analytical costs only (labor and equipment costs are based on the test duration) and, for MSBA and tray aeration, are also intended to account for intermediate sampling to evaluate the number of stages or trays. Exhibit 20 shows the sampling protocol that EPA used as a basis for estimating costs of analyzing water samples. 
Additional costs include labor to design the test, contact vendors, install equipment, collect water quality samples, monitor the operation of the system, analyze the sampling results and adjust the system accordingly, and develop a final report containing recommendations for the design and operation of a full-scale system. For labor, EPA used the assumptions noted in Section 3.3. Exhibit 21 provides a summary of the cost components for bench and pilot studies by system size.
Exhibit 20. Sampling Assumptions for Low-Profile Aeration Bench and Pilot Tests
                                   Parameter
                                  Feed Water
                                 Treated Water
Alkalinity
                              Twice per study[1]
                              Twice per study[1]
Calcium
                              Twice per study[1]
                              Twice per study[1]
Chloride
                              Twice per study[1]
                              Twice per study[1]
DO
                           Once per condition tested
                           Once per condition tested
Hardness
                              Twice per study[1]
                              Twice per study[1]
HPC
                              Twice per study[1]
                              Twice per study[1]
Iron (total)
                           Once per condition tested
                           Once per condition tested
Iron (dissolved)
                           Once per condition tested
                           Once per condition tested
Magnesium
                              Twice per study[1]
                              Twice per study[1]
Manganese (total)
                           Once per condition tested
                           Once per condition tested
Manganese (dissolved)
                           Once per condition tested
                           Once per condition tested
pH
                           Once per condition tested
                           Once per condition tested
Sulfate
                              Twice per study[1]
                              Twice per study[1]
TDS
                           Once per condition tested
                           Once per condition tested
Temperature
                                  Continuous
                                  Continuous
TOC
                              Twice per study[1]
                              Twice per study[1]
Target Contaminant
                      Three times per condition tested[2]
                      Three times per condition tested[2]
Sources: adapted from Lowry and Lowry (2002), Kinner et al. (1988) and Alabdula'aly and Maghrawy (1999).
DO = Dissolved Oxygen, HPC = Heterotrophic Plate Count, TDS = Total Dissolved Solids, TOC = Total Organic Carbon.
1. Pre- and post-testing.
2. Includes verifying that steady-state has been reached.

Exhibit 21. Bench and Pilot Study Costs for Low-Profile Aeration Systems (2019 dollars)
                                   Component
                           Small Systems
(<1 mgd)
                           Medium Systems
(1-10 mgd)
                          Large Systems
(>10 mgd)
MSBA/Tray Aeration Systems
Bench study
                                      NA
                                      NA
                                      NA
Pilot study - Equipment
                                    $3,400
                                    $4,300
                                    $4,300
Pilot study - Lab analysis
                                    $2,300
                                    $19,200
                                    $49,900
Pilot study - Labor
                                    $10,000
                                    $11,100
                                    $11,100
Total
                                    $15,600
                                    $34,600
                                    $65,400
Diffuse Aeration Systems
Bench Study
                                      NA
                                      NA
                                      NA
Pilot study - Equipment
                                    $2,000
                                    $2,000
                                    $2,000
Pilot study - Lab analysis
                                    $2,300
                                    $19,200
                                    $49,900
Pilot study - Labor
                                    $10,000
                                    $11,100
                                    $11,100
Total
                                    $14,300
                                    $32,300
                                    $63,100
Lab analysis costs reflect groundwater sampling frequency. 
Detail may not add to totals because total amounts are rounded to report costs to the nearest $100.
NA = Not applicable.
 
Spray Aeration
Drinking water systems sometimes install spray aeration systems in pre-existing storage tanks to remove volatile contaminants, most commonly TTHMs, in the distribution system. The performance of spray aeration can be predicted using computer models and design equations (e.g., Brooke and Collins, 2011; Cecchetti et al., 2014). Systems often rely on desktop analysis using these tools, along with vendor recommendations, to determine appropriate design and operating parameters (Fiske et al., 2011; Medora Corporation, 2014a; PAX Water Technologies, 2012; Reed, 2013; Tseng and Kim, 2014). Desktop analysis alone may be sufficient to guide full-scale installation (Medora Corporation, 2014a), but it is also common to follow this analysis with additional testing. The most common approach is to install a spay aeration system in a single full-scale tank for evaluation prior to adopting the technology throughout the distribution system (Clunie et al., 2014; Fiske, 2013; Medora Corporation, 2014b; PAX Water Technologies, 2012; Reed, 2013; Rosenfeldt et al., 2014; Tseng and Kim, 2014; Williams, 2013). There are only a few examples of pilot studies that included bench- or small-scale tests (e.g., using an aquarium tank or small pipe to simulate a full-scale tank) (Spacek, 2009; Walfoort et al., 2008; Williams, 2013). 
Based on the literature, EPA assumed that small systems could rely on desktop analysis alone. Pilot costs for desktop analysis incorporate eight hours of labor time for a civil engineer. This labor includes the time taken by a system engineer to send data on water quality parameters to a process design company or vendor, the costs of modeling performance based on the data, and time to report the results of the study.
EPA assumed that medium and large systems, which are more likely to have multiple storage tanks, would follow desktop analysis with testing in a single existing tank before deploying more widely. The purpose of this testing could include the following:
Verifying modeled or vendor-predicted performance
Optimizing air-to-water ratio
Examining the effects of ventilation and/or mixing on performance
Determining optimal aerator placement (e.g., with respect to tank walls).
The duration of tank testing can range from three to 10 months (Clunie et al., 2014; Fiske et al., 2011; Reed, 2013; Rosenfeldt et al., 2014; Tseng and Kim, 2014; Williams, 2013). Typically, the duration of this phase is selected to cover the system's season of highest TTHM concentrations, although it also can depend on the number of operating conditions to be examined. EPA assumed tank testing would last four months for medium systems and six months for large systems. 
Systems typically do not rent equipment for pilot testing. Instead, they purchase components (e.g., pipes, pumps, nozzles) and fabricate the aeration system in-house (Fiske, 2013; Reed, 2013), purchase a package unit for the study tank from a vendor (Medora Corporation, 2014b; PAX Water Technologies, 2012), receive a sample unit from a vendor with an option to buy at the end of the pilot (Williams, 2013), or use some combination of these approaches (Clunie et al., 2014; Tseng and Kim, 2014). In all of these cases, the pilot equipment ultimately becomes part of the final installed full-scale aeration process. Therefore, EPA excluded equipment costs from the spray aeration pilot study cost estimate to prevent double counting capital expenditures that are already included in the WBS engineering model for spray aeration.
During tank testing, systems compare analytical results for spray aeration to control results. The control results can be for the same tank during periods with the aeration system turned off (e.g., Rosenfeldt et al., 2014; Tseng and Kim, 2014) or for similar tanks in the system that do not have aeration installed (e.g., Clunie et al., 2014; Medora Corporation, 2014b). Regardless of the source of the control samples, Exhibit 22 shows the sampling protocol that EPA used as a basis for estimating costs of analyzing water samples.
Exhibit 22. Sampling Assumptions for Spray Aeration Pilot Tests
                                   Parameter
                                    Control
                              With Spray Aeration
Free Chlorine
                                    Weekly
                                    Weekly
pH
                                    Weekly
                                    Weekly
Temperature (online)
                                    Weekly
                                    Weekly
TOC
                                    Weekly
                                    Weekly
Target Contaminant (e.g., TTHMs)
                                    Weekly
                                    Weekly
Sources: based on Clunie et al. (2014), Gosh et al. (2015), Reed (2013), Tseng and Kim (2014), and Williams (2013).
TOC = total organic carbon; TTHMs = total trihalomethanes

Additional costs include labor to design the test, contact vendors, install equipment, collect water quality samples, monitor the operation of the system, analyze the sampling results and adjust the system accordingly, and develop a final report containing recommendations for the design and operation of a full-scale system. For labor, EPA used the assumptions noted in Section 3.3. Exhibit 23 provides a summary of the cost components for bench and pilot studies by system size.
Exhibit 23. Bench and Pilot Study Costs for Spray Aeration Systems (2019 dollars)
                                   Component
                           Small Systems
(<1 mgd)
                           Medium Systems
(1-10 mgd)
                          Large Systems
(>10 mgd)
Desktop analysis
                                     $600
                                     $600
                                     $600
Pilot study - Equipment[1]
                                      NA
                                      NA
                                      NA
Pilot study - Lab analysis
                                      NA
                                    $8,600
                                    $12,800
Pilot study - Labor
                                      NA
                                    $18,200
                                    $22,800
Total
                                     $600
                                    $27,300
                                    $36,300
Lab analysis costs reflect sampling in water storage tanks. 
Detail may not add to totals because total amounts are rounded to report costs to the nearest $100.
NA = Not applicable.
1. Equipment cost are not included to prevent double counting. See text for further discussion.

Biological Treatment
The WBS biological treatment model can design and cost the following three types of biological treatment systems:
Systems where the bioreactor is a pressure vessel with a fixed media bed
Systems where the bioreactor is an open concrete basin with a fixed media bed
Systems where the bioreactor is a pressure vessel with a fluidized media bed.
Although bench-scale studies for biological treatment processes are occasionally conducted simultaneously with pilot testing to evaluate removal mechanisms and/or guide piloting decisions, they are not typical (Brown, 2011). For pilot systems, EPA obtained equipment rental costs from a company that leases biological pilot testing equipment. These equipment rental prices do not include the cost of pilot test consumables, which include acetic acid, phosphoric acid, hydrogen peroxide, and polymers. Consumption rates for these consumables were extrapolated from pilot test operating data from a 10-month demonstration study conducted to treat perchlorate-contaminated groundwater using fixed-bed bioreactor technology, as reported in U.S. DoD (2008). 
A biological treatment pilot study can range in duration from four to 10 months, regardless of the size of the system being tested. The pilot test duration depends on the target contaminants (Brown, 2011). For the purpose of estimating average costs, EPA calculated costs of biological treatment pilot scale tests based on the following assumptions:
bench scale tests are not typically conducted
small systems (with design flows less than 1 mgd) would run a 4-month pilot test using equipment on the scale of a 6 gpm system
medium systems (with design flows of 1 to 10 mgd) would run a 6-month pilot test using equipment on the scale of a 25 gpm system
large systems (with design flows over 10 mgd) would run a 10-month pilot test using equipment on the scale of a 25 gpm system.
A typical sampling protocol for a biological treatment pilot test includes sampling the influent, bioreactor effluent, post-treatment effluent, and disinfection tank. The specific analytes depend on the study objectives, contaminants of concern, and other variables. Exhibit 24 shows the sampling protocol that EPA used as a basis for estimating costs of analyzing water samples. Sampling parameters and frequency are based on sampling records for the aforementioned 10-month fixed-bed bioreactor demonstration study, where perchlorate was the target contaminant (U.S. DoD, 2008).
Additional costs include labor to design pilot tests, contact vendors, install pilot equipment, collect water quality samples, monitor the operation of the system, analyze the sampling results and adjust the system accordingly, and develop a final report containing recommendations for the design and operation of a full-scale system. For labor, EPA obtained hour estimates for initial setup of biological treatment pilot test equipment (60 hours) and monthly pilot test operation (184 hours) from the company that provided the pilot equipment rental costs (Brown, 2011), and used the labor cost per hour for an engineer as discussed in Section 3.3. Exhibit 25 provides a summary of the cost components for pilot studies by system size.
Exhibit 24. Target and Non-Target Contaminant Sampling Parameters and Frequencies for Biological Treatment Pilot Study
                                   Parameter
                                  Feed Water
                              Bioreactor Effluent
                            Post-Treatment Effluent
                             Chlorine Contact Tank
Target Contaminant
                              24 times per month
                                       
                                       
                                       
Ammonia
                                   bi-weekly
                                   bi-weekly
                                      --
                                      --
BOD
                               3 times per study
                                       
                                       
                                       
BDOC
                                   bi-weekly
                                    weekly
                                    weekly
                                      --
Chlorate
                                      --
                                    weekly
                                      --
                                      --
Chlorite
                                      --
                                    weekly
                                      --
                                      --
DOC
                                twice per week
                                twice per week
                                twice per week
                                twice per week
DO
                                  continuous
                                  continuous
                                  continuous
                                  continuous
Fecal Coliform
                                    weekly
                                    weekly
                                    weekly
                                    weekly
Free Chlorine (online)
                                      --
                                      --
                                      --
                               3 times per week
HAAs
                              10 times per study
                                       
                                       
                                       
HPC
                                    weekly
                                    weekly
                                    weekly
                                    weekly
Hydrogen Sulfide (H2S)
                                      --
                                    weekly
                                    weekly
                                      --
Iron (total)
                                    monthly
                                    monthly
                                    monthly
                                      --
Manganese (total)
                                    monthly
                                    monthly
                                    monthly
                                      --
Nitrate
                               3 times per week
                               3 times per week
                                      --
                                      --
Nitrite
                                    weekly
                                    weekly
                                      --
                                      --
pH
                                  continuous
                                  continuous
                                  continuous
                                  continuous
Phosphate
                                   bi-weekly
                                   bi-weekly
                                      --
                                      --
Sulfate
                                    weekly
                                    weekly
                                      --
                                      --
TDS
                               3 times per study
                                       
                                       
                                       
Temperature
                                  continuous
                                      --
                                      --
                                      --
Total Coliform
                                    weekly
                                    weekly
                                    weekly
                                    weekly
TSS
                               3 times per study
                                       
                                       
                                       
TTHMs
                              10 times per study
                                       
                                       
                                       
Turbidity
                                  continuous
                                  continuous
                                  continuous
                                  Continuous
Volatile Suspended Solids (VSS)
                               3 times per study
                                       
                                       
                                       
Source: U.S. DoD (2008).
-- = Not sampled. BOD = biochemical oxygen demand; BDOC = biodegradable organic carbon; DOC = dissolved organic carbon; DO = dissolved oxygen; HAAs = haloacetic acids; HPC = heterotrophic plate count; TDS = total dissolved solids; TSS = total suspended solids; TTHMs = total trihalomethanes

Exhibit 25. Bench and Pilot Study Costs for Biological Treatment Systems (2019 dollars)
                                   Component
                          Small Systems 
(<1 mgd)
                           Medium Systems
(1-10 mgd)
                          Large Systems
(>10 mgd)
Bench Study
                                      NA
                                      NA
                                      NA
Pilot study - Equipment
                                    $16,000
                                    $41,500
                                    $69,200
Pilot study - Lab analysis (target contaminant)
                                    $19,200
                                    $28,800
                                    $48,000
Pilot study - Lab analysis (non-target contaminant)
                                    $34,500
                                    $49,800
                                    $80,400
Pilot study - Labor
                                    $58,300
                                    $85,200
                                   $139,100
Total
                                   $128,000
                                   $205,300
                                   $336,700
Lab analysis costs reflect groundwater sampling frequency. 
Detail may not add to totals because total amounts are rounded to report costs to the nearest $100.
NA = Not applicable.

Biologically Active Filtration
Drinking water systems use biologically active filtration (BAF) to remove various organic compounds by allowing (or even encouraging) the growth of naturally occurring microorganisms on filter media. Thus, BAF removes natural organic matter both through filtration and as a result of consumption or metabolization by the microorganisms. In addition to natural organic matter in general, BAF treatment also sometimes targets more specific compounds including specific disinfection byproduct precursors, taste and odor compounds, cyanotoxins, trace organic compounds, and/or manganese (AWWA, 2011; Carter et al., 2014; Evans et al., 2016; Gerrity et al., 2011; Jakubowski et al., 2011; Rattier et al., 2012; Sawade et al., 2012; Schneider and LeChevallier, 2017; Selbes et al., 2017).
Generally, pilot testing is used to (Carter et al., 2014; Evans et al., 2016; Song, 2016; Urfer et al., 1997):
Confirm consistent removal effectiveness (particularly given seasonal changes in water quality and temperature and following shut-downs)
Demonstrate that a proposed design meets regulatory requirements
Evaluate the effect of various design and operating parameters on performance including:
    Empty bed contact time
    Media type
    Backwash conditions
    Varying pre-treatment (coagulant addition or pH adjustment) and/or pre-oxidation (ozone or aeration) conditions
    Supplemental nutrient addition.
Some of the studies described in the literature (e.g., Krasner et al., 1993; Schneider and LeChevllier, 2017) are designed to guide conversion of existing filters to biological filters. These evaluations are typically full-scale demonstrations where a filter is allowed to remain in service for an extended time to encourage biological growth, then compared to control filters subject to conventional chlorination practices and/or media replacement frequency. The WBS engineering model, on the other hand, considers a scenario where BAF is designed and installed as new construction. There are also examples in the literature of BAF pilot testing as part of a proposed potable reuse train (e.g., Gerrity et al., 2011; Hwang et al., 2016). These studies tend to use large-scale, permanent or semi-permanent demonstration facilities that test results of a full reuse treatment train. 
Studies designed to guide BAF as new construction in typical drinking water applications use pilot-scale columns and last from three to 24 months, with a median duration of eight months (Brown et al., 2016; Chakmak et al., 2015; Evans et al., 2016; Jakubowski et al., 2011; Page et al., 2016; Song, 2016). They are sometimes preceded by a desktop evaluation to identify pilot testing parameters (Jakubowski et al., 2011; Song, 2016). Note that BAF pilot study durations must incorporate sufficient acclimation time for microbial growth to establish and reach steady-state conditions (Song, 2016; Urfer et al., 1997). Based on lessons learned from facilities participating in the Biofiltration Knowledge Base survey, Evans et al. (2019) recommend a total pilot study duration of six to 12 months.
EPA assumed a pilot study duration for large systems of eight months (including acclimation period), the median reported in the literature pertaining to BAF as new construction in typical drinking water applications and within the range recommended by Evans et al. (2019). Because all of the examples in the literature were for large systems, EPA assumed that pilot study duration for small and medium systems would be six months (including acclimation period), the minimum recommended in Evans et al. (2019). EPA also assumed:
Systems would conduct a desktop evaluation prior to the pilot study 
Based on the desktop analysis, small systems would test the single, most promising treatment approach
Medium systems would evaluate two different treatment conditions 
Large systems would evaluate four different treatment conditions
The pilot study duration would include a two-month acclimation period.
Exhibit 26 shows the sampling protocol that EPA used as a basis for estimating costs of analyzing water samples. Note that this protocol assumes the target contaminant is a trace organic. Where the treatment target is manganese, disinfectant byproduct precursors, or natural organic matter in general, this protocol will result in a conservative estimate of analytical costs (erring on the side of higher cost), because measures of these treatment targets are included among the routine monitoring parameters.
Exhibit 26. Sampling Assumptions for Biologically Active Filtration Pilot Tests
                                   Parameter
                                  Feed Water
                                 Treated Water
              Acclimation Period (2 months for all system sizes)
Nitrate
                                    Weekly
                             Weekly per condition
Nitrite
                                    Weekly
                             Weekly per condition
TOC
                                    Weekly
                             Weekly per condition
Testing Period (4 months for small and medium systems; 6 months for large systems)
Alkalinity
                                 Twice monthly
                          Twice monthly per condition
Aluminum
                                 Twice monthly
                          Twice monthly per condition
Ammonia
                                 Twice monthly
                          Twice monthly per condition
AOC
                                 Twice monthly
                          Twice monthly per condition
ATP biomass in treatment media
                                      --
                        Four times total per condition
Bromate
                                 Twice monthly
                          Twice monthly per condition
Bromide
                                 Twice monthly
                          Twice monthly per condition
Chlorine demand (simulated distribution system)
                                 Twice monthly
                          Twice monthly per condition
Color
                                 Twice monthly
                          Twice monthly per condition
DO
                                 Twice monthly
                          Twice monthly per condition
DOC
                                 Twice monthly
                          Twice monthly per condition
HAAs (simulated distribution system)
                                 Twice monthly
                          Twice monthly per condition
Hardness
                                 Twice monthly
                          Twice monthly per condition
Heterotrophic plate count
                                 Twice monthly
                          Twice monthly per condition
Iron
                                 Twice monthly
                          Twice monthly per condition
Manganese
                                 Twice monthly
                          Twice monthly per condition
Nitrate
                                 Twice monthly
                          Twice monthly per condition
Nitrite
                                 Twice monthly
                          Twice monthly per condition
pH
                                  Continuous
                                  Continuous
Phosphate
                                 Twice monthly
                          Twice monthly per condition
Sulfide
                                 Twice monthly
                          Twice monthly per condition
Taste and odor compounds
                                 Twice monthly
                          Twice monthly per condition
Temperature
                                  Continuous
                                  Continuous
TOC
                                    Weekly
                             Weekly per condition
TTHMs (simulated distribution system)
                                 Twice monthly
                          Twice monthly per condition
Turbidity
                                  Continuous
                                  Continuous
UV-254 absorbance
                                    Weekly
                             Weekly per condition
Target contaminant (e.g., trace organics)
                                 Twice monthly
                          Twice monthly per condition
Sources: based on Chakmak et al., 2015; Evans et al., 2016; Halevy et al., 2013; Jakubowski et al., 2011; Page et al., 2016; Song, 2016; Urfer et al., 1997.
AOC = assimilable organic carbon; ATP = adenosine triphosphate; DO = dissolved oxygen; DOC = dissolved organic carbon; HAAs = haloacetic acids; TOC = total organic carbon; TTHMs = total trihalomethanes

For equipment costs, EPA used rental cost estimates for a skid-mounted pilot filtration system from a vendor that provides pilot testing services. Additional costs include labor to design the test, contact vendors, install equipment, collect water quality samples, monitor the operation of the system, analyze the sampling results and adjust the system accordingly, and develop a final report containing recommendations for the design and operation of a full-scale system. For labor, EPA used the assumptions noted in Section 3.3, plus eight hours of labor time for a civil engineer to conduct the initial desktop analysis. Exhibit 27 provides a summary of the cost components for pilot studies by system size.
Exhibit 27. Bench and Pilot Study Costs for Biologically Active Filtration Systems (2019 dollars)
                                   Component
                           Small Systems
(<1 mgd)
                           Medium Systems
(1-10 mgd)
                          Large Systems
(>10 mgd)
Desktop analysis
                                     $600
                                     $600
                                     $600
Pilot study - Equipment
                                    $45,700
                                    $45,700
                                    $60,500
Pilot study - Lab analysis
                                    $39,000
                                    $58,600
                                   $144,300
Pilot study - Labor
                                    $22,800
                                    $22,800
                                    $27,500
Total
                                   $108,100
                                   $127,700
                                   $232,900
Lab analysis costs reflect surface water sampling frequency. 
Detail may not add to totals because total amounts are rounded to report costs to the nearest $100.
NA = Not applicable.

Summary and Discussion
This paper describes the method and assumptions used to estimate pilot and bench study costs for use in EPA's WBS models, which estimate costs for drinking water treatment facilities to implement new technologies to comply with new regulations. The costs, test durations, and sampling protocols used here do not correspond to a specific target contaminant or rulemaking. For a given target contaminant or rulemaking, it might be appropriate to change certain key assumptions. Adjustments that might be appropriate could include changes to the assumption of a $200 sampling cost for a target parameter, the protocol for nontarget parameter samples, and the assumptions relating to labor time and rates. However, because bench and pilot study costs are a small component of the overall compliance costs, the impact of such rule-specific changes would likely be minimal.
Exhibit 28 provides a summary of total bench and pilot study cost estimates in 2014 dollars by technology category and system size. The WBS models include these bench and pilot testing costs as an add-on cost. The models escalate these costs to current year dollars using the Consumer Price Index for equipment rental the Bureau of Labor Statistics Employment Cost Index for labor and laboratory analysis.
Exhibit 28. Estimated Bench and Pilot Study Costs for the WBS Models (2019 dollars)
                                  Technology
                          Small Systems 
(<1 mgd)
                           Medium Systems
(1-10 mgd)
                          Large Systems
(>10 mgd)
Adsorptive Media
                                   $15,000 
                                   $67,000 
                                   $67,000 
Lime Softening
                                     $600 
                                     $600 
                                     $600 
Membranes (MF/UF)
                                   $55,900 
                                   $100,600 
                                   $207,400 
Membranes (RO/NF)
                                    $1,200 
                                   $63,300 
                                   $131,600 
Membranes (EDR)
                                     $800 
                                   $63,900 
                                   $134,700 
Conventional Filtration / Direct Filtration
                                    $2,600 
                                   $82,700 
                                   $152,100 
Ozone Pre-oxidation
                                    $2,400 
                                    $3,600 
                                    $4,800 
Ozone Disinfection
                                   $30,500 
                                   $46,100 
                                   $46,100 
Ultraviolet Disinfection
                                      $0 
                                   $19,100 
                                   $37,400 
Ultraviolet Advanced Oxidation (UVAOP)
                                   $23,200 
                                   $59,300 
                                   $108,600 
Aeration (PTA)
                                     $600 
                                     $600 
                                     $600 
Aeration (MSBA/Tray)
                                   $15,600 
                                   $34,600 
                                   $65,400 
Aeration (Diffuse)
                                   $14,300 
                                   $32,300 
                                   $63,100 
Aeration (Spray)
                                     $600 
                                   $27,300 
                                   $36,300 
Biological Treatment
                                   $128,000 
                                   $205,300 
                                   $336,700 
Biologically Active Filtration (BAF)
                                   $108,100 
                                   $127,700 
                                   $232,900 
Amounts are rounded to report costs to the nearest $100.

References
Alabdula'aly, A.I. and H.B. Maghrawy. 1999. "Radon Removal from Water Supplies by Diffused Bubble Aeration System." Journal of Radioanalytical and Nuclear Chemistry 241(1): 3-9.
Adham, Samer, et al. 2005. Development of a Microfiltration and Ultrafiltration Knowledge Base. AWWA Research Foundation (AWWARF) Project 2763, Report 91059.
Ahmed, Zeyad T., David W. Hand, and David L. Perram. 2014. Removal of Volatile Organic Contaminants via Low Profile Aeration Technology. Jointly Sponsored by U.S. Environmental Protection Agency, Water Environment Research Foundation, Water Research Foundation. Web Report #4439. 
American Water Works Association (AWWA). 1995. AWWA M38: Electrodialysis and Electrodialysis Reversal. American Water Works Association, Denver, CO.
AWWA. 2005. Microfiltration and Ultrafiltration Membranes for Drinking Water: Manual of Water Supply Practices M53. Denver, CO: American Water Works Association.
AWWA. 2007. Reverse Osmosis and Nanofiltration: Manual of Water Supply Practices M46. Denver, CO: American Water Works Association.
AWWA. 2011. Water Quality & Treatment: A Handbook on Drinking Water Quality. Sixth Edition. J.K. Edzwald, Ed. New York: McGraw-Hill.
Bilello, L.J. and J.E. Singley. 1986. "Removing Trihalomehtanes by Packed-Column and Diffused Aeration." Journal AWWA 78(2): 62-71.
Boulos, Lina, et al. 2013. Electrolysis in Water Treatment, Design Concepts, and Treatment Costs. Presented at the AWWA Water Quality Technology Conference. Long Beach, CA. November 4.
Brooke, Ethan and M. Robin Collins. 2011. "Posttreatment aeration to reduce THMs." Journal AWWA 103(10): 84-96.
Brown, Jess. 2011. Carollo Engineers. Email communication with D. Fortney, SAIC.
Brown, J., R.S. Summers, M. LeChevallier, H. Collins, J. Alan Roberson, S. Hubbs, and E. Dickenson. 2015. "Biological Drinking Water Treatment? Naturally." Journal AWWA 107(12): 20-30.
Brown, J., G. Upadhyaya, J. Carter, T. Brown, and C. Lauderdale. 2016. North American Biofiltration Knowledge Base: Appendices A and B. Project #4459. Denver, CO: Water Research Foundation.
Calgon Carbon. 2020. UV Oxidation for Dummies. Retrieved from:  https://www.wateronline.com/doc/uv-oxidation-for-dummies-0001, November 17, 2020.
Carter, Jason T., Jess Brown, and Chance Lauderdale. 2014. Best Practices from Biofiltration Knowledge Base. Presented at AWWA Water Quality Technology Conference, New Orleans, Louisiana. November 16-20.
Cecchetti, Aidan R., Harrison Roakes, and M. Robin Collins. 2014. "Influence of selected variables on trihalomethane removals by spray aeration." Journal AWWA 106(5): E242-E252.
Chakmak, A., D. White, A. Ghosh, and C. Seidel. 2015. Research Does Not Necessarily Translate to Easy Implementation: Piloting Biological Treatment for a Challenging Source. Presented at the AWWA Annual Conference & Exposition, Anaheim, California, June 7-10.
Chiu, C.-A., P. Westerhoff, and A. Ghosh. 2009. Removal of Organic Nitrogen and Other DBP Precursors by Granular Activated Carbon (GAC)  -  Rapid Small-Scale Column Tests (RSSCTs) to Evaluate Disinfection By-product Control. Presented at the AWWA Water Quality Technology Conference & Exposition, Seattle, Washington, November 15-19, 2009.
Chuang, Y., A. Szczuka, F. Shabani, J. Munoz, R. Aflaki, S.D. Hammond, and W.A. Mitch. 2019. "Pilot-scale comparison of microfiltration/reverse osmosis and ozone/biological activated carbon with UV/hydrogen peroxide or UV/free chlorine AOP treatment for controlling disinfection byproducts during wastewater reuse." Water Research 152: 215-225.
Cleveland, C. and D. Mount. 2011. When GAC is not Enough for DBP Compliance: City of West Sacramento's Bryte Bend WTP. Presented at the AWWA Annual Conference & Exposition, Washington, D.C. June 12-16.
Clunie, B., P. Ranslow, and S. DeFrancesco. 2014. Finished Water Spray Aeration for Reduction of Total Tri-halomethanes: A Case Study. Presented at the AWWA Annual Conference & Exposition, Boston, Massachusetts. June 8-11.
Cotton, Christine, 2004. Malcolm Pirnie. Personal communication with L. Petruzzi, SAIC.
Crittenden, J.C., R.R. Trussell, D.W. Hand, K.J. Howe, and G. Tchobanoglous. 2012. MWH's Water Treatment Principles and Design. Third Edition. John Wiley & Sons, Inc., Hoboken, New Jersey. ISBN 978-0-470-40539-0.
Diefenthal, George. 2020. Neotech Aqua Solutions. Personal communication with C. Altizer, Abt Associates.
El-Behlil, M.A. and S.A. Adma. 2012. Volatile Organic Chemicals Removal from Contaminated Water Using Low Profile Sieve Tray Air Stripping Towers. International Water Technology Journal, Vol. 2, No.2, pp. 165-177. June.
Evans, Ashley, Doug Wagner, Kirk Nowack, and Jason Carter. 2016. Optimizing BAF in Concert with Ozone Downstream of Softening to Combat Cyanotoxins. Presented at AWWA Annual Conference & Exposition, Chicago, Illinois. June 19-22.
Evans, A.N., J.T. Carter, J. Brown, and C. Lauderdale. 2019. "North American biofiltration practices." AWWA Water Science 2019: e1152. https://doi.org/10.1002/aws2.1152
Fiske, Peter S., Jason Oppenheimer, Randall Moore, and Randy Everett. 2011. "In-tank Aeration Predicts and Reduces THMs." Optflow. November. 
Fiske, Peter S. 2013. "Spray Aeration System Reduces THM Levels by 53% in and 8 MG Water Tank." Chesapeake Magazine. Summer. 
Fitz, Allen. 2004. Economic and Engineering Services, Inc., Bellevue, WA. Personal communication with R. Fetter, SAIC.
Ford, R., M. Carlson, and W.D. Bellamy. 2001. "Pilot-testing with the End in Mind." AWWA Journal. May.
Gerrity, Daniel, Sujanie Gamage, Janie C. Holady, Douglas B. Mawhinney, Oscar Quinones, Rebecca A. Trenholm, and Shane A. Snyder. 2011. "Pilot-scale evaluation of ozone and biological activated carbon for trace organic contaminant mitigation and disinfection." Water Research 45: 2155-2165.
Glover, C.M., O. Quinones, and E.R.V. Dickenson. 2018. "Removal of perfluoroalkyl and polyfluoroalkyl substances in potable reuse systems." Water Research 144: 454-461.
Gosh, A., C. Seidel, E. Townsend, R. Pacheco, and C. Corwin. 2015. Reducing Volatile Disinfection By-Products in Treated Drinking Water Using Aeration Technologies. Web Report #4441. Denver, CO: Water Research Foundation.
Gurian, P.L., B. Dvorak, D. Bitter, and S. Gasteyer. 2010. Arsenic Rule Implementation: Research and Dissemination Needs. Presented at the AWWA Annual Conference & Exposition, Chicago, Illinois, June 20-24
Halevy, Patrick Z., William B. Anderson, and Peter M. Huck. 2013. Pilot-Scale Investigation of Advanced Water Treatment Technologies in a Eutrophic River in Southern Ontario. Presented at AWWA Water Quality Technology Conference. Long Beach, CA. November 3-7.
Hokanson, D.R., D.W. Hand, and J.C. Crittenden. 1999. Aeration Systems Analysis Program (ASAP(TM)) Manual. Michigan Technological University.
Hulsey, Bob. 2004. Black and Veatch, Inc., Kansas City, MO. Personal communication with S. Ragland and R. Fetter, SAIC.
Hwang, M., J. Biggs, L. Schimmoler, and J. Lozier. 2016. Potable Reuse for Inland Applications: Pilot Testing Results Tucson, AZ. Presented at the AWWA Water Quality Technology Conference, Indianapolis, Indiana. November 13-15.
Jaeger Products, Inc. 2003. Jaeger Frequently Asked Questions. Houston, TX. Retrieved from: http://www.jaeger.com/faq.htm.
Jakubowski, Craig, Paul Thompson, and Mike Brooks. 2011. Replacement of Powdered Activated Carbon with Ozone and Biological Filtration for Taste, Odour and Cyanotoxin Removal. Presented at 74th Annual Water Industry Engineers and Operators' Conference. September 6-8.
Kini, S., P. Schwer, C. Lekven, and P. Seitz. 2011. Maui Organic Carbon Reduction Study. Presented at the AWWA Annual Conference & Exposition, Washington, D.C. June 12-16.
Kinner, N.E., G. Shell, P. Quern, and C. Lessard. 1988. Radon Removal from Drinking Water Using Granular Activated Carbon, Packed Tower Aeration and Diffused Bubble Aeration Techniques. Environmental Research Group, Department of Civil Engineering, University of New Hampshire.
Kommineni, Sunil, Ph.D. 2003 and 2004. Malcolm Pirnie. Personal communication with R. Fetter, SAIC.
Kohlmeier, D.B. Resolving DBP Violations at Small Water Systems Using GAC. 2011. Presented at the AWWA Annual Conference & Exposition, Washington, D.C. June 12-16.
Krasner, S.W. M.J. Sclimenti, and B.M. Coffey. 1993. "Testing Biologically Active Filters for Removing Aldehydes Formed During Ozonation." Journal AWWA 85(5): 62-71.
Kwan, P., K. Schnadt, G. Van Orsdol, and C. Murray. 2010. Simultaneous Lead and Copper and Arsenic Rules Compliance with Snowmelt and Hot Water. Presented at the AWWA Water Quality Technology Conference, Savannah, Georgia. November 14-18.
Lekkerkerker-Teunissen, K., A.H. Knol, J.G. Derks, M.B. Heringa, C.J. Houtman, C.H.M. Hofman-Caris, E.F. Beerendonk, A. Reus, J.Q.J.C. Verberk, and J.C. van Dijk. 2013. "Pilot Plant Results with Three Different Types of UV Lamps for Advanced Oxidation." Ozone: Science & Engineering 35: 38 - 48.
Linden, K. 2015. Some Things to Keep in Mind About UV AOP. Presented at International Ultraviolet Association and WateReuse 30[th] Annual WateReuse Symposium. Seattle, Washington. September 13.
Lowry, Jerry D. and Sylvia B. Lowry. 2002. Oxidation of As(III) by Aeration and Storage. Prepared for National Risk Management Research Laboratory, Office of Research and Development, U.S. Environmental Protection Agency. EPA/600/R-01/102. January.
Mead, E., and J. Leibbert. 1998. "A Comparison of Packed-Column and Low-Profile Sieve Tray Air Strippers." Proceedings of the 1998 Conference on Hazardous Waste Research, 328-334.
Medora Corporation. 2014a. Summarized Case Study (PW). USOHPW-LOC1027.001, Last updated: September 16, 2014. Retrieved from: www.medoraco.com.
Medora Corporation. 2014b. Summarized Case Study (PW). USCOPW-LOC1080.001; Last updated: July 16, 2014. Retrieved from: www.medoraco.com.
Montana University System Water Center. 2001. Demonstration Project Summary: The Use of Biological Filtration for the Removal of Soluble Iron from Drinking Water. Bozeman, MT: Small Systems Technical Assistance Center. Principal Investigator: Dr. Anne Camper. Retreived from: http://water.montana.edu.
National Academy of Sciences (NAS). 1997. Safe Water From Every Tap: Improving Water Service to Small Communities. Washington, D.C.: National Research Council, Committee on Small Water Supply Systems. Retrieved from: http://stills.nap.edu/html/safe/index.html#con 
National Science Foundation (NSF). 2003a. Environmental Technology Verification Protocol: Protocol for Equipment Verification Testing for Inactivation of Microbiological Contaminants. Ann Arbor, MI. 03/9204/EPADWCTR. 
National Water Research Institute (NWRI). 2000. Treatment Technologies for Removal of Methyl Tertiary Butyl Ether (MTBE) from Drinking Water. Second Edition. Report for the California MTBE Research Partnership. NWRI-99-06. February. 
NSF. 2003b. Environmental Technology Verification Protocol: Protocol for Equipment Verification Testing for Arsenic Removal. Ann Arbor, MI. 03/9201/EPADWCTR.
Neemann, Jeff. 2004. Black and Veatch. Personal communication with R. Fetter, SAIC.
Neofotistos, Paris. 2004. Infilco Degremont. Personal communication with L. Petruzzi, SAIC.
Overbeck, Paul. 2004. GDT Corporation and International Ozone Association. Personal communication with R. Fetter, SAIC.
Page, S., C. Russell, and A. Ohrt. 2016. Impact of Biofiltration on NDMA Precursor Concentrations. Presented at AWWA Water Quality Technology Conference, Indianapolis, Indiana. November 13-15.
PAX Water Technologies. 2012. Case Study: In-tank THM Removal System Eliminates THMs for Rural Consecutive System. March.
Pech, Adrian. 2020. Trojan Technologies. E-mail communication with C. Altizer, Abt Associates.
Pico, Tristan M. 1997. Treatability Test of a Stacked-Tray Air Stripper for VOCs in Water. Environmental Restoration Division, Lawrence Livermore National Laboratory.
Rattier, Maxime, Julien Reungoat, Wolfgang Gernjak, and Jurg Keller. 2012. Organic Micropollutant Removal by Biological Activated Carbon Filtration: A Review. Urban Water Security Research Alliance Technical Report No. 53.
Reed, Hal. 2013. Localized Control of DBPs By Spray Stripping in Storage Tanks. Presented at American Water Works Association Annual Conference & Exposition, Denver, Colorado. June 10.
Reed, Joseph A. 2002. Application of an Electrodialysis (EDR) Process to Remediate a Brine-Impacted Aquifer. Presented at the IPEC Conference, Albuquerque, NM. October 22-25.
Robinson, K. 2016. UV/Chlorine AOP for Potable Reuse: Lower Cost Option. Presented at Water Reuse in Texas. San Marcos, Texas. June 15.
Roccaro, J. 2016. Water Treatment Pilot Studies at SCWA. Presentation.
Rosenfeldt, E., A.K. Boal, J. Springer, B. Stanford, S. Rivera, R.D. Kashinkunti, and D.H. Metz. 2013. "Comparison of UV-mediated Advanced Oxidation." Journal AWWA 105(7): 29-33.
Rosenfeldt, E., N. Sajdak, M. Roberts, and M.R. Collins. 2014. Beyond the Tanks: Assessing Impact of Tank Aeration for THM Reduction in Distribution Systems. Presented at the AWWA Annual Conference & Exposition, Boston, Massachusetts, June 8-11.
Sarathy , S.R., M.I. Stefan, A. Royce, and M. Mohseni. 2011. "Pilot-scale UV/H2O2 advanced oxidation process for surface water treatment and downstream biological treatment: effects on natural organic matter characteristics and DBP formation potential." Environmental Technology 32(15): 1709 - 1718.
Sawade, Emma, Melanie Vache, Lionel Ho, Tim Tang, and Gayle Newcombe. 2012. Implementation of biological filtration for the treatment of cyanobacterial metabolites. Presented at AWWA Water Quality Technology Conference, Toronto, Ontario. November 4-7.
Schers, G.-J., K. Hink, S. Sansom, and S. Heady. 2011. State-of-the-Art Treatment for a Large Florida City. Presented at the AWWA Annual Conference & Exposition, Washington, D.C., June 12-16,.
Schneider, Orren D. and Mark W. LeChevallier. 2017. "A Cost-Effective Treatment Process for Producing High-Quality Drinking Water." Journal AWWA 109(3): 39-47.
Selbes, Meric, Jess Brown, Chance Lauderdale, and Tanju Karanfil. 2017. "Removal of Selected C- and N-DBP Precursors in Biologically Active Filters." Journal AWWA 109(3): E73-E84.
Sherant, S.R. 2008. Trihalomethane Control by Aeration: A Thesis in Environemtnal Pollution Control. The Pennsylvania State University. Master of Science Thesis.
Song, A. 2016. Treatment Process Selection and Piloting Deep Bed Filters and Biological Filtration for Denver Water's New North System Renewal Water Treatment Plant (NSRWTP). Presented at the AWWA Annual Conference & Exposition, Chicago, Illinois. June 19-22.
Spacek, David. 2009. "Areation Decreases THMs." Optflow. September.
Tseng and Kim. 2014. Managing Total Trihalomethane with an Aeration System: Anaverde tank pilot study. Presented at AWWA Water Quality and Technology Conference, New Orleans, Louisiana. November 17.
Urfer, Daniel, Peter M. Huck, Stephen D.J. Booth, and Bradley M. Coffey. 1997. "Biological filtration for BOM and particle removal: a critical review." Journal AWWA 89(12): 83-98.
U.S. Army Corps of Engineers (USACE). 2001. Engineering and Design: Air Stripping. Design Guide No. 1110-1-3. October 31.
U.S. Bureau of Labor Statistics (U.S. BLS). 2020a. May 2014 National Industry-Specific Occupational Employment and Wage Estimates, NAICS 2213  -  Water, Sewage and Other Systems. Washington, D.C.: U.S. Bureau of Labor Statistics. Retrieved from: http://www.bls.gov/oes/current/naics4_221300.htm#17-0000, November 24, 2020.
U.S. BLS. 2020b. Employment Cost Trends, Databases: Employer Cost for Employee Compensation. Washington, D.C.: U.S. Bureau of Labor Statistics. Retrieved from: http://www.bls.gov/ncs/ect/, November 24, 2020.
U.S. Department of Defense (U.S. DoD), Environmental Security Technology Certification Program (ESTCP). 2008. Final Report: ER-0544: Direct Fixed-bed Biological Perchlorate Destruction Demonstration. September 25.
U.S. Environmental Protection Agency (U.S. EPA). 1983. Treatment of Volatile Organic Compounds in Drinking Water. Cincinnati, OH: Municipal Environmental Research Laboratory. EPA-600/8-83-019. May.
U.S. EPA. 1998. Small System Compliance Technology List for the Non-Microbial Contaminants Regulated Before 1996. Washington, D.C.: Office of Water. EPA 815-R-98-002.
U.S. EPA. 2000. Arsenic Removal from Drinking Water by Ion Exchange and Activated Alumina Plants. Columbus, OH: U.S. Environmental Protection Agency, Office of Research and Development. EPA/600/R-00/088.
U.S. EPA. 2003a. Analytical Methods for Drinking Water. U.S. Environmental Protection Agency, Office of Groundwater and Drinking Water. 
U.S. EPA. 2003b. Draft Membrane Filtration Guidance Manual. Washington, D.C.: Office of Water. EPA 815-D-03-008.
U.S. EPA. 2003c. Ultraviolet Disinfection Guidance Manual. Washington, D.C.: Office of Water. EPA 815-D-03-007.
U.S. EPA. 2003d. Arsenic Treatment Technology Evaluation Handbook for Small Water Systems.  EPA 816-R-03-014. Washington, D.C.: U.S. Environmental Protection Agency.
U.S. EPA. 2003e. Economic Analysis for the Long Term 2 Enhanced Surface Water Treatment Rule. Proposal Draft. Prepared by The Cadmus Group, Arlington, VA. 
U.S. EPA. 2005. Membrane Filtration Guidance Manual. Washington, D.C.: Office of Water. EPA 815-R-06-009.
U.S. EPA. 2020. Membrane Separation, Drinking Water Treatability Database. Retrieved from: https://tdb.epa.gov/tdb/treatmentprocess?treatmentProcessId=-2103528007, November 27, 2020.
Viessman and Hammer. 1998. Water Supply and Pollution Control. Menlo Park, CA: Addison Wesley.
Wachinski, Tony. 2003. Pall Corporation. Personal communication with R. Fetter, SAIC.
Walfoort, Carol, Michael J. Messina, and Dean Miner. 2008. "Storage Tank Aeration Eliminates Trihalomethanes." Optflow. May.
Wang, D., J.R. Bolton, S.A. Andrews, and R. Hofmann. 2015a. "UV/chlorine control of drinking water taste and odour at pilot and full-scale." Chemosphere 136: 239-244.
Wang, D., J.R. Bolton, S.A. Andrews, and R. Hofmann. 2015b. "Formation of disinfection by-products in the ultraviolet/chlorine advanced oxidation process." Science of the Total Environment 518 - 519: 49 - 57.
Washington State Department of Health (WADOH). 2003. Guidance Document: Arsenic Treatment for Small Water Systems. DOH PUB. #331-210.
Westerhoff, Paul. 2003. Arizona State University. Personal communication with R. Fetter, SAIC.
Williams, Gary E. 2013. Reducing THM via Aeration: A Case Study. Presented at Southeast American Water Works Association District Meeting. November 21.
Zhang, Z, Y. Chuang, A. Szczuka, K.P. Ishida, S. Roback, M.H. Plumlee, and W.A. Mitch. 2019. "Pilot-scale evaluation of oxidant speciation, 1,4-dioxane degradation and disinfection byproduct formation during UV/hydrogen peroxide, UV/free chlorine and UV/chloramines advanced oxidation process treatment for potable reuse." Water Research 164: 114939.
