Water: Regulatory Guidance
Appendix B -Detailed Listing of Comments
Table of Contents
- 2.1 Defining the Breakpoint for Small Systems
- 2.2 Defining the Treatment Objective
- 3.1 Use of Peak Production Flow for Design
- 3.2 Redundancy
- 3.3 Storage Use for Design Capacity Mitigation
- 5.1 Types of Permits
- 5.2 Waste Disposal of Brine
- 5.3 Costs to Obtain Permits
- 5.4 Treatability and Pilot Testing
- 7.1 Basic Footprints
- 7.2 Value of Land
- 8.1 Key Characteristics of "Median" Large Water Systems
- 8.2 Key Characteristics of "Median" Small Water Systems
- 8.3 Key Characteristics of "Outlier" Large and Small Water Systems
- 9.1 Miscellaneous
- 9.2 Use Existing Models or Use WBS Build-Ups Using RS Means
- 9.3 Level of Detail if WBS Used and Additional Indirect Costs to be Included
- 10.1 Labor Rates and Indirect Costs for O&M
- 10.2 Labor Requirements of Various Technologies
- 11.1 General
- 11.2 Membrane Technologies
- 11.3 Anion Exchange (AE)
- 11.4 Cation Exchange (CE)
- 11.5 Activated Alumina (AA)
- 11.6 Packed Tower Aeration (PTA)
1. Introduction to Appendix
Appendix B contains an attempt to characterize the sum total of major points made by various participants. The points expressed in this Appendix generally do not reflect the concensus of the group. Numerous suggestions of appropriate cost techniques based on individual experience were presented, but not carried forward to the main report either because there was a lack of consensus, or simply because some participants did not have the necessary information before them to judge the merits of the suggested design factors. The Appendix is provided to convey a fuller impression of the issues and perspectives raised during the meeting.
2. Background on Cost Modeling Assumptions
2.1 Defining the Breakpoint for Small Systems
Is there a point below or above which EPA could assume PWSs would choose fully engineered or packaged treatment?
- Systems between 1 and 10 MGD choose both packaged and fully engineered treatment.
- The breakpoint is probably closer to 1 MGD than 10 MGD.
- Should the break be based on the number of customers served?
- One might not see any breakpoint for RO and other forms of membrane filtration. If there was a breakpoint, it would at least be higher than 10 MGD.
- Where would the breakpoint be for systems that have, for example, 30 ground water wells that each look like a 1 MGD system?
- When you deal with a larger system with lots of ground water wells, would you treat the wells separately when there is distance between them?
- Multiple wells have been an issue with corrosion control.
- There are situations where a system will have wells all over and situations where all of the wells will be confined to a specific area.
- There are lots of systems with both surface water and ground water. It seems EPA would need to consider that both situations are occurring.
O&M or Capital Intensive Technologies for Small Systems:
Should EPA be biased toward O&M or capital intensive technologies for small systems?
- Small systems often do the opposite of what is in their (or their customer's) best interest. In reality, they tend to choose technologies that are low in capital costs but high in O&M costs.
- Small systems tend to be risk averse in their decision making. They do not want to see any new complexities if they can avoid them.
- Before the Ground Water Disinfection Rule (GWDR), small PWSs are operating pump stations. After the GWDR, these PWSs will be operating a treatment system. This will make land a big issue.
- One would have to look at a system's future viability. In the past, there has been a tendency of these systems to choose technologies, namely packaged technologies, that are O&M intensive rather than capital intensive. Should this situation be accepted when it is known that many of the systems will not be able to supply the O&M resources required?
Influence of State Regulators:
Will States establish a breakpoint or require small systems to choose fully engineered or packaged technology?
- States will not decide whether a small system should use a fully engineered or a packaged technology.
- Small systems will need to propose their own choice to the State and provide the appropriate cost information.
- Some States may try to discourage packaged treatment if they see a consistent pattern of failure regarding Wamp;M.
- Some small systems are beginning to hire more skilled system operators.
- Some regulators will be raising questions when reviewing small system treatment plans and will be encouraging the conduct of total life cycle analyses.
2.2 Defining the Treatment Objective
Should EPA assume PWSs will blend their source water for some contaminants?
- One participant commented that small systems may use blending less once they get into producing public reports. At that time, they might start receiving more public pressure to perform full treatment.
- It was noted by another participant that if blending is ruled out, costs for many regulatory treatment options will be flat across all regulatory options.
- One participant noted that they had seen quite extensive use of reverse osmosis with side stream blending.
- Other participants noted that they had witnessed blending for total dissolved solids (TDSs), fluoride, and radium.
3. Sizing the System
3.1 Use of Peak Production Flow for Design
Should EPA use peak production flow over one year for design capacity for all system sizes? Some commissions use peak flow on an annual basis. One analysis using peak flow revealed that peak flow is about 50 percent above the average daily flow. Should EPA use this number as an assumption for modeling capital costs?
- Initially, one participant indicated that EPA should use maximum daily flow for all systems, except the very small.
- For very small systems, one participant noted that desings need to look at peak hourly pumping rate (because of limited storage capacity in the system).
- It was noted by one participant that some small systems' peak flows would routinely exceed twice the average flow.
- Group members were generally comfortable with the notion that EPA should use maximum design flow for large and medium systems and use peak hourly flow for small systems. If hourly flows weren't available, maximum design flows were seen as the next best option.
- It was noted that storage may become the viable treatment alternative for some systems when the only other alternatives are expensive technology options.
- One participant indicated that his firm tracks peak daily and average daily flows. His experience was that the flows run between 1.2 and 2.0. For design purposes, his firm uses actual data on a system-specific basis.
- Another participant indicated that these types of systems were probably the outliers. In general, this individual thought that EPA could assume the peaking factor to be 2 for systems less than 1.5 MGD and approximately 1.5 for systems greater than 1.5 MGD.
- It was noted by another group member that in a recent State study that was conducted, 2.3 was found to be the average peaking factor for small and medium systems. The State settled on 2 as the average.
- Another participant commented that in his firm's review of a number of studies on the topic, they were generally seeing a 2.0 spread between average flow and design flow for both small and medium groundwater and surface water systems. This individual noted that in a recent American
- Water Works Association (AWWA) study, it was not until the system size categories of 5 MGD and above that the peaking factors of 1.5 were being realized.
- One participant asked whether EPA should consider groundwater and surface water facilities separately in their analyses of flow. There were some additional comments made about the role of irrigation in affecting peak flows.
- Finally, there was general agreement with the use of 2 as a minimum peaking factor for larger systems for estimating design flows.
How should EPA define redundancy? It might be defined as a system's ability to meet capacity when its largest unit is out of service. How can EPA's cost models be designed to take redundancy into account, and how many units would be typical to assume?
- One State participant indicated that redundancy is not really an issue with noncommunity water systems. He noted that these systems typically close down while repairs are made. Redundancy is more an issue for community water systems. It is then a function of size and storage. Very small systems will typically have 1 day of storage to use while repairs are being made. Systems serving populations greater than 50,000 are the systems that really require redundancies.
- It was noted by another participant that all States tend to have requirements of 10 State Standards on their books; however, there is variability in terms of how the standards are enforced.
- This same individual indicated that one would not expect to see less than 3 units and that, typically, there would be 4 units to address operational needs.
- Another participant indicated that system size is an issue. To clarify his point, he indicated that it would be hard to install 4 units in < 100,000 gallon plant.
- Redundancy will be based in part on the nature of the contaminants being treated; that is, whether they are acute or chronic contaminants. If the contaminant is arsenic, systems are likely to build in lots of redundancy.
- One participant asked how redundancies for POU and POE devices should be handled? Another participant indicated that redundancies will vary for POU and POE devices also. Each technology or process will vary by how far they can be pushed. Some technologies can be run over capacity for a short time period and no real performance problems will be realized. However, if you pushed this same technology beyond capacity for a long period of time, you would see negative impacts on performance (e.g., microfiltration). The point being inferred was that the systems would be retired early as opposed to being over designed.
- One participant indicated that they were not sure redundancy is as big an issue as how frequently a system is going to have problem.
- It was noted by one group member that they do not build in full redundancy at peak.
- One participant noted that they were required to have 25 percent redundancy at average day for surface water systems. (This requirement, however, is well below typical design practices.)
- There was general agreement on the point made by one participant that it is critical to have redundancy or backup to ensure the reliability of pumpage and chemical feed rates.
- One participant indicated that it is difficult to identify representative redundancies. Redundancy decisions need to be based on the failure mode of the technology in consideration and the turn-around time to repair problems.
- Many participants thought that EPA should use the 10 State Standards as its guide for developing model criteria. There seemed to be agreement on this point. There was also agreement that a wide range of "applications" within the 10 State Standards are possible.
- It was pointed out by one participant that the 1997 edition of the 10 State Standards is available. The new edition includes a policy on automation.
3.3 Storage Use for Design Capacity Mitigation
Do small systems use storage to address design capacity issues?
- It was noted that storage should not be used as a tradeoff for capacity.
- One participant indicated that it is not good operating practice to use storage for design capacity; however, if regulators allow it, some systems are likely to use.
- Another comment was made that systems need storage time for some treatment technologies.
- One State official commented that they are not typically seeing systems oversize their storage for capacity.
- Another State official indicated that systems are generally using storage to address contact time (CT) requirements under the Surface Water Treatment Rule (SWTR).
- Another participant noted that they typically see 1 day of storage for systems beyond storage needs for CT.
- Some participants noted that they are seeing systems install tanks to keep up water pressure. They further noted that it is rare to see storage being installed except for fire safety reasons.
- Finally, it was noted that if a system strictly disinfects, they must go beyond the minimum CT requirements.
- Workshop participants agreed with the comment that storage is a viable treatment technique for radon.
4. Differences in Assumptions Between Packaged Systems and Engineered Systems
What assumptions should EPA be using in its cost models to distinguish between small and large PWSs and their respective uses of packaged and fully engineered treatment technologies? A draft table of assumptions that had been developed for the workshop was reviewed by participants as part of this discussion topic.
- One participant indicated that for filtration package units, his experience has been that engineering for site work is involved.
- Another participant questioned the definition of a packaged unit (i.e., whether it includes pumps and pipes). It was noted that pumps and piping are included for some technologies but not for others. It was then recommended that some allowance for engineering costs for packaged systems (thus for small systems) be included. It was noted that packaged treatment requires some site engineering.
- It was recommended by one participant that the Agency distinguish between engineering that is provided by a contractor and that which is provided by a manufacturer.
- Workshop participants noted that manufacturers perform some of the engineering for packaged systems.
- It was noted that small systems tend to use contractors for engineering when they are building up from scratch.
- One State official noted that their State requires an engineering stamp on any plan. However, it did not matter to the State whether the engineer was with the plant, the manufacturer, or was a contractor.
- A comment was made that more economies of scale are realized on engineering for large systems.
- Another participant indicated that he disagreed with the assertion in EPA's table that there are minimal contractor costs for packaged units. There was some discussion that these types of costs may eventually be included as part of total package costs. Some manufacturers presently market their technologies in such a fashion, but such marketing does not occur on a uniform or global basis. The general consensus of the group from this discussion was that EPA should consider engineering costs for packaged plants.
- While it was acknowledged that R.S. Means is a reasonable source for engineering costs, it was noted that it only applies to balance of plant costs and not process costs. Workshop participants agreed that engineering costs need to be incorporated as part of both process and balance-of-plant costs. As such, EPA should not rely exclusively on R.S. Means.
- It was noted by one workshop participant that owners and regulators approach costing differently and therefore use different costing components. With regard to overhead costs, R.S. Means looks at general contractor overhead and these costs are covered by the 54 percent figure in EPA's table. Regulators and owners put other items into overhead that are not considered by R.S. Means. This same participant noted that Means is a good benchmark for a fully engineered system.
- Another workshop participant indicated that EPA needs to be clear about the components that are being included under overhead costs when it puts forth its models to the public. People will then be able to inform the Agency as to the reasonableness of the components being included and the corresponding percent figure being used.
What permitting activities should EPA consider in its cost models? Draft technology tables of permitting assumptions were reviewed by participants as part of the discussion. See Section 5 of Discussion Summary for actual listing of permits proposed to be considered by EPA.
5.1 Types of Permits
- It was recommended that EPA include OSHA/EPA 112R risk management permits for chlorine under the appropriate technologies. It was noted that a local fire department might get involved in risk management when a building permit is pulled, especially if the building of the facility or treatment component of an existing facility is to take place in a residential area.
- Some participants noted that thresholds based on population density exist with regard to which systems have to contend with OSHA/EPA 112R permits. The major point being made was that risk management is a highly site-specific issue and not one that can be sorted out in a national level analysis.
5.2 Waste Disposal of Brine
What permitting issues are associated with regard to the disposal of brine?
- It was noted that brine is not just an issue or only covered under NPDES permits. Brine becomes an UIC issue for facilities that do not have surface water in which to discharge the brine.
- One State official noted that they had found the issue of brine disposal to be a major deal breaker.
- This comment was supported by other workshop participants who also indicated that often there is no cost-effective solution for dealing with brine.
- Other participants indicated that decisions regarding brine disposal are typically dependent on the size of the system and the stream to receive the brine. It was noted that if a receiving stream is small, the issue of brine disposal is bigger than if the receiving stream is large.
- Workshop participants agreed that brine and its relationship to permitting should be part of EPA's decision trees.
- If was asked that if EPA costs out deep-well injection of brine, the consequence will be that RO can not be BAT across the country for any contaminant. There was discussion that RO is applicable in some locations, for example, those facilities that can discharge to the ocean. In addition, it was noted that not all RO technologies produce a brine (i.e., the TDSs are not concentrated). If TDSs are concentrated under an RO or NF application, brine can be a major cost of the installation.
- The handling of brine tends to be geographic. In Florida, deep wells are used for disposal. Brine lines are being used in California. In Texas, the brine is discharged into rivers or French drains. In still other areas, the brine is being discharged to publicly or privately owned treatment works. It is more likely that the brine, however, is being discharged to public rather than private treatment works. Disposal to a wastewater treatment facility works well when the drinking water and wastewater personnel represent the same municipality.
- For public systems, EPA might need to anticipate some discharge of brine to sewers. One participant thought a generic cost could be derived for public systems to put brine back in rivers. Another participant felt that discharge philosophies vary too much by district to allow a generic cost to be developed.
5.3 Cost to Obtain Permits
What types of costs are involved to obtain the permits associated with PWS retrofits or new construction?
- Permit issuance can involve significant costs for a State. Costs are probably somewhat driven by the size of the system and the extensiveness of the modifications being made. The flat labor rate charged by California for plan review is $76.45 per hour. For a relatively small system with a small permit change, approximately 8 hours would be required of the system in terms of preparing the permit application and accompanying materials. The permit review process, however, would involve approximately 40 to 80 hours for State personnel. Systems serving less than 1000 connections are charged a flat rate of $253 for treatment changes.
- Some States charge a plan review fee while others do not. Of those that charge, the costs probably vary widely. EPA would probably need to go with some general percentage increase. The old Water
- Cost model also used a percent, probably 1 or 2 percent.
- Permitting costs are technology-dependent.
- One participant noted that they were not sure EPA should include piloting. The individual was not certain piloting costs could be sufficiently distinguished from permitting costs.
- Another participant indicated that piloting will not be uniform so it is likely to be a variable cost, whereas permitting is likely to be a fixed cost.
- It was noted that legal fees are not presently included in EPA's costing assumptions for permitting. Some participants felt that legal fees should be included. The extent to which legal fees are incurred depends on size of the project. In a small project, legal fees can be a large proportion of the total cost. For larger projects, it is typically a smaller proportion of total costs.
- EPA's old models excluded permitting. Workshop participants debated about whether the full range of permitting and piloting costs would be covered by the three percent on the average.
- It was noted that a large number of permits are required. As such, the percentage would need to take the number of permits into consideration. However, it was also noted that there is too much variation in permit costs to breaking down cost models to look at individual types of permits.
- Some participants concluded that it would probably make more sense for EPA to use 3 percent as a floor or baseline. Another suggestion was that EPA add on a couple of extra percents to cover legal fees associated with permitting. Another suggestion was made that than when systems serving less than 500 people are considered, EPA should use a fixed cost for permitting. Finally, another suggestion was made for EPA to use a flat fee of $2,500 for systems less than 10K. For systems higher than 10K, EPA should use the 3 percent figure. EPA noted that it would analyze systems less than 10K using the 3 percent figure to see how close the systems came to the $2,500 figure.
5.4 Treatability and Pilot Testing
To what extent will treatability and pilot testing be conducted and for which technologies?
- There is a NSF and EPA pilot verification effort. Under the effort, a survey of States was made (with an approximate response rate of 28 States) on methods for reducing the numbers of pilot tests. It was suggested that survey findings be reviewed by EPA.
- It was noted that piloting costs can add up for small systems. If piloting is considered in EPA's models, it can outweigh the cost of the technologies themselves. One participant noted that he had experiences wherein the pilot study cost more than the equipment.
- Some workshop participants indicated that manufacturers are frustrated that they have to redo work via piloting. They are trying to sort out their one-time and new costs to hopefully reduce their piloting costs in the future.
- A barrier to the manufacturers is that States never want to accept data that has been used elsewhere.
- There was general hope expressed that the NSF/EPA pilot verification project will move States forward in granting reciprocity.
- Concern was expressed by some participants about the effect of piloting costs on small systems. The following general question was posed by one participant: If a small system spends its resources on piloting only to find out that the technology does not work, in what shape is the system left?
- Workshop participants agreed that piloting is needed to some degree just to ensure that a technology is viable. They noted that piloting will not go away, however, the degree to which piloting is performed could be reduced.
6. How to Apply POU and POE Devices as Technology of Choice
What are the scenarios under which POU and POE devices might be considered treatments of choice?
What are the differences in the design and/or application of both POU and POE devices?
- EPA officials noted that the Agency's models will assume that PWSs will provide oversight regarding the use and maintenance of POU and POE devices.
- One workshop participant indicated that POU/POE devices were recently used successfully for fluoride in Suffolk County, VA. However, this individual noted that fluoride is a relatively easy contaminant to control.
- EPA officials indicated that POU and POE devices cannot be used for acute contaminants. They can only be used with chronic contaminants. If POU/POE devices are to be used for such chronic contaminants as arsenic and radon, the type of units considered will involve carbon filtration or membranes.
- It was suggested that EPA review the Suffolk County study because it might suggest cost component items that could be used by the Agency.
- It was noted that EPA's model will need to assume monitoring of some number of units over some specified time periods. Monitoring is likely to involve significant labor hours.
- It was noted that a PWS would not monitor all devices. The system would only have to examine a statistically drawn sample of devices.
- The comment was made that POU and POE devices are not probably applicable for use by larger systems. The sheer number of units a PWS would have to manage and monitor would pose implementation and management difficulties. Small systems are probably the most likely users of POU and POE devices. Even small systems might choose these devices as a last option due to the complexities involved in dealing with individual households.
- It was stressed by some workshop participants that EPA requests for monitoring and inspection of POU/POE devices would best be delivered via guidance. It was noted that States and PWSs will require flexibility for implementing monitoring plans.
Are EPA's draft footprint calculations and assumptions on the use of land reasonable? When is land acquisition and availability an issue for each technology and at what system size?
7.1 Basic Footprints
- Historically, EPA's assumption has been that land is already owned by a facility is a sunk cost. EPA officials noted that it is not clear how land will be handled in the models that are being developed. Land can be a large component of capital cost or little cost at all.
- The issue of a basic footprint for waste handling was raised. Participants noted that a footprint was needed for storage of chemicals and waste. The footprint would also need to address the time period waste is kept on-site at a facility. The draft footprints presented to the group assumed a 30-day supply for chemicals and approximately a 1-week storage period for waste. When technologies such as anion exchange are considered, it was noted that the footprint needs to consider wash for brine at some specified frequency.
- It was recommended that EPA add onsite lagoons as one possible disposal option for brine in arid areas. If systems are landlocked, then lagoons could not be assumed.
- It was recommended that EPA increase the equipment access factor in the footprints for new systems. It was noted that in the real world, one would see more than 1 « times the footprint for chlorine gas, for example. It was further noted that accessibility to equipment is both site- and treatment-specific.
- Regarding the equipment access factor, one participant recommended that EPA take the raw area of the footprints for each piece of a process and sum them and then, depending upon the scale of the project, add some buffer factor, which would be a function of the scale of the project. This approach was being proposed in lieu of EPA's unit process by unit process approach. It was noted that this approach is more typical of standard design engineering practices.
- It was then noted that if EPA's task is to evaluate the incremental costs of proposed regulatory requirements, they need to look at equipment access on a unit process by unit process basis.
- Another participant commented that because there are so many factors that must be considered under each technology, it was probably reasonable for EPA to use its proposed menu-based approach.
- It was noted by another participant that EPA's approach appeared to be more conservative than a composite-based approach.
- One participant commented that under the GWDR, EPA would be looking at new facilities in the case of hypochlorite.
- Another participant indicated that the 30-day chemical storage EPA was proposing for hypochlorite was not adequate even though that is the time period in the 10 State Standards. It was noted that the time period would really vary across the country.
- Another participant recommended that EPA revisit the square footage being assumed for small systems under most of the technologies (systems less than 1 MGD).
- It was noted that land for spill containment or chemical delivery need to be considered under the footprints. It was further noted that the costs for spill containment would be sunk costs if the facility already had the land. Spill containment is an issue for new facilities.
- Another workshop participant indicated that land may be an issue for both existing and new facilities to achieve compliance with the Americans with Disabilities Act (ADA). Often times, compliance with new code requirements will require that additional land be acquired/used by existing facilities.
- It was indicated that EPA will probably have to look at the components of each technology as each regulatory package is developed. At that time, the Agency can give consideration to the buffer factor that should be applied. It is hard to determine the buffer factor a priori.
- Another participant indicated that the menu-based approach is probably the best EPA can do for retrofits. However, he cautioned that some existing facilities will have situations that requiring starting anew.
- It was noted that sinking a new well and adding treatment would be a new cost. The situation would not come into play if a facility had an existing well and an old pipe and replaced the old pipe while redoing the wellhead. This type of cost would be considered part of a lifetime cost for the system.
- When the general question was posed as to whether the types of systems being discussed were outliers, workshop participants were generally unable to respond. One participant recommended that facilities be surveyed on the issue.
- The issue of land for systems with wells throughout a town was raised. One participant noted that if a system has a well field, it would be forced due to spacing to already have land. If the facility did not have well fields, actual wells would be in people's backyards. This latter situation would require that land be obtained.
- There was some disagreement regarding whether land is an issue for groundwater systems. Some participants have not seen land as a problem for groundwater systems, whereas other participants had considerable experience where wells were actually located in people's backyards.
- When asked if there was an entry point above which being land-locked becomes an issue, one participant commented that if there were a source with 5 entry points, then the system probably would have space for some treatment, but how much treatment was not clear. It was further noted that some technologies require more space than others.
- Another comment was made that there are clearly situations where land availability is a problem. In most cases, it is probably not an issue. However, the group seemed to be saying that the issue of being land-locked was important enough that it should be considered under EPA's models. The group could not suggest the size entry point above or below which land availability becomes an issue.
7.2 Value of Land
- The issue of what value should be assigned to different types of land (i.e., urban, suburban, rural) was raised. The question was posed whether EPA should assume that as the price of land increases, systems will move away toward land that is not so highly priced and valued.
- One workshop participant indicated that the cost of piping can sometimes be more expensive than the price of land for systems with a small entry point.
- It was noted that if a system has a large entry point, it is likely to have land.
- A comment was also made that the cost of land varies by State.
- One participant indicated that some land costs are up to around $100,000 an acre. He noted that it would seem like a system could do a lot of piping at that price (1,000 feet) and get out of people's backyards.
- Another participant indicated that $100,000 would probably only get the system out of people's backyards but not out of neighborhoods.
- One suggestion was made that EPA use average land values based on zip codes as an easy first step. EPA could estimate costs based on occurrence data and zip codes.
- Another participant recommended that property tax assessment data be used for land values.
- Another suggestion was made that EPA use Census data to determine land availability based on population density by zip code. It was later pointed out that the pixel size of the Census data is variable and often very large, thus, ruling out the use of this data for drinking water assessment purposes.
- It was recommended by one participant that for systems with multiple entry points, EPA should throw piping costs on top of land costs.
- Another participant indicated that this situation would likely vary by contaminant.
- Another group member believed these types of systems would be outliers.
- One person believed that the most extreme cases with respect to land would be currently untreated systems on groundwater.
- It was noted that the contaminants of concern would include radon and arsenic.
- Many problems in urban areas occur due limited space. The issues that must be addressed include deciding how treatment should be performed within the limited space, how land needs should be determined and acquired, the sound barriers or other aesthetic structures that are needed and where they will be constructed, and managing the permitting process for noise and air pollution.
- Another comment was made that if no land is available, adjacent land may need to be condemned. There are legal components to condemning land.
8. Breakout Sessions
8.1 Key Characteristics of "Median" Large Water Systems
- Breakout group members agreed that large water systems should include those systems with flows greater than 1 MGD.
- Use design capacity for the plant.
- For very large treatment plants serving large cities whose populations have diminished significantly (particularly cities in the northeast), maximum daily flow should be used for costing purposes.
- For large systems, redundancy is not an issue. Prudent design practices cover downing scenarios and redundancy.
- When systems approach flows of 1 MGD, they must add key units. No specific cut-off point was recommended.
- For systems with flows significantly greater than 1 MGD, single point failure equipment should be redundant (e.g., special metering for chemical feed pumps and process pumping).
- For large systems, land is generally available onsite but often cannot be used because of wetland issues. So land must ultimately be acquired.
- Breakout group members agreed that 3 percent of construction value should be used for estimating permitting costs for large systems. It was recommended that EPA conduct a cost estimation study to validate the 3 percent figure.
- Large systems will conduct pilot tests in all cases for treatment optimization.
- Pilot testing undertaken will achieve compliance. Pilot testing, however, will not be driven by compliance.
- Representative large systems will have land available.
- Median large water systems will incur mitigation costs for wetlands.
- Costs need to be considered for wastewater treatment by calculating the footprint for treating waste concentrates and backwash water.
- Large systems will use packaged technology (modular units) for aeration, IE, and membrane and other pressure systems.
- Large systems will incur costs for technologies that disrupt their hydraulic profile (e.g., GAC and ozone).
- There is less civil engineering around a small package with automation even though a building and a roadway are involved. For large treatment plants, it is a significant integration exercise to automate. Engineering costs increase for large plants because the engineering content increases.
8.2 Key Characteristics of "Median" Small Water Systems
- The breakpoint for small systems should be those systems with flows less than 1 MGD.
- There was some discussion regarding further breakpoints based on populations served within the small system category. The breakpoints discussed included the following:
-3,300 to 500 people
-500 to 100 people
- <100 people
- On the contaminant occurrence side, most of the systems will be clustered around the low end of the occurrence curve. The cost effort should reflect this reality in evaluating treatment options and benefits.
- With regard to occurrence, there will always be a few extreme cases on the high side (e.g., high values for radon at 50,000 Pci/L). There is a need for EPA to conduct separate analyses for waters with extreme contaminant levels. Arsenic is a good example of the type of contaminant that should be evaluated separately. The basis of the point being made was that at very high occurrence levels, two stage chemical precipitation treatment will likely be needed to achieve effluents below the MCL.
Technologies and Decision Trees:
Breakout group participants made a list of the types of technologies that are suitable for median small systems to achieve possible compliance levels regardless of their occurrence situations. The group then evaluated the technologies to determine their likely costs and compared the costs to those that would be incurred for other alternatives such as blending and switching to an alternate source. Finally, they looked at possibilities for modeling the factors that would influence system decision making with regard to treatment. That is, would systems make decisions on an economic basis, regional basis, water source basis, or some combination of these factors?
- Blending is used by a fair number of small systems that have multiple wells and in instances where the final product will have contaminant levels below the MCL. Blending can be done with all contaminants except for microbial contaminants.
- Systems will also look to new sources whether these be outside the current system, the drilling of a new well, etc.. These options need to be incorporated into the decision trees.
- Package technologies are appropriate for use by small systems and should be costed.
- Small systems would use POE and POU for chronic contaminants but not for acute contaminants such as the microbials.
- EPA needs to estimate the cost of each technology alternative versus various occurrence targets to determine which approach is affordable.
- EPA should consider central treatment.
- States should make the determination if POE or POU are BAT.
- POU devices might turn out to be very expensive for systems in terms of the labor cost involved for monitoring.
- O&M requirements listed by EPA are typical for small systems.
- EPA should cost technologies that more suitable for automation like package treatment.
- EPA should consider the amount of capital available to small systems and the sources for these funds. For publicly-owned systems, there is the SRF (State Revolving Fund). Private systems, in general, have to borrow money. In some States, however, they can obtain resources from the SRF. This situation makes costing an even more complex activity.
- One suggestion was made that EPA use rural utility service data (FHA). This organization has a loan program for loans out to 40 years at very low interest rates.
- The comment was made that one of the requirements under the SDWA is that EPA set affordability criteria. This action has not yet been taken. When these criteria are set, a comparison could be done that might kick out certain technologies (e.g., RO, IE, or AA for arsenic), particularly for systems serving less than 100 people. These systems would then need to be removed from EPA's bigger costing analysis.
- EPA needs to identify existing practices that have an impact on reducing a contaminant level and provide credits for this removal.
8.3 Key Characteristics of "Outlier" Large and Small Water Systems
High Side Extreme Systems:
- It was noted that most of the extreme cases on the high side would be those systems using groundwater. As a result, it was recommended that EPA consider groundwater systems for the high side extreme cases.
- The high side extreme cases will involve combinations of contaminants rather than single contaminant scenarios. For example, there will be a number of high side extreme cases with radon, arsenic, and disinfection problems.
- The high side extreme cases will involve water systems that are located in urban areas where land is unavailable.
- Land will be the major cost element in extreme cases.
- EPA needs to consider retrofitting for high side extreme systems. Retrofitting should include the removal of old technologies.
- Systems falling into the low side extreme category are assumed to be those with clean groundwater sources that have constant water quality on a year-round basis. Moreover, these systems are assumed to use a package unit that requires minimal engineering costs.
- It was recommended that EPA conduct an analysis to determine the degree to which permitting must be undertaken for various system size categories. Participants in the breakout group noted their belief that EPA needs to provide data to support the 3 percent cut-off factor discussed earlier by workshop participants, especially since the percentage is not based on RS Means.
- In addition to requiring land for the technologies themselves, extreme case systems will also require land for other structures such as sound barriers and aesthetic structures. Therefore, EPA's costing with respect to land needs to look broader than at the technologies themselves.
- The installation of treatment requires permits to be acquired for air pollution and noise. As such, permit acquisition will drive the technology installation process for the extreme case systems.
- Permits might not be granted due to public opposition. Therefore, EPA may want to consider relocation costs in its models. If a system needs to have a treatment plant, land have to be found. Additional costs are likely to be incurred for public relations and consumer education.
- The location of some facilities might limit the use of some technologies (e.g., the use of RO in semi-arid and arid areas).
9. Other Capital Costing Issues
- It was noted that many sites have existing systems but might need to replace old technologies with new technologies. The opinion that retrofitting is often complicated was expressed. In some circumstances, the old system needs to stay in service while the new system is being built. These situations lead to increased costs.
- There is agreement on the need to deal with retrofitting costs but no clear consensus on how to deal with them.
- Retrofit costs must accommodate infrastructure changes within the plant. This includes cost adjustments factors on different things such as chemical feeds, adding new technologies such as new chemical feeds or adding new piping.
- Repumping costs need to be considered for GAC and other technologies where pressure is broken.
- During retrofit, construction costs might go up.
- Need to differentiate between capital costs and retrofitting costs. There is a need to look at existing construction cases to get data for strawman values.
- Issue: retrofit costs require some factor in the model due to the problem of dealing with existing plant and construction and maintenance of plant operations. This is higher with surface water plant than with groundwater plant.
- The frequency of occurrence of retrofitting in surface water systems is relatively high so it should be an easy task to find necessary data. The factors in the model may be much more appropriate in representative cases than in extreme cases.
- An issue for existing retrofit projects is how to parcel out actual expenditures. Expenditures could have occurred due to compliance reasons or could have occurred related to other activities. There should be a study to determine these allocations and the purpose of each, although not a simple task.
- In addressing the retrofitting issue, one must deal with cost models. When costs are added to model, there needs to be a reasonable basis for adding them. Don't want to add or remove too many technologies to the model. Need good intermediate strategy.
- Stakeholder meetings can add costs, which should be addressed in case studies.
- Education of the public should be included in cost estimates.
- Public involvement can be addressed as part of permitting process.
- There are significant regional cost differentials in labor.
- Location specific factors will influence labor costs.
- Areas with sensitive populations have additional costs. For example, a hospital or nursing homes will affect treatment operations.
- Group discussed differences between low cost and high cost estimates. High cost estimates are 95 percentile, not 99 percent.
9.2 Use Existing Models or Use WBS Build-Ups Using RS Means
- The fundamental difference between existing models and WBS models is that WBS involves a "build up" approach based on design parameters (pre-building process). Existing models are based on post-building process, which look at costs of facilities "as built."
- In existing models, some of the assumptions might not be representative of what the system actually is and percentages may not be a representative reflection of the average. There is a need to know how the numbers in the existing models were developed, if they are to be used.
- To calculate cost, design variables (user input) include the size of the pumps, the size of pipes, and the size of the basins. The user can look at different technologies to determine cost.
- WBS approach allows to build up from ground zero. The user can see all details that went into costing of each technology.
- Participants felt existing models should be relied upon for the coming three rules, but should be updated.
- People have a lot of experience dealing with WaterCost model.
- Transitioning to the use of the WBS model entails starting from ground zero in collecting data.
- To use existing models would require model enhancements.
- The time frame for change is one year.
- Issues of concern include
- time frame for WBS model availability, particularly data for assumptions required,
- comfort level in use of existing model,
- level of validation of existing model (comparison to what actually has taken place).
- Ideally, need to develop supplementary information that updates technologies and cost curves.
- One weakness of existing models is that cost estimates are based on one consulting engineers' views of water treatment plant process and design.
- Many design innovations have occurred since model (WaterCost, Water) inception. Should look at emerging technologies such as membrane, aeration, and ozone to get better data.
- Performance costs need to be updated to reflect each individual process.
- WBS offers the best framework for estimation and validation of cost.
- WBS allows the opportunity to have site estimate and can relay the data up to national model.
- Data that is captured in WBS can be utilized for other analyses or studies.
- WBS allows for stating assumptions up front and they can be included in model.
- Even rudimentary WBS estimates have added value to projects.
- In ideal world, WBS is the way to go.
- Concern exists regarding the amount of time it would take to put together WBS structures for all the scenarios in one year.
- Depending on resource availability, WBS may be the best tool for material cost estimation in the long term.
- WBS assumptions need to be improved. People need more information.
- Need WBS at least at level 3 and perhaps crosswalk to validate across models. Get finer resolution and more data over time.
9.3 Level of Detail if WBS Used and Additional Indirect Costs to be Included
- Existing models have hard wired percentages for indirect costs. There is a need to incorporate technology information properly in models since they are big cost drivers.
- An intermediate and potentially useful alternative would be to limit use of WBS to capture activities such as waste disposal, variability, and retrofitting.
- Take original cost curves and refit them with more up to date data to improve them statistically.
- The purpose is to cost out a representative design. WBS is site specific. Given this goal, WBS may not enhance modelling capability.
- Drinking water treatment will never have same goal as restoration. Cost model should only be used on site-specific basis.
- Specific elements of WBS require more detail and those may go to Level 4.
- There has to be a screen or certain level of impact to move those to Level 4.
- Are existing models suited for full plant design? WBS gives more flexibility in retrofit scenario.
- Quite a few WBS factors are low. For example, a non power intensive technology would require extensive rework if new technology is retrofitted.
- Power requirements at switchboards and control centers must be considered in determining electrical needs.
- Instrumentation per se is missing from WBS.
- Suggestion that electrical and instrumentation be put together and be 15-20 percent of capital costs.
- If we know the cost per linear foot to install piping (part of site work) then that can be included in the cost model.
- Road work, fencing, landscaping are included in site work.
- Group estimated 5-10 percent of construction costs for site work. This will be higher for retrofits. It may be appropriate to have multiplier for retrofit projects.
- Suggestion was made to consider site work as a percentage of construction costs with a basement.
- One participant suggested to key everything off of construction costs, using 8 or 10 major components of construction.
- WBS uses 16 different categories. Apply land incrementally.
- Input from panel on percentages of construction costs:
- Site work: 5-10 percent
- Electrical work and instrumentation: 15 percent
- Engineering Design: 15 percent (range is from 10 percent to 20 percent)
- Need a category for construction/project management. This is not included in administrative costs or contractors overhead costs: 10 percent
- Percentages depend on state of design.
- Process, that is whatever it is your installing, is the largest portion of costs.
- On smaller end of projects there are three major components of design builds: construction management, design, project management tasks. Contractors, builders, and manufacturer entities are involved. Actual price of system is 40 percent of whole project.
- In package plants, engineer costs are somewhat minimal.
- One case showed utility expense to oversee contractor is 3 percent.
- Costs may be approximated as 40 percent, 40 percent, 20 percent for process (including capital elements), construction (including mechanical, insurance and contingencies), and engineering Including design and construction management) respectively for smaller systems.
- Design/build is when owner decides not to go out to bid. There is no bid phase between design and build.
- Regarding design/build, one participant believes small and large systems are similar. In small systems, with design/build, systems tend to meet regulatory requirements and price reflects that.
- In large systems, there are other things that design/build process meets in addition to regulatory requirements such as facility for tours and training rooms.
- Between owner and project manager, owner has between 3-10 percent of total project costs.
- Permitting is part of design.
- Normally add 22-29 percent as an incremental add-on to construction cost for large private systems.
- Contractor costs should be considered to be included in construction costs.
- There is a difference in public and private sector.
- There is currently a document out developed by the Engineering Joint Council that discusses engineering fees as percentage of construction.
- One recommendation is that EPA use 25 percent for the low end private sector and 25-35 percent for the public sector for non-construction costs.
- Non-construction costs for small FHA water and wastewater plants in one state ranged from 25-35 percent of construction costs.
- 25 percent is probably low end for public sector.
- Nonconstruction costs for small FHA water and wastewater plants in one state range from 25-35 percent of construction costs.
- It was recommended that some percent needs to be added to capture costs to owner for their financial costs related to bonding.
- 3 percent legal and administrative fee would include owners cost of issuing bonds.
- 3 percent originated from small systems cost.
- Use 25-35 percent for large public sector systems for non-construction costs.
- 10 percent for contingencies should be applied after engineering and construction costs
10. O&M Costs
10.1 Labor Rates and Indirect Costs for O&M
- With regard to operator rates, $52-75 (loaded cost) for licensed operators was discussed. It was noted that costs can vary by system and by the qualifications required of the operators.
- One participant indicated that small systems do not always have full time operators.
- EPA used a $28 per hour basis, including fringe of 50 percent in its draft tables. Wage data for State of Minnesota were evaluated.
- It was noted that if a new project is undertaken with federal dollars, the wage must be the prevailing wage.
- One participant suggested a $14 per hour (unloaded) rate for small systems.
- It was noted that many small systems have contract operators.
- One participant indicated that the National Rural Water Association (NRWA) should have survey data on labor rates.
- It was noted that Ohio and Florida have statistics on labor rates, including reviews of prevailing labor rates.
- Another participant indicated that the Bureau of Labor Statistics (BLS) publishes monthly reports that include fully burdened labor costs that can be analyzed on a regional basis.
- It was recommended that EPA review and compare BLS and State labor rate material.
10.2 Labor Requirements of Various Technologies
- It was noted that different technologies require different levels of skilled/trained operators.
- A comment was made that reverse osmosis requires advanced operator skills, whereas microfiltration requires basic operator skills.
- Another comment was made that EPA's models seem to be more sensitive to hours than they are to rates. It was pointed out that EPA should be more concerned with hours for smaller systems.
- It was noted that it is often necessary for systems to make adjustments in staffing or shift their activities as operating hours increase.
- There was general agreement that EPA should only consider the incremental cost of maintenance for medium to larger systems.
- It was noted that each technology will have automation impacts; however, these impacts may not be felt by a large number of people.
- It was recommended that the calibration of turbidity meters should be listed under O&M.
- It was also noted that EPA should assume that only certified operators are doing the work.
- Comments were made that States may change their regulations if changes are made to operator certification guidance issued by EPA.
- It was noted that O&M should be considered as the only additional labor that comes with a new process. All other labor is substitute labor.
- It was noted that preventive maintenance is almost nonexistent in small systems. Repairs are made as problems are found. Small systems do not typically renovate or modernize.
- One participant indicated that the State of Washington attempted to identify operational labor requirements per technology under the Surface Water Treatment Rule (SWTR).
- The State of Washington also attempted to compare automation versus the use of labor based on SWTR requirements. They found small systems to require 1 to 2 hours per day and approximately 3 hours on Saturdays for technologies used to comply with the SWTR.
- One participant indicated that a weekly inspection visit is needed for groundwater systems required to disinfect.
11. Technology Design
Are EPA's technology specific assumptions and equations reasonable? What should the design defaults be for each technology? Are there additional design needs that should be identified?
- A recommendation was made that EPA look at an AWWA Journal article in 1996 that compared the construction and O&M costs of MF and UF.
- It was noted that AWWA has a manual of practices on NF and RO that should be issued shortly. The Association is supposedly working on a similar document for MF and UF. EPA was encouraged to incorporate these as they are finalized.
- RO, GAC, PTA, and ozonation require the extra step of repumping due to disruptions of the hydraulic pressure. It was noted, however, that many systems do not provide enough pressure to begin. As such, EPA should not have to cost the repumping since these systems should be providing the pressure in the first place.
11.2 Membrane Technologies
All Membrane Technologies:
- Pretreatment increases the life of the membrane and should be assumed.
- The figure in EPA's draft tables for maximum operating TDSs (1,000 ppm as NaCl) is possible but not realistic.
- EPA should assume that spiral membranes will generally be used.
- Hollow fibers have been used for brackish or sea water.
- Temperature can have a significant impact on the filters.
- EPA's first equation is for permeate flow. As such, should it not be depicted as Qp instead of Qd?
- For all processes, EPA needs to decide which design criteria it is going to use to determine costs. Rather than two pages of requirements, participants saw only five factors that needed to be evaluated for unit costing purposes. An example was provided for a membrane plant: (1) design flux (10 or 15gal/sq ft-day), (2) design pressure drop, (3) design cleaning frequency (at least 8 weeks), (4) design percent recovery (80-85 percent, the latter if aggressive), and (5) life of membrane (4 or 5 years).
- Cleaning is an important aspect of membrane technologies. Monthly cleaning was considered too frequent. One participant strongly felt that 8 weeks would be an acceptable range and that if a facility could not go this long between cleanings, it would likely be looking for another type of membrane to use. Another participant said that if cleaning is required more frequently than monthly, the system would be in trouble. Some systems can extend cleaning out to 10 to 12 weeks.
- Automation is not real feasible for membrane replacement. Replacement of the membrane can be one of the most significant components of a facility's total O&M costs .
- Workshop participants indicated that EPA's first assumption about pretreatment is a good one. At issue is how to determine what activities actually constitute adequate pretreatment.
- One set of adequate pretreatment activities are cartridge filters for groundwater and microfiltration for surface water.
- Membrane flux and pressure varies according to sources.
- The Yuma plant experience is that membrane useful life is unaffected by amount of use and that a certain amount of operating expenses are associated with maintaining the technology regardless of whether it is producing water.
- One workshop participant noted that he uses a default assumption of 80 or 85gfd for the recovery rate.
- Proposed flux rates ranged between 15 and 25 gfd. Workshop participants agreed that 15gfd was the best proportion to consider.
- Labor requirement values should be representative.
- Some discussion in support of EPA's reverse osmosis example: membrane life is 5 years; operating pressure is 200-1200 psi which is used for sizing of the pumps (default is 200 psi); maximum TDS is 1,000 ppm as NaCl; various type of membranes.
- Automation should not be assumed for RO cleaning.
- Regarding RO, pH adjustments should be assumed.
- Membranes that can tolerate higher pH levels are more costly.
- Waste generation depends on reject rates.
- One participant recommended 100gfd flux for MF and 60 gfd flux on UF.
- In MF model, EPA should think about the impact of the hydraulic backwashing cycle when sizing a unit. Individual units should be sized to reduce the overall required membrane design capacity.
- One participant indicated that recovery is less in UF than MF because of the use of a water versus an air-water wash. This participant commented that one could generally expect to see 95 percent recovery for MF and 90 percent recovery for UF.
11.3 Anion Exchange (AE)
- Anion exchange can be useful for arsenic removal.
- General agreement on assumption of 30-day supply of chemicals and 7-day storage time for wastes.
- Need to choose resin type on which to base design.
- An O&M issue raised by one participant is that certain resins, like activated alumina, can degrade and accumulate in the distribution system. This eventually will require the distribution to be flushed. Another participant indicated that the cited situation might be unique.
- One recommendation was that EPA assume a resin life of 2 years for both anion and cation exchange.
- Some utilities are looking at ion exchange for TOC removals. This application warrants different design assumptions if it is ultimately applied.
- Critical cost element to assume is the exchange capacity. For anion exchange, the exchange capacity is 2 meq/ml. Exchange capacity is dependent on how many gallons of water required to backwash one square foot of resin in relation to the backwash frequency, the regenerate requirements, the service rate, and the resin life.
- Resin life for anion exchange was discussed as 4 or 5 years, rather than decades. Agreement was reached by the group that 4 years was the most representative value.
- Participants agreed on a bed expansion rate of 75 percent.
- Backwash volume is important because volumes affect the size of the tanks that are installed.
11.4 Cation Exchange (CE)
- EPA's model does not include a softening component because it was designed for use with radium.
- CE is not the most effective technology for removing radium.
- Use 2 « times the nominal exchange capacity volume to calculate how much salt is required for CE regeneration for radium.
- The backwash rate is two times higher for CE than AE.
11.5 Activated Alumina (AA)
- Same design parameters as ion exchange except regeneration.
- Capacity decreases with regeneration in this technology.
- No clear indication of how much resin is lost over a 1-year time period.
- Capacity decreases and resin diminishes with time using AA. This affects treatment costs.
- Participants will provide EPA with background information on their experiences with AA.
11.6 Packed Tower Aeration (PT)
- Volumetric water loading rate of 25 gpm/ft2 is a very reasonable number.
- Primary design parameter is Henry's Constant for contaminant that is being removed.
- Air/water ratio is very low for radon.
- One way to cost PT is to use water application rate.
- If air/water ratio is correct, the packing translates into the removal required.
- In 10 State Standards, the ratio of the column diameter to packing is 10:1, the minimum air to water ratio is 25:1, and the maximum air to water ratio is 80:1.
- Some parameters must remain constant.
- Treatment for air streams may need to be considered.
12. Economic Issues
Are EPA's assumptions regarding equipment financing, escalation (or inflation) values, and cost allocations for premature retirements and outmoded facilities realistic?
- A comment was made that the cost of capital in the private sector is easier to determine than in the public sector because it was in PUC records.
- It was recommended that EPA look at requirements under FHA and SRF programs for input on its assumptions.
- It was noted that the duration of financing is dependent upon negotiations between the owner and the financing institution. Banks and lending institutions do not care about life of a process versus the life of equipment. They, instead, are concerned with a facility's assets and liabilities.
- Several participants felt EPA should make an effort to develop a life cycle based analysis for estimating capital needs.
- A comment was made that it is important to know what is in the base before costing a regulation.
- It was reported that the duration of private source financing is between 20 and 30 years for large systems and between 5 and 10 years for small systems.
- The comment was made that the condition and age of an existing plant are important factors in determining the cost of some regulations. For example, under the Enhanced Surface Water Treatment Rule (ESWTR), some under drains are to be modified. For old plants they are supposed to be modified anyway. Therefore, this modification is not a legitimate regulatory cost. A legitimate regulatory cost is one that requires new equipment or services (e.g., a new power line).