Watershed Management for Potable Water Supply: Assessing the New York City Strategy (2000)

January 21, 1997, marked an important event in the history of American water management: the signing of the mammoth New York City Watershed Memorandum of Agreement (MOA) which provides a legal framework for protecting the drinking water supply of nine million people. The culmination of years of negotiation between upstate and downstate interests, the MOA commits New York City to a long-term watershed management program that combines land acquisition, new watershed rules and regulations, and financial assistance to watershed communities to promote environmental quality and their local economies. Most important for New York City, the agreement currently satisfies provisions of the Environmental Protection Agency's (EPA) Surface Water Treatment Rule (SWTR) that will allow the City to avoid filtering its upstate Catskill/Delaware water supply until at least 2002.

Immediately following the signing of the MOA, the National Research Council (NRC) was asked by the New York City Comptroller's Office to provide a scientific evaluation of the watershed management program. The goal of the NRC study was to determine whether the MOA is based on sound science and to recommend improvements to strengthen watershed management for this large unfiltered supply (see Chapter 1 for the complete Statement of Task). This report is intended to inform New York City and other public water suppliers that are trying to maintain the purity of their existing water sources through proactive watershed management (regardless of whether they presently utilize filtration).

The NRC committee was specifically asked to address the following provisions of the MOA: (1) the use of setback distances to protect bodies of water from nonpoint source pollution, (2) the Total Maximum Daily Load (TMDL) program, (3) siting and technology requirements for wastewater treatment plants and septic

systems, (4) the phosphorus offset pilot program, (5) the enhanced monitoring program conducted by the New York City Department of Environmental Protection (NYC DEP), (6) New York State's antidegradation policy, and (7) NYC DEP's Geographic Information System (GIS). In addition to these tasks, the committee also evaluated the Land Acquisition Program and the MOA's requirement for comprehensive land use planning. The role of active disease surveillance in watershed management was explored, and a microbial risk assessment was conducted. Finally, the committee considered the potential impact of future changes in federal regulations regarding safe drinking water. As a result of its studies, the committee determined that the following issues should be taken into consideration as New York City carries out watershed management. More detailed conclusions and recommendations are found in this executive summary and throughout the report.

The watershed management program of New York City should be prioritized to place importance first on microbial pathogens, second on organic precursors of disinfection byproducts, third on phosphorus, and fourth on turbidity and sediment. The prime focus of New York City's efforts in protecting the health of its consumers should be on pathogens in the water supply and on developing means for their control. Because pathogenic microorganisms are the primary target of the SWTR, they should be the focus of any watershed management program designed to gain filtration avoidance. Currently, the main focus of New York City's watershed management strategy is phosphorus, because of its role in eutrophication and its contribution to the creation of disinfection byproducts (DBPs). Considerably less effort has been expended developing monitoring and modeling tools for microbial pathogens, which pose a more significant and direct threat to public health. After pathogens, New York City should focus on the reduction of DBPs, both existing and emerging. Because disinfection via chlorination is planned to continue, control of DBPs will require precursor (organic carbon) reduction. Precursor control can be partially accomplished by the current strong efforts to control phosphorus loading to the reservoirs, which should continue. Nonpoint sources should be the primary focus of future phosphorus control. Finally, greater effort should be made to control sources of inorganic turbidity and sediment at their points of origin rather than in downstream aqueducts and reservoirs. Controlling inorganic turbidity at its source has multiple benefits in relation to other priority pollutants.

The concept of balancing watershed rules and regulations with targeted support of watershed community development is a reasonable strategy for New York City and possibly other water supplies. The City's commitment of financial resources to promote environmentally sensitive development in watershed communities is a fair quid pro quo for enhanced land use planning and regulation intended to protect the drinking water supply. It is possible that resulting economic development will affect water quality. However, existing information convinced the committee that population growth in the Catskill/

Delaware watershed is limited. The potentially adverse effects of population growth and increased economic activity can be offset by careful planning, directed development, more extensive environmental regulation, and improved wastewater management, as provided in the MOA. Such measures will help to maintain high water quality in the Catskill/Delaware reservoirs over the next several years, assuming growth rates do not increase substantially.

The committee encourages New York City and all other water supplies to be receptive to the possibility of additional treatment options. The need for additional treatment may arise because of uncertainties about pollutant sources, better scientific understanding of the impacts of pollutants on human health, and changing regulations. At some point in the future, implementation of the MOA may not be sufficient to avoid additional treatment beyond disinfection. For example, New York City may be compelled to treat its water by coagulation/filtration or equivalent technologies to comply with more stringent regulations for DBPs and microbial pathogens in spite of its best efforts under the MOA. Management of these constituents may be infeasible with current infrastructure simply because increasing disinfectant concentrations to improve pathogen removal simultaneously increases the quantity of DBPs.

New York City should lead in efforts to quantify the contribution of watershed management to overall reduction of risk from waterborne pollutants. Watershed management is an essential component of a modern water supply system, but its direct contribution to risk reduction is particularly difficult to quantify. It is much less difficult to estimate the risk-reducing effect of specific treatment options, such as coagulation/filtration, membrane technologies, and alternative disinfection. Because of the comprehensive approach embodied in the MOA and because it has gained national prominence, New York City is in a unique position to demonstrate the overall risk-reducing effects of watershed management. This requires developing the concept of risk reduction as a metric of overall program success and conducting regular risk assessments.

New York City obtains its drinking water from three upstate watersheds—the Catskill, Delaware, and Croton systems. These watersheds and a complex infrastructure of reservoirs, aqueducts, and tunnels encompass 1,970 square miles, contain 600 billion gallons of usable storage, and provide as much as 2 billion gallons of water per day (Figure ES-1). Operational flexibility is afforded by interconnections among the systems that allow bypass of reservoirs and blending of water. In particular, water from the Catskill and Delaware systems is generally combined in the Kensico Reservoir, making this terminal reservoir critical to maintaining high-quality water. The combined system serves nine million people in New York City and adjacent Westchester County, the second largest service area of a single water supplier in the United States.

Since the mid-nineteenth century when it turned to the large and sparsely populated Croton River watershed, New York City has recognized the importance of high-quality source water in providing drinking water and protecting the public health. Because the water from the upstate systems historically has met drinking water standards, it has been delivered to consumers without coagulation/filtration, using only disinfection by chlorination. In recent years, however, the need for filtration of surface water supplies has been increasingly emphasized by regulators and public health experts. Since 1989, the SWTR has required every water supplier to filter its surface water sources prior to disinfection unless its source water meets specific water quality criteria and it establishes a watershed management program. In accordance with this regulation, New York City is preparing to filter water from the more heavily developed Croton watershed, which delivers approximately 10 percent of the City's total water supply. However, thus far the City has been able to satisfy regulatory criteria for water quality and gain a waiver from filtration for the combined Catskill/Delaware system, which delivers the remaining 90 percent of the supply. To maintain its waiver, New York City has embarked on an ambitious strategy of watershed management unprecedented in scope, scale, and cost.

Seventy-four percent of the Catskill/Delaware watershed is in private ownership. With a population that varies from 50,000 to 200,000 depending on the season, there is a fear that human activities, such as increased residential and commercial development, will degrade water quality. National concerns about risks of microbial pathogens such as Cryptosporidium, Giardia, and viruses in drinking water make the issue of water quality all the more sensitive. Balancing New York City's interest in protecting the watershed for high-quality water with the economic, recreational, and other interests of watershed residents has proven to be a significant challenge. Conflict between the City and upstate residents has been generated by efforts to protect the upland water supply source through land purchase and extensive regulation in the watersheds. The conflicts have largely been resolved by the 1997 signing of the MOA.

Multiple concerns about drinking water quality have spawned a plethora of federal, state, and local laws, regulations, and policies. Perhaps the most prominent concern is that drinking water will endanger human health, either through acute infections caused by microbial agents or through chronic exposure to chemicals. Although a wide variety of microbes have been linked to waterborne illness, at the present time the microbial pathogens of greatest concern are the protozoa Giardia and Cryptosporidium because of their resistance to chlorine disinfection. Chemical contamination of drinking water can also produce adverse health effects in humans. The most notable contaminants are DBPs such as trihalomethanes and haloacetic acids, formed in drinking water when chlorine interacts with dis-

solved organic matter in raw water. Exposure to DBPs has been shown to be related to certain cancers and to miscarriages (Chapter 3).

These concerns, as well as aesthetic considerations, have prompted the creation of environmental laws that directly and indirectly govern the supply of drinking water and watershed management in New York City and across the country. The Safe Drinking Water Act of 1974 (SDWA) sets allowable maximum contaminant levels (MCLs) for multiple contaminants in finished water. The SWTR, pursuant to the SDWA, was created to address the health risks of drinking water derived from surface supplies, particularly those risks associated with microbial pathogens. It describes the criteria that must be met by surface water supplies that want to forgo filtration, including the creation of a watershed management program, meeting standards for turbidity and for fecal and total coliforms, providing adequate disinfection, and avoiding any waterborne disease outbreaks.

The Clean Water Act (CWA) requires the nation's waters to meet water quality levels that are fishable and swimmable. In accordance with the Act, states must designate waters according to use classifications and set specific water quality criteria for those classifications. For those waters not meeting water quality standards, the Act requires that TMDLs be developed to identify sources of pollution and describe mechanisms for reducing them. The National Pollutant Discharge Elimination System (NPDES) issues permits to wastewater treatment plants (WWTPs) and stormwater discharges, ensuring that effluent standards are maintained and best available control technologies are used. Thus, while the SDWA targets public health and the quality of drinking water at or near the point of use, the CWA focuses more on the quality of source waters and discharges into those waters.

State and local environmental regulations also play a role in providing high-quality drinking water to New York City. The Watershed Rules and Regulations of the MOA fill gaps between the CWA and the SDWA by placing specific restrictions on watershed activities. Additional oversight of watershed activities is provided by the New York State Environmental Quality Review Act (SEQR), which creates a role for public decision-making regarding proposed projects through its requirement for an environmental impact statement.

Watershed management is recognized as a key part of the multiple-barrier approach to providing safe drinking water (see Chapter 4). Source water protection specifically refers to management of water supply watersheds, both surface and groundwater. An ideal source water protection program contains several components. First, it should have defined program goals and objectives. An inventory of the watershed to assess existing and potential contamination is also necessary. Once sources of contamination have been identified, protection strat-

egies must be chosen to reduce contamination and prevent future pollution. Protection strategies require proper implementation, a component of source water protection that is often overlooked and sometimes responsible for program failure. A final component of any source water protection program is effectiveness monitoring and evaluation of monitoring data. To be successful, stakeholders must be involved in all stages of a source water protection program. These components exist to varying degrees within the MOA and other watershed management programs.

Common pollutants in the New York City watersheds include microbial, chemical, and physical parameters. The pathogenic protozoa Giardia and Cryptosporidium have been detected at low levels in the Kensico Reservoir (see Figure 6-6) and in stream sites throughout the Catskill and Delaware watersheds. Although both nitrogen and phosphorus can cause eutrophication of waters, phosphorus is the more important nutrient in the New York City reservoirs because it is limiting to algal growth during most of the year. Natural organic carbon compounds, which serve as precursors for many undesirable DBPs, are found in all the reservoirs. Sediment transport to reservoirs is primarily a problem in the Schoharie and Ashokan watersheds, where steeper slopes and fine-textured soils are more prevalent. Finally, toxic compounds are used in specific regions of the Catskill/Delaware watershed.

The pollutants described above are derived from both point and nonpoint sources. Point sources, consisting primarily of 41 WWTPs, occur throughout the Catskill/Delaware watershed. Nonpoint sources of pollution are more wide-spread and abundant in the watershed. Undisturbed forests, comprising almost 70 percent of the Catskill/Delaware watershed, are sources of nutrients, natural organic carbon compounds, sediment, and pathogens. Agriculture, the second largest land use in the Catskill/Delaware watershed, is a potentially significant contributor of microbial pathogens, nutrients, and pesticides. Although urban areas comprise a relatively small percentage of land in the Catskill/Delaware watershed, they can be significant sources of nutrients, pathogens, and pesticides, as well as of heavy metals and other toxic compounds. The Catskill/Delaware watershed includes over 30,000 septic systems (on-site sewage treatment and disposal systems, or OSTDS) that can be sources of both microbial pathogens and nutrients. Finally, atmospheric deposition over land and water is a probable source of nitrogen and mercury to the watersheds that is extremely difficult to quantify and control.

Measures of reservoir water quality indicate a chronic problem with the eutrophic health of the water supply reservoirs, particularly the Cannonsville Reservoir and the Croton system reservoirs. All 19 reservoirs are classified as

either mesotrophic or eutrophic, and some have variably high concentrations of phosphorus, chlorophyll a, and turbidity. The Kensico Reservoir had elevated levels of fecal coliforms between 1990 and 1993, which were attributed to bird populations. In addition, new commercial development and road expansion in the Kensico watershed may threaten water quality in this reservoir in the near future. Despite these conditions, source water and drinking water in New York City have historically been and continue to be in compliance with the SDWA. Catskill/Delaware source water and associated tap water are meeting requirements for total and fecal coliforms and trihalomethanes, and there have been no documented outbreaks of waterborne disease in New York City since 1974.

Water quality monitoring has four general objectives. Compliance monitoring evaluates physical, chemical, and pathogenic biological parameters for compliance with local, state, and federal regulations. In almost all cases, the water supply has met the requirements for physical, chemical, and biological parameters. Operational monitoring is conducted on a broad set of parameters needed to assess the ambient quality of water and reservoir dynamics, and to determine the sources of pollution that influence water quality. New York City's monitoring program measures a wide range of parameters in streams, subsurface flow, reservoirs, aqueducts, and the distribution system. Performance monitoring is used to evaluate the effectiveness of watershed management practices and policies and to isolate design factors that influence pollutant removal. This type of monitoring in the New York City watersheds is confined to a few specific studies. Finally, monitoring data can be used to support modeling of projected changes in water quality under different conditions related to land use change or watershed management actions. Monitoring for model calibration and validation involves both intensive and extensive sampling of a reservoir and its watershed to define parameter values, set initial conditions, and physically characterize the watershed.

In general, the committee finds the monitoring program to be informed, extensive, and of high quality, as is appropriate for a water supply of this size. An extensive review of the enhanced monitoring program led to the following suggestions for improvement.

Monitoring should be conducted on the basis of discharge-mediated volume rather than on fixed intervals for stream flow, shallow subsurface flow, WWTP effluent, and precipitation analyses. Even-based or flow-proportional sampling is needed to capture rapid variation in water flow and water quality constituents, including microbes, total and dissolved organic carbon, sediment, and nutrients. Although some flow-proportional monitoring is under way, more efforts and resources should be directed toward converting fixed-frequency sampling to storm-event sampling.

Shallow subsurface and groundwater parameters should be monitored regularly throughout the reservoir watersheds. This is particularly important given the prominent role of agriculture, the high density of OSTDS, and the potential for leaking sewer lines. At the present time, groundwater sampling occurs only in the Kensico watershed. New routine groundwater sampling should be integrated with direct experiments on the efficacy of OSTDS and riparian buffer zone management.

With the exception of dissolved organic carbon, the analytical methods employed for physical, chemical, and biological parameter monitoring are generally adequate. Pathogen-monitoring techniques must be improved to achieve significantly higher numbers of nonzero measurements for (oo)cysts, to assess overall recovery efficiency on a weekly basis, and to build split samples into the routine sampling plan. Additional microbial parameters should be considered for routine monitoring, including E. coli, F⁺ coliphage, Clostridium perfringens, and cyanobacteria that produce toxins.

Monitoring and other special pollution prevention activities in the Kensico Reservoir and its watershed should continue at their present high intensity given the importance of this watershed in controlling Catskill/Delaware drinking water quality. Similar measures should be considered for the West Reservoir and watershed, which is undergoing more rapid population growth.

Active disease surveillance for giardiasis and cryptosporidiosis began in New York City in the mid-1990s. This activity involves the compilation of case rates from City medical laboratories followed by case interviews to obtain information about suspected risk factors. Current endemic rates of these infections are significantly lower than rates in previous years, primarily because of fewer infections among immunocompromised persons.

Active disease surveillance is generally too slow and insensitive to detect an ongoing disease outbreak, unless the outbreak is very large. To better detect outbreaks in real time, New York City is monitoring sales of antidiarrheal medications, surveying health of nursing home patients, and monitoring laboratories for the number of stool samples submitted. These complementary systems can indicate sudden increases in gastrointestinal illness, which may be indicative of a waterborne disease outbreak.

Linking observed disease to drinking water as the route of transmission can only be accomplished with specific epidemiologic studies. New York City has conducted a case-control study for giardiasis and a cross-sectional study of cryptosporidiosis, in which no relationships between drinking water habits and infection were observed, possibly because the study populations were not large

enough. In general, New York City is making concerted efforts to conduct active surveillance and outbreak detection for giardiasis and cryptosporidiosis.

To improve the sensitivity of active disease surveillance, health care providers and laboratories should make Cryptosporidium testing part of all routine stool examinations. Active disease surveillance as practiced by New York City and other communities is too insensitive to detect waterborne disease outbreaks on a real-time basis. Data from other communities suggest that only 1 in 22,000 cases are diagnosed and reported via disease surveillance during waterborne disease outbreaks.

New York City should determine the lowest incidence of disease that can be detected by the current outbreak detection program and increase the sensitivity by studying specific populations. It is not clear what size of outbreak can be detected using the combined system of monitoring sentinel nursing home populations, sales of antidiarrheal medications, and submission of stool specimens. To increase the sensitivity and population base covered by these systems, New York City should consider monitoring school absenteeism related to gastroenteritis in sentinel schools, monitoring Health Maintenance Organization nurse hotline calls to track increased incidence of gastroenteritis, monitoring hospital emergency room visits for gastroenteritis, and monitoring gastrointestinal symptoms in a network of sentinel families.

To determine the role of tap water as a vehicle of infection, New York City should conduct additional epidemiological studies. New York City is in an excellent position to conduct a randomized household intervention trial using some type of in-home water treatment device. Although such studies are complex and expensive, they are a rigorous and definitive way to quantify the potential for waterborne disease, and they provide valuable documentation of the safety of the water supply.

Active disease surveillance is one analytical tool to estimate potential disease impacts from pathogens in water. A complementary tool is quantitative microbial risk assessment. Although risk assessment is not being done by New York City on a regular basis, there are sufficient data being collected to use this technique to estimate potential disease impacts.

The risk assessment conducted by the committee used data on Cryptosporidium oocyst concentrations in the Kensico Reservoir from 1993 through July 1998. The daily risk estimate considering all years of data and all sources of uncertainty was 3.4 × 10^-5, with a 95 percent confidence interval ranging from 3.4 × 10^-7 to 21.9 × 10^-5. For an exposed population of 7.5 million, this would translate into an estimated 255 infections per day. See Chapter 6 for details and a description of several caveats that must be kept in mind when

reviewing and interpreting these results. The following recommendations regarding risk assessment are made.

It is recommended that a Cryptosporidium risk assessment be performed on a periodic basis for New York City. The goal of these efforts should be to help determine the contribution of watershed management (vs. other treatment options and management strategies) to overall risk reduction. Data that are sufficient for these purposes are currently collected as part of the NYC DEP Pathogen Studies. Prior to commencing this regular effort, a decision must be made as to what level of risk is deemed to be acceptable to the regulatory agencies, the City, and the affected parties. This level should be arrived at after full and open discussion with the various stakeholders.

An ongoing program of risk assessment should be used as a complement to active disease surveillance. Risk assessment allows one to ascertain the level of infection implied by a very low level of exposure that would go undetected by active surveillance, thus acting as a complementary source of information about public health. In general, periodic risk estimates should be examined for concordance with prior computed risks and observed illness rates in formulating subsequent water treatment and watershed management decisions.

The goal of the Land Acquisition Program is to solicit up to 355,000 acres of sensitive lands in the Catskill/Delaware watershed. Land must be obtained on a willing buyer/willing seller basis, and conservation easements are an option in lieu of direct purchase. To date, the program has over 15,000 acres under contract, and it has made offers on over 100,000 acres. Other program elements to promote water quality include a flood damage buyout program, agricultural easements, and stewardship initiatives. Areas for acquisition within the Catskill/Delaware watershed have been prioritized based entirely on their proximity to bodies of water and the distribution system. In addition, the program sets minimum size limits from one to ten acres for all priority areas.

The committee recommends that greater use be made of the GIS and available land use/nonpoint source computer models to determine more precisely which areas are critical to water quality and thus should be protected through acquisition.

Lower limits (from one to ten acres) on the size of parcels that can be solicited may exclude environmentally sensitive lands from acquisition. It is recommended that these limits be relaxed and eventually dropped as the number of larger parcels diminishes and small adjacent parcels become available.

A local land trust for the Catskill/Delaware watershed region would be beneficial to furthering the goals of the land acquisition program. Currently, the methods for managing the lands acquired by the City, both through outright purchase and conservation easement, are complex and require oversight from the

City and State. A land trust can move more quickly than the Land Acquisition Program to acquire or accept gifts of land, and it may be attractive to those landowners who simply do not wish to sell to New York City.

The MOA Watershed Protection and Partnership Programs, most of which are administered by the Catskill Watershed Corporation (CWC), are intended to maintain and improve water quality while preserving the character and economic vitality of watershed communities. Land use plans are prerequisite to participation in some of the programs, in particular the new sewage treatment infra-structure program, the sewer extension program, and the phosphorus offset pilot program. Although each of these programs is in the initial stages, areas for improvement are apparent. First, the planning efforts under way are fragmented. That is, none of the efforts are truly comprehensive (i.e., housing, transportation, public utilities, environmental conditions, and other relevant factors are considered). This fragmentation is compounded by the numerous small constituencies responsible for plan development. Second, planning efforts do not require enough citizen participation to foster success. Finally, strategies for successfully implementing the plans have not been described.

The committee recommends that the CWC review local and county plans for the watershed area to make sure they are compatible. To coordinate planning efforts currently applied under the MOA, one alternative would be to create comprehensive plans for those watershed towns that do not have them, and to review and update existing plans to be in accord with the newly created plans. At the next level, counties should attempt to integrate the plans for towns and villages within their jurisdictions, and the CWC should review the county plans for compatibility and workability.

Expanding the mission of the CWC would allow it to play a greater role in local planning efforts of the smaller communities in the Catskill/Delaware region. The CWC could act as a forum for citizen involvement, provide technical assistance to support local implementation and enforcement, and provide a comprehensive framework for local watershed plans.

As required by the City's filtration avoidance determination, TMDLs of the priority pollutant phosphorus have been calculated for all of the New York City water supply reservoirs in two phases (I and II). Several Croton reservoirs and the Cannonsville Reservoir consistently exceed their TMDL. According to land use model results, reductions in phosphorus loading from both point and nonpoint sources will be required to meet Phase II TMDLs in these reservoirs.

The Phase II TMDL calculations are a significant improvement over Phase I calculations because more accurate land use models were used and a more appropriate guidance value for total phosphorus in the reservoirs was chosen. Phase III

TMDLs are expected to be fully time-variable and spatially variable, and models to accomplish this are under development.

The 15-µg/L total phosphorus guidance value is appropriate for the Phase II TMDL process. The Phase I goal of 20 µg/L was not adequately conservative for a drinking water supply, as it is based on ecological and aesthetic considerations. Conservatism in the choice of a phosphorus standard is necessary because data for some of the New York City reservoirs (e.g., Cannonsville) show that algal productivities, estimated from average and maximal summertime algal biomass as chlorophyll a, are in excess of recommended values for drinking water systems.

NYC DEP should place a high priority on implementing all necessary nonpoint source control measures to reach Phase II TMDLs. The most frequently cited criticism of TMDL programs across the country is the failure of states to implement pollution control measures following TMDL calculations. Currently, it is unclear what specific measures are being taken to reduce phosphorus loading to the Cannonsville Reservoir and several Croton reservoirs other than WWTP upgrades.

Phase III of the TMDL program should focus more on public health protection by developing models that link phosphorus to DBP precursors and other relevant parameters. New York City is currently constructing input–output models for total phosphorus, measurements of dissolved and particulate phosphorus loadings and concentrations, and assessment of individual phosphorus retention coefficients in the Cannonsville Reservoir. Deterministic models are needed to link phosphorus and dissolved organic matter derived from inreservoir and terrestrial sources to DBP formation potential, algae, chlorophyll, and taste and odor.

The MOA includes a five-year phosphorus offset pilot program that allows for the construction of up to six new WWTPs in pollution-sensitive basins. The purpose of this program is to allow for continued growth in these watersheds, while preventing a net increase in phosphorus loading. The offset program is similar to watershed-based pollutant trading programs for water and emissions trading programs for air, in which a discharger of pollution is allowed to increase its discharge if another party will concomitantly reduce its discharge. The offset ratio indicates the amount of pollutant reduction that must occur to balance the new pollutant discharge, either 2:1 or 3:1 in the New York City program. A variety of offset mechanisms can be used to reduce pollutant loadings from other point and nonpoint sources. Recommendations for program improvement include the following.

Baseline, minimum requirements for phosphorus reduction must be in place and operating effectively before additional reductions can be defined

as surplus. This refers to such activities as all planned upgrades to WWTPs that will reduce phosphorus loadings and to phosphorus reduction mechanisms that are required as part of a Stormwater Pollution Prevention Plan, among others. Criteria are needed to determine whether baseline requirements are in place and operational and to further define surplus reductions that can be used as offsets.

New York City should reevaluate the 3:1 and 2:1 ratios and develop a technical basis for the ratios that reflects the unique conditions associated with specific proposed offset mechanisms. In addition to providing a safety margin for variability in the performance of best management practices (BMPs) and error in data collection, the offset ratio should reflect conditions present in the New York City watershed such as the spatial and temporal variability of offset mechanisms, the relative locations of the offset mechanism and the WWTP, and the different forms of phosphorus produced in the effluents of WWTPs versus the offset mechanisms. Because the ratios do not explicitly take these issues into consideration, they may not afford sufficient protection.

There is no evidence that the phosphorus offset pilot program will result in a net reduction in phosphorus loadings to the water supply reservoirs, because the offset ratios were not established with this intent. If a net reduction in phosphorus loading becomes a goal of the program, the offset ratio should be made more conservative.

Antidegradation is a federal regulation related to the CWA stating that waters must not be allowed to degrade in quality. It deals primarily with ''assimilative capacity"—the amount of additional pollution that a waterbody can receive without exceeding its water quality criteria. States must develop an antidegradation policy that distinguishes three levels of water quality: Tier 1 (the lowest level of quality), Tier 2 (fishable and swimmable waters), and Tier 3 (outstanding natural resources). Tier 1 waters cannot be allowed to degrade any further, and their existing uses must be maintained. In Tier 2 waters, water quality is only allowed to degrade to the level of fishable/swimmable and only if important social or economic reasons for the lowering of water quality are given. Tier 3 waters, which have exceptional recreational and ecological significance, cannot be allowed to degrade at all. In New York State, antidegradation is implemented via three state regulations: stream reclassification, the State Pollutant Discharge Elimination System (SPDES) permitting program, and SEQR.

The SPDES Permit Program and the SEQR process are adequate tools for implementing antidegradation, but they could be improved with some revisions. For the most part, these regulations take into account the reasonable alternatives to a proposed action, the social and economic benefits of a proposed action, and the significance of potential environmental impacts, all of which are required under federal regulations. However, additional guidance on the types of

economic and social benefits that should be part of environmental impact statements is needed. In addition, explicit consideration of a receiving water's assimilative capacity should be required as part of environmental impact statements. Because SEQR is the primary avenue for regulating new nonpoint sources that will adversely affect water quality in New York State, this requirement for addressing assimilative capacity is critical.

New York State should better define what a significant lowering of water quality is in a Tier 2 water. That is, it should set quantitative criteria for altering the assimilative capacity of a body of water. Other state antidegradation programs suggest that a 5 percent to 25 percent change in a water's assimilative capacity is significant.

In addition to various watershed management activities, the MOA requires a series of studies on coagulation/filtration of the Catskill/Delaware system and a study on alternative disinfectants that might be used in lieu of filtration. Simultaneously carrying out watershed management and planning for additional treatment, such as coagulation/filtration and alternative disinfection, is a relatively new policy concept that has been dubbed the dual-track approach. This approach was designed to ensure no time is lost if other treatment processes are later determined to be necessary. Preliminary results from these studies led to the following recommendations regarding future additional treatment for the Catskill/Delaware supply.

The dual-track approach allows New York City to focus most of its resources on watershed management at this time. Watershed protection reduces source water pollutant concentrations that would be treated by future additional treatment. Thus, watershed management can reduce the cost of filtration. For this reason, future decisions to construct treatment processes beyond disinfection should in no way deter New York City from pursuing watershed management.

The results of the filtration pilot plant study show the New York City water supply can be effectively treated via coagulation/filtration. Treated water from direct filtration had a low turbidity and low total trihalomethane and haloacetic acid formation potential. A removal rate of at least 3 log for Cryptosporidium oocysts is expected for the entire treatment train (corresponding to a 3-log reduction in risk from potential waterborne Cryptosporidium). This level of treatment effectiveness depends upon high source water quality. Finished water quality would be expected to be even better with improvements in source water quality.

Additional studies to assess the potential of ozone as a treatment technique are needed. Ozone has the potential to increase levels of biodegradable organic carbon in finished water and foster regrowth of heterotrophic plate count

organisms and possibly coliforms, although distribution system disinfectant residuals may counter this phenomenon. Thus, ozone should not be adopted as the only treatment other than a distribution residual unless the City undertakes an investigation to satisfy itself that bacterial regrowth can be kept in check.

Nonpoint source pollution is, by definition, widely dispersed in the environment and is associated with a variety of human activities that produce pollutants such as nutrients, toxic substances, sediment, and microorganisms. Activities that generate nonpoint source pollution cause changes in vegetative cover, soils, or flow paths that reduce the ability of the land to naturally remove pollutants in stormwater. Three programs within the MOA that address nonpoint source pollution are reviewed in this report: (1) the Watershed Agricultural Program, (2) the Watershed Forestry Program, and (3) Stormwater Pollution Prevention Plans. For each, implementation and long-term maintenance of BMPs will be critical to program success.

The Watershed Agricultural Program (WAP) is a voluntary program intended to standardize and improve environmental practices among farmers within the watershed. Unlike most other activities, agriculture is exempted from the Watershed Rules and Regulations in the MOA, including requirements for setback distances, discharge permits, and rules regarding pesticide application. The WAP is intended to "substitute" for such regulations, and its continuance is required by EPA's filtration avoidance determination for New York City.

The goal of the WAP is to design and implement Whole Farms Plans—comprehensive strategies for controlling pollution at individual farms in the watershed. Because dairy farms are the most common farm type in the Catskill/Delaware watershed, phosphorus and pathogens are the pollutants of concern, and common BMPs include barnyard stormwater management, improved manure storage, and the separation of calves from cows. Currently, program success is directly related to the numbers of practices installed and the numbers of farms participating.

A scientifically based phosphorus load reduction goal that will achieve the desired phosphorus concentration in the water supply reservoirs is needed for the WAP. The purpose of having this load reduction goal is to be able to apportion reductions among different phosphorus outputs from individual subwatersheds and farms. Simulation models that take into account in-reservoir generation of phosphorus, phosphorus cycling, and different forms of phosphorus can be used for this effort.

Although the WAP has met its required milestones (including over 95

percent farmer participation), these metrics do not indicate the net effect of the WAP on water quality. To accomplish this critical task, additional monitoring is needed at various scales. First, the use of demonstration farms for evaluating whole farm plans should be expanded beyond the two sites now being studied. Second, monitoring to determine the effectiveness of BMPs in reducing pathogen and phosphorus concentrations in nearby streams is urgently needed.

Lands enrolled in the newly established USDA Conservation Reserve Enhancement Program should be prioritized based on (1) frequency of flooding, (2) vegetation type, and (3) whether the landowner will voluntarily exclude livestock from riparian zones. If prioritization is not possible, rental and cost-share incentives should be increased to retire frequently flooded farmland into riparian forest buffers and to exclude livestock from streams.

Modeled after the WAP, the voluntary Watershed Forestry Program (WFP) was created in 1996 to improve the economic viability of forestland ownership and the forest products industry in ways compatible with water quality protection and sustainable forest management. Although the WFP is new, progress toward program objectives has been rapid and substantial, mainly in areas such as program administration, education of potential program participants and foresters on the use of appropriate BMPs, and development (by independent scientists) of monitoring techniques to assess BMP effectiveness. To date, 78 landowners and a combined area of 21,000 acres are enrolled in the program.

Water quality monitoring should be implemented on the model forests and in conjunction with NYC DEP and New York State Department of Environmental Conservation (NYS DEC) monitoring networks on other tributaries. The WFP is encouraged to merge landowner, management plan, and field assessment data with water quality data to evaluate program performance.

The committee recommends that New York State consider tax policy changes that would promote sustained management of private forestland including those forests on relatively small parcels. A major impediment to the sustainability of forestry in the Catskill/Delaware region is the imbalance between taxes and expected revenues, such that short-term financial pressures discourage long-term investments in forest conservation. Because the benefits of the WFP are likely to extend beyond the local political boundaries, impacts on town revenues should be evaluated and mitigation payments to local governments by the State should be made as an integral part of the program.

Most types of new, large-scale (greater than five acres) development in the New York City watershed region are required to submit a Stormwater Pollution

Prevention Plan (SPPP) for controlling the quantity and quality of stormwater runoff generated by new impervious cover. SPPPs specify BMPs that will prevent erosion and sedimentation during construction and prevent any increase in the rate of pollutant loading via stormwater after construction. They must include a quantitative analysis demonstrating that stormwater quality from post-construction conditions will be better than or equal to that of pre-construction conditions.

Prior to the MOA, fewer activities required the drafting of an SPPP, and the regulatory oversight was spread among multiple agencies. SPPPs have spawned confusion among engineers, developers, and local and state agencies about exactly how the plans should be interpreted and implemented, as most of these organizations have had no prior experience with stormwater quality control and/or stormwater BMP design. The report includes a variety of specific suggestions to improve the quality of SPPPs and promote their successful implementation.

For removal of both phosphorus and bacteria from stormwater runoff, current structural BMPs advocated by SPPP guidance material are only moderately effective. In almost all cases, they cannot reduce post-development loading to pre-development levels. At present, the only stormwater retrofit BMPs whose performance is quantifiable are ponds, wetlands, sand filters, and swales. To counteract these deficiencies in structural urban stormwater BMPs, SPPPs should also encourage use of nonstructural BMPs that will limit the amount and impact of impervious surfaces.

New York City should embrace a performance-based approach to stormwater management rather than the permit-based approach embodied by the current SPPPs. The SPPPs should rely less on quantitative, theoretical calculations and more on performance monitoring and strict requirements for BMP size and treatment efficiency. Guidance material for such a new approach should include information on performance monitoring of stormwater BMPs for a variety of pollutants.

One of the most prevalent features of the MOA Watershed Rules and Regulations is the use of setback distances for separating waters from potentially polluting activities. For activities such as storage of hazardous waste and petroleum products, siting of septic systems and landfills, and creation of new impervious surfaces, 25–1,000 ft of land must separate the activity from nearby reservoirs, reservoir stems, controlled lakes, wetlands, and watercourses.

The committee assessed the adequacy of these setback distances in three ways. First, it compared them with similar regulations found elsewhere. Second, information on pollutant removal in buffer zones was used to predict the effectiveness of setbacks. (Buffer zones are natural or managed riparian areas that protect waterbodies from adjacent nonpoint sources of pollution by retaining or transforming pollutants and producing a more favorable environment for aquatic

ecosystems.) Finally, a predictive model analysis of pollutant movement through a setback was used to simulate the functioning of setbacks. All three of these evaluations are necessarily indirect because of the absence of actual monitoring data on pollutant concentrations in stormwater emanating from different land uses in the Catskill/Delaware watershed.

The setbacks prescribed in the MOA are not equivalent to buffer zones. Setback descriptions do not discuss the characteristics of setback land known to influence pollutant removal in buffer zones, such as slope, hydraulic conductivity, soil moisture, vegetation or surface roughness, and flow rates. When delineating setback areas, New York City should consider setting a slope threshold above which land cannot be included in setback considerations. Areas with a slope greater than 15 percent generally do not function as effective buffer zones.

Active management of setbacks is necessary to achieve the pollutant removal efficiencies ascribed to buffer zones. To improve the functioning of setbacks, BMPs that will convert channelized flow to sheet flow should be installed upslope from the setback. Setbacks should have appropriate vegetation to retain nutrients, sediment, and other pollutants. If setbacks are managed as buffers, periodic harvesting of vegetation may be necessary to remove accumulated pollutants and renew assimilative capacity.

Based on an analysis of travel times, existing literature, and a model analysis, the following setbacks are judged most likely to be inadequate: (1) the 100- and 500-ft setbacks for hazardous wastes, (2) the 100- and 500-ft setbacks for petroleum underground storage tanks, (3) the 100- and 500-ft setbacks for heating oil, (4) the 250- and 1,000-ft setbacks for landfills, and (5) the 100-ft setbacks for septic systems, WWTPs, and impervious surfaces because of microbial pathogens.

Because of their variable pollutant removal abilities, buffer zones should not be relied upon to provide the sole nonpoint source pollution control and are instead best used when integrated with appropriate source controls on pollutant releases. Although riparian buffers can ameliorate nonpoint source pollution in some circumstances, they are most effective when used as part of an overall pollution control or conservation plan.

Treatment and disposal of wastewater in the Catskill/Delaware watershed is a major factor in determining the quality of New York City drinking water. This is because almost all wastewater from the region is discharged either directly into streams tributary to the water supply reservoirs or into the subsurface where it can eventually reach the reservoirs. Individual septic systems (OSTDS) and centralized WWTPs are the major sources of wastewater in the watershed. The committee (1) used a qualitative approach to determine whether the watershed can sustain new WWTPs and OSTDS without declines in water quality, (2) evalu-

ated the effluent standards for WWTPs mandated by the MOA and technologies for WWTPs and OSTDS, and (3) considered whether the rules governing the locations of new WWTPs and OSTDS are adequate.

Upgrades to WWTPs mandated by the MOA and the use of best available control technology for OSTDS should be effective in reducing effluent loadings of phosphorus, total suspended solids, coliforms, viruses, Giardia cysts, and Cryptosporidium oocysts. Although declines in water quality related to WWTPs and OSTDS will occur with population growth, in most cases from 40 to 100 years will pass before contaminant contributions from these treatment systems reach present levels. To arrive at these conclusions, best available control technology for OSTDS was assumed to be aerobic treatment units with essentially a zero failure rate.

Current technologies being used for new and replacement OSTDS in the Catskill/Delaware watershed are not adequate; they do not represent best available control technologies. Implementation of aerobic treatment systems for OSTDS, including a significant enforcement effort. could substantially reduce effluent concentrations of Giardia, Cryptosporidium, fecal coliforms, and viruses in all Catskill/Delaware watersheds. Therefore, aerobic treatment units should be mandated for new or replacement OSTDS, and enforcement efforts should include annual inspections. This recommendation is especially important for the Kensico watershed, because of its critical location in the water supply and because OSTDS serves a large percentage of the population.

To ensure that the Continuous Backwash Upflow Dual Sand Filtration units being used at WWTPs represent the best available control technology, these units should be subjected to rigorous long-term monitoring to verify that equivalency with microfiltration is maintained. Monitoring of particle counts and turbidity should be conducted to determine the effectiveness of the filtration process and backwashing, and it should detect operational problems that may occur, such as clogging of the filter.

The current 60-day value for siting WWTPs does not appear to be supported by available knowledge. A scientific rationale for such a zone of protection should be developed. Limited research suggests 60 days may be inadequate for significant inactivation of Cryptosporidium oocysts and Giardia cysts, which are known to be resistant to disinfection, particularly at low temperatures. New York City should undertake further studies of pathogen fate and transport that will refine this value.

The New York City MOA, which outlines an ambitious program of watershed management, is a unique document in the history of water resource management. The program it advances is a prototype of the utmost importance to all water supply managers. The MOA provides for an extraordinary financial and

legal commitment from New York City to (1) prevent existing and potential contaminants from reaching the source reservoirs, (2) monitor a broad range of water quality and drinking water parameters, (3) conduct new research on water-borne health risks and water quality, and (4) promote the economic and social well-being of the Catskill/Delaware watershed communities in an environmentally sustainable manner. The committee is encouraged by the evidence of progress achieved to date in realizing the goals of the MOA. Although its limitations are noted in this report and it is not a guarantee of permanent filtration avoidance, the MOA is a template for watershed management that, if diligently implemented, will maintain and improve source water quality.