Watershed Management for Source Water Protection
Watershed management can be defined as any program or collection of strategies that positively influence activities and land characteristics within a drainage basin. Watershed management is necessary for many reasons, including meeting requirements of the Clean Water Act (CWA) (to maintain the physical, chemical, and biological integrity of the nation's waters), the Safe Drinking Water Act (SDWA) (to protect drinking water supplies), and state programs, or meeting local community needs. The purpose of this chapter is to provide an overview of watershed management for the protection of drinking water supplies—that is, source water protection—as a basis for review and critique of the watershed strategies included within the Memorandum of Agreement (MOA).
Source water protection programs are essentially watershed management programs with the specific goal of protecting drinking water supplies. In recent years, it has been more widely recognized that source water protection is a critical first step in the multiple-barrier approach to providing safe drinking water advocated by the Environmental Protection Agency (EPA) and the American Water Works Association (AWWA, 1997; EPA, 1997a). Examples of barriers include selecting the highest-quality source water, practicing watershed management, using the best available treatment technologies, maintaining a clean distribution system, practicing thorough monitoring and accurate data analyses, having well-trained operators, and maintaining operating equipment. Although the term "multiple barriers" is a recent one, the concept is old, as evidenced by the following quotation:
"This all tends to bring into prominence the necessity of using barriers of protection to the best possible advantage. The first of these is the reservoir system through which all raw waters should flow, even if the source is of sufficiently good quality to be by-passed to the works direct. The effect of promoting a continuous flow throughout storage systems in which water is sedimented results in an enormous reduction of bacterial and organic components. Next, an efficient filter barrier is interposed and, lastly, a chemical process which should be afforded time for effective action, either in contact tanks or in the mains before the draw-off. Although, on the one hand, one counsels the use of chemicals as a finishing process, and as a supplementary final barrier, it is necessary with such effective agents as the chloramines to voice a warning as regards their position in the scheme of water purification. The aim of application should be as a corrective to counter an emergency rather than the be-all and end-all of water purification, otherwise there is a definite risk of allowing our filtration, sedimentation, and storage barriers to fall to a less important position and to depend upon chloramine entirely." (Harold, 1934)
As perceptive as this 65-year-old quotation is, it omits a critical part of the equation—the active management of reservoir watersheds such that reservoir water quality is enhanced (watershed management). Today, it is well understood that water utilities should both control contaminant discharges into watersheds and prevent any unwise land use in watersheds that would adversely influence downstream water quality.
Source water protection programs include actions, policies, and practices to protect and enhance sources of drinking water (AWWA, 1997; EPA, 1997b). Figure 4-1 depicts the components of a watershed management framework for source water protection advocated by the committee. Several components or steps are necessary: (1) establishment of goals and objectives, (2) an inventory of the watershed and assessment of possible contaminant sources, (3) development and (4) implementation of protection strategies, and (5) monitoring and evaluation of program effectiveness. Stakeholder involvement in each of these components is key to ensuring that watershed management programs are effective. The remainder of this report analyzes sections of the New York City watershed management strategy that correspond to these components. For example, Chapter 6 focuses on the New York City Department of Environmental Protection's (NYC DEP) enhanced monitoring system and data analysis (effectiveness monitoring and evaluation). It also evaluates several important tools used in site and contaminant assessment. Chapters 7, 9, 10, and 11 discuss different protection strategies and their implementation, from structural best management practices (BMPs) used for stormwater management to nonstructural activities such as land acquisition.
Watershed management can be implemented at varying scales, from small subbasins (e.g., first-order streams or small lakes) to very large and complex systems (e.g., higher-order streams, large lakes, or estuaries). Previous experi-
ence has found that watershed management is most effective when it is conducted at the scale of medium-sized tributary and subtributary watersheds rather than at larger or smaller scales (NRC, 1999).
The watershed management process is dynamic and iterative, and the steps in Figure 4-1 are repeated as conditions within the watershed evolve and new information becomes available. It should be kept in mind that this chapter describes an ideal water management strategy that might be used for source water protection. It is rare to find a community in which all these components are successfully carried out, although many of the elements can be found in the watershed management program of the Tennessee Valley Authority (Box 4-1).
The Tennessee Valley Authority (TVA), an independent federal agency established in 1933, is responsible for flood control, energy production, economic development, and natural resource conservation in the Tennessee Valley. TVA serves a geographic area of 80,000 sq mi that includes 125 counties in seven states. The Tennessee River is 652 miles long and is the fifth largest river system in the United States. The necessary components of watershed management for source water protection, as described in Figure 4-1, are aptly demonstrated by several activities of the TVA.
Goal and Objective Setting: The TVA has recently placed a greater emphasis on water quality and has adopted a goal of making the Tennessee river system the cleanest and most productive commercial river system in the United States by the year 2000. To this end, it adopted the Clean Water Initiative in 1993 and, using a watershed approach, it has divided the region into seven subbasins, each with its own self-directing interdisciplinary River Action Team.
Watershed Inventory and Contaminant Assessment: To pinpoint water-quality problems, the River Action Teams are using rapid biological assessment methods and conventional physical and chemical methods in conjunction with aerial photographs and Geographic Information System (GIS) maps. Data and input from other agencies and individuals are being used to characterize the region and improve the health of the watershed.
Protection Strategies: A variety of structural and nonstructural protection strategies are being used to improve water quality in the region, from creation of wetlands for removal of phosphorus in wastewater to installation of pump-out stations on lakes. The strategies that are used differ, depending on the type of pollutant present, the value of the water resource (evaluated by six indicators: human health, ecological integrity of the resource, human use of the resource, ecological integrity of the downstream resource unit, downstream human use, and economic sustainability), and the likelihood of success (as determined by assessing local political, economic, and regulatory realities).
Implementation: The River Action Teams ensure implementation by partnering with federal and state agencies and individual landowners affected by the choice of protection strategies. Successful projects are
used by the River Action Teams as examples to demonstrate the benefits of the watershed approach.
Effectiveness Monitoring and Evaluation: There is an extensive biological, chemical, and physical monitoring system to track water quality. Results are fed to the River Action Teams, which use the information to further develop and refine protection strategies.
Stakeholder Involvement: The TVA's holistic, multiobjective watershed approach to the Tennessee River and its tributaries relies on the input of multiple stakeholders. Projects undertaken to improve water quality are joint actions of individuals, businesses, local and state agencies, and other federal agencies, along with TVA. Public education and open disclosure of the environmental problems are also part of the solution. For example, RiverPulse, a colorful, easy-to-read document, reports annually on the conditions for swimming and fishing and on overall ecological health of the waters in the Tennessee Valley, and it discusses the progress of some of the projects under way in the watershed.
GOAL AND OBJECTIVE SETTING
Goals provide general direction for source water protection programs by broadly stating the intent of the management plan. A primary goal of all municipal water suppliers is to provide an adequate supply of high-quality water, as defined by its taste, odor, color, clarity, and concentration of contaminants. Beyond this general goal, specific goals can be tailored to a watershed's physical characteristics, existing water quality concerns, contaminant sources, and regulatory constraints (AWWARF, 1991). Detailed numeric objectives often complement general goals by providing quantifiable and measurable direction for source water protection programs.
Source Water Protection Goals
Source water protection goals generally reflect the specific needs and conditions of the watershed and the entity implementing the program. Beyond providing high-quality source water, such programs may strive to (1) reduce or limit sources of contamination, (2) minimize the risk of hazardous chemicals entering
the water supply, (3) mitigate the effects of natural disasters, (4) provide flexibility within water system operations, (5) minimize treatment costs, and (6) comply with regulatory requirements. Realistic goals must recognize the need for balance and compromise among competing and often conflicting demands for various uses within the watershed (AWWARF, 1991), such as protection of aquatic life, recreation, water supply, agriculture, forestry, and urban development. This need for compromise is particularly pressing in water supply watersheds that are not substantially or wholly owned and managed by the water supplier. In addition, the limitations of regulatory authorities that implement and enforce program goals must be acknowledged. Goals must have the necessary supporting legal and regulatory authority to ensure effective implementation.
One example of a goal is ''to protect the water quality and supply reliability by seeking to balance the watershed uses such as the rights of private property owners and public recreational activities with the protection and management of natural resources" (Santa Clara Valley Water District, 1995). The Santa Clara Valley Water District goal recognizes the need for public support and for cooperation from other stakeholders to ensure a successful source water protection program. The Upper South Platte Watershed Protection Association (1998) in Colorado has established a similar goal that reflects a desire to balance activities and effectively engage stakeholders. The Salt Lake City Watershed Management Plan is more strongly focused on maintenance of water quality over other factors: "[Watershed] management emphasis prioritizes water quality first and multiple use of the watershed second. The Wasatch Canyons are protected to maintain a healthy ecological balance with stable environmental conditions, healthy streams and riparian areas, and minimal sources of pollution" (Salt Lake City Department of Public Utilities, 1998).
Specific Water Quality Objectives
In addition to basic goals, water agencies may also establish specific numeric or narrative objectives for their drinking water sources. As required under the CWA, most states or EPA regions have established use designations and associated water quality criteria for the protection of drinking water supplies. To date, water quality criteria to protect drinking water supplies have been limited and somewhat inconsistent from one state to another (Table 3-7). Based on a recent survey of western states, existing criteria generally focus on nitrate, metals, and a few other inorganic constituents; some organic and radiological constituents; and fecal coliform bacteria (Paulson and Vlier, 1997). Other constituents of concern in drinking water supplies, such as organic carbon and specific pathogens, have not yet been widely addressed by water quality criteria.
Maximum contaminant levels (MCLs) have been specified by the SDWA for application to treated water supplies and are not directly applicable as water quality objectives for source waters prior to treatment. Natural waters, even
pristine supplies, may not be able to meet MCLs for many constituents. MCLs can, however, serve as a basis for the derivation of water quality objectives for source water supplies. This is done by accounting for the effects of transport, storage, and treatment, a challenging task because treatment levels can vary and water quality transformations are not always fully understood. Hypothetical numeric objectives that might be set for source water are given in Table 4-1.
WATERSHED INVENTORY AND CONTAMINANT ASSESSMENT
An inventory of the watershed and an assessment of potential contaminants help to define the boundary conditions and form the basis for source water protection programs. A watershed inventory begins with a delineation of the source water protection area and an evaluation of existing water quality conditions. A complete inventory includes information on natural characteristics, land uses, activities, and other factors that can affect water quality. Land uses and ownership are particularly important because they help determine the need for regulatory authority to support the source water protection plan. The potential suscep-
TABLE 4-1 Example Numeric Objectives for a Surface Water Supply Reservoir
tibility of water supplies to natural and human sources of contamination within the source protection area is then evaluated in a contaminant assessment.
Source Water Protection Area Delineation
For surface water systems, the source water protection area generally includes the watershed area upstream of a water supplier's intake. It is delineated by the boundaries of drainage basins that supply streams, lakes, and reservoirs that serve as source water. Basins can also be divided into smaller subbasins that drain to tributary systems. In areas with transbasin diversions, the entire source water protection area may include watersheds that are geographically far removed from the point of use.
For groundwater systems, the source water protection area, also known as the wellhead protection area, is defined as the zone of recharge around a well. The wellhead protection area can be delineated using one of several methods, including the following: an arbitrary radius around a well (e.g., 2–3 miles), a calculated fixed radius that is determined as a function of hydraulic gradients, analytical modeling, or hydrogeologic mapping (Colorado State Department of Public Health and Environment, 1998; EPA, 1989). In delineating the source water protection area, zones and pathways through which contaminants could migrate and reach surface or groundwater systems must be considered.
Existing Water Quality Conditions
Existing water quality conditions can be assessed by comparing all available data on physical, chemical, and biological parameters to water quality objectives for source protection. In some states, the water quality objectives to which data should be compared are encompassed by a water's use classification and water quality criteria. Published assessments of water quality conditions (e.g., CWA Section 305(b) reports and Section 303(d) listings) can serve as readily available summaries of water quality conditions relative to use classifications and water quality criteria. For source water protection, some constituents are of particular concern because of their potential impact on water supply treatment and finished water quality. Some of these constituents are listed in Table 4-1.
In addition to specific constituents of concern, other chemical, physical, and biological measures can also provide strong indications of watershed ecosystem "health" (Meyer, 1997). For example, the relative condition of biological communities, particularly benthic macroinvertebrate communities, can serve as a screening tool to identify potential water quality problems that could be overlooked in chemical-specific monitoring. It should be noted that macroinvertebrate monitoring is most useful for meeting aquatic health goals and is somewhat limited as an indicator of water quality for drinking water sources.
In a preliminary assessment, water quality data can be compared with source
water quality objectives to determine where objectives are being met and to identify areas for improvement.
Natural watershed characteristics define conditions in the absence of human impacts. These can serve as a baseline against which changes in water quality associated with human development are monitored. An inventory of natural conditions within a watershed includes information on hydrology, topography, soils, vegetative cover, erosion potential, wetland and riparian areas, wildlife, and disturbance potential. Hydrologic information is needed for all surface waters (e.g., major rivers, tributaries, lakes, and reservoirs) and groundwater (water level contours, deep aquifers) in order to delineate basin and subbasin boundaries. Erosion potential can be determined by considering data on slope, soils, and land cover (Universal Soil Loss Equation; see Wischmeier and Smith, 1978). Disturbance potential refers to forest fires, hurricanes, and insect and disease outbreaks, among other things.
Land Cover and Uses
Source water quality is directly and profoundly influenced by land cover and land uses, both natural and human. For this reason, quantifying land cover and land uses is a necessary step in watershed management. Information on land ownership, land jurisdictions, and water rights is also needed to help determine the potential to manage land uses and mitigate impacts.
Important categories of land cover to measure include forests, open spaces, bodies of water, agricultural cover (e.g., pastures, row crops), and impervious surfaces. Activities within watersheds can affect water quality by producing contaminants from discrete point sources or from diffuse nonpoint sources. Those deserving special consideration in a watershed inventory include industrial and municipal wastewater treatment plants, individual sewage disposal systems, permitted stormwater discharges, agricultural activities, forestry, mineral extraction, and the generation, storage, and disposal of hazardous materials. Recreational activities, both within watersheds and on bodies of water, can also affect water quality.
Existing and potential contaminant sources identified during the watershed inventory should be assessed for their future impact on source water quality. Factors such as location of the contaminant source, relative toxicity and mobility of the contaminant, and size and quantity of the contaminant source must be considered. In addition, existing levels of the contaminant measured in the
watershed and their relationship to water quality goals can help determine the relative significance of various contaminant sources. Although work in this area is limited to date, both qualitative and quantitative contaminant assessment can help prioritize water quality problems and better direct limited resources.
Geographic Information Systems (GIS) and water quality models can be applied independently or jointly to support source water protection efforts. In combination, these tools are particularly powerful, with GIS as a tool to streamline data input and facilitate effective presentation of modeling results. As technology develops, specific linkages for these tools will continue to evolve.
Geographic Information Systems
Watershed inventories require the integration of many types of data and spatial location of that data to effectively evaluate potential impacts on source water supplies. This is best facilitated by a GIS, a computerized system for the storage, display, and manipulation of geographic data. It has been widely applied in source water protection and watershed planning initiatives across the nation (EPA, 1996).
A GIS can be built to house multiple "layers" of information to support a watershed inventory. Data layers, such as topography, water features, roads, and vegetation type, can be stored in two forms. Linear features such as roads or streams are often stored in vector form as a series of compass directions or azimuths1 and distances tracing the path of the landscape feature. Other attributes such as topography, soil type, land use, or vegetation type may be represented with vectors enclosing a polygon(s) or as matrices of position and attribute data in primary layers. Secondary layers are formed with attribute data (as real numbers) pertaining to a primary layer. For example, a primary layer of data on soil type can be used to generate secondary layers of data on soil thickness, infiltration capacity, and permeability. GIS layers also may be comprised of data on such points as buildings, wells, septic systems, or other features of limited or discrete size.
One of the powerful features of GIS is that point, line, and area data can be combined to map and model interrelationships, calculate areas, or create new layers. For example, a slope layer can be created by calculating the change in elevation between adjacent grid cells. The slope layer can then be used to estimate flow paths. These derivative layers are invaluable for watershed modeling and management. It should be noted that the quality and value of derived layers is dependent on the available grid size used in the GIS.
Information stored in a GIS can be linked to water quality databases and can be used to display water quality conditions at various locations, as part of the inventory. GIS can also be applied, in conjunction with water quality modeling, to help trace a problem identified downstream back to its source in the watershed. For example, GIS land cover data are used in New York City in conjunction with a terrestrial runoff model to determine the relative importance of phosphorus sources. In assessing the susceptibility of source water supplies to contaminants, a GIS can be used to locate areas that have the greatest potential for adverse impacts, quantify relative impacts for various subbasins, and determine where to concentrate management efforts.
Water quality models can be an important tool to support source water assessment and protection initiatives, particularly in areas where data are limited, or to project future conditions. Water quality modeling objectives can range from predicting impacts of future development scenarios, to identifying key impact areas, to looking at the effectiveness of various control strategies. A wide array of water quality models has been developed (EPA, 1992) that can be roughly divided into two general categories: (1) loading models, which address pollutant loads from various point and nonpoint sources within a watershed (e.g., P8, Watershed Management Model, AGNPS), and (2) receiving water models, which address the impacts of various loading conditions on water quality in rivers (e.g., QUAL2E), lakes (e.g., eutrophication models), and groundwater (e.g., Modflow). Recent research and development efforts have emphasized the linkage of GIS and parametric methods to build spatially distributed models.
DEVELOPMENT OF PROTECTION STRATEGIES
Comprehensive source water protection includes multiple structural and nonstructural protection strategies to remove or reduce contaminant sources of surface and groundwater supplies. Effective protection strategies focus on high-priority watershed activities and contaminants and target the ones with the greatest potential to affect water sources. The challenge is to select the appropriate combination of practices to prevent or treat point and nonpoint pollution sources in the watershed. The potential structural and nonstructural practices that are often considered in watershed management are listed in Table 4-2.
Structural practices are defined as those that treat or reduce pollution discharges from an existing source. Examples include upgrades to wastewater treatment plants (WWTPs) to reduce point source pollution and BMPs to reduce
TABLE 4-2 Structural and Nonstructural Practices to Be Considered in a Watershed Management Strategy
nonpoint source pollution. BMPs can mitigate nonpoint source impacts from agriculture, forestry, mining, construction, and urban runoff. Nonpoint source practices for agriculture generally address erosion control and sediment transport in disturbed areas, including the use of buffer zones, and management of significant activities such as grazing and irrigation, with special emphasis on controlling nutrients, pesticides, and herbicides (EPA, 1993). Forestry BMPs address timber harvest, road construction and maintenance, revegetation of disturbed areas, management of chemicals and fire hazard, and maintenance of streamside buffer zones (EPA, 1993). Urban stormwater BMPs include several conventional practices that have been widely applied: vegetated swales and filter strips, stormwater infiltration and percolation, extended detention (dry) basins, retention ponds (wet), and constructed wetlands (WEF and ASCE, 1998). More detail on the design and performance of BMPs can be found in the literature (Brown and Schueler, 1997; EPA, 1993; Schueler et al., 1992; WEF and ASCE, 1998) and in Chapters 8-11.
Nonstructural controls prevent or reduce potential pollution discharges from future sources. Many of the nonstructural controls involve institutional or policy tools used during the local land use planning process that governs how and where new development can occur. Public education can also be an important
nonstructural tool to change specific behaviors that create pollution discharges (e.g., lawn care, pet waste cleanup, and pesticide application).
Nonstructural practices (also known as source control measures) are often the most effective tools in addressing potential water quality problems, both in terms of cost and reliability. Depending on the ownership of, jurisdiction over, and authority for the land within the watershed, nonstructural strategies may include land use controls, land acquisition, and limits on activities and on the presence of hazardous materials. In addition, special management practices can be applied to limit the impacts of agriculture, forestry, and mining activities within key watershed areas. Some of the nonstructural practices found in the New York City MOA are evaluated in Chapter 7.
Several steps are necessary to ensure that the combinations of structural and nonstructural practices selected to protect the watershed are effectively implemented on the ground and maintained over the long term. The seven most frequently used elements to support implementation, often administered at the local or regional level, include performance criteria, review and inspection, maintenance, compliance enforcement, education and training, funding and incentives, and tracking.
1. Performance Criteria. Specific minimum engineering criteria, standards, or specifications are needed for each practice to ensure it is properly applied and can achieve its target performance for pollutant treatment or prevention. Using on-site sewage treatment and disposal systems (OSTDS) as an example, this may involve specific soil tests to determine whether subsurface disposal is feasible at a site, tank and field sizing criteria, setback distances, reserve fields, and technology requirements. Performance criteria are typically included in local, regional, or state manuals.
2. Review and Inspection. Ensuring that the watershed protection practices are designed properly and installed correctly occurs in two stages. In the first stage, development plans are scrutinized in a local review process to ensure that they meet applicable performance criteria and are appropriate for the site. In the second stage, field inspections are conducted at the site to ensure that the practices have been installed or constructed according to the plan. In the case of OSTDS, a developer might submit a plan for on-site wastewater disposal for local review, followed by an inspection at the site during construction to confirm that the system is installed properly and works as designed.
3. Maintenance. Structural and nonstructural practices must be continuously maintained in order to ensure effectiveness over time. Maintenance tasks for structural practices include routine pollutant cleanouts and rehabilitation of the practices as they approach the end of their design life. With OSTDS, routine maintenance entails periodic cleanouts of the septic tank and repairs to the distri-
bution field. OSTDS rehabilitation occurs as the system approaches the end of its design life and may include switching to a reserve field and/or tank replacement. For nonstructural practices such as a conservation easement, maintenance can involve routine inspections and vegetative management measures. Each watershed practice requires a legally binding maintenance agreement that identifies the party responsible for future maintenance and clearly outlines maintenance tasks, costs, and schedules.
4. Compliance Enforcement. An enforcement authority is usually needed to ensure that all owners or developers are in compliance with watershed regulations or criteria. Enforcement may be needed in the event an owner avoids watershed requirements, fails to properly implement plans, or does not adequately maintain a practice. Periodic site inspections are needed to detect watershed violations, and a range of enforcement tools can be used to induce compliance (correction notices, fines, stop-work orders, and even criminal penalties). In the case of OSTDS, compliance issues could include failure to install the system, improper installation or location, or failure to perform routine cleanouts or periodic rehabilitation.
5. Training and Education Outreach. Effective implementation of structural practices often requires a community investment in intensive training on how to design and apply watershed practices, since some consultants, engineers, and planners may be unfamiliar with current design practice. For example, contractors and engineers may require training on new OSTDS testing or technology requirements, and outreach may also be needed for OSTDS owners to improve operation and maintenance. Education and technical assistance are also critical in the implementation of nonstructural practices that involve private stewardship, pollution prevention, and land use management.
6. Funding and Incentives. The costs of applying and maintaining watershed practices can be significant for landowners or developers. Consequently, to achieve more widespread implementation, it may be desirable to provide innovative financing systems and/or economic incentives. This is particularly important for nonstructural protection strategies, such as conservation easements and other stewardship of private lands. Innovative financing systems such as stormwater or septic system "utilities" are often useful in providing long-term funding for the maintenance of structural watershed protection tools.
7. Watershed Tracking. Over time, implementation results in the application of practices at hundreds of development sites, farms, properties, and other locations within a watershed. In many cases, dozens of local, state, and federal agencies, utilities, land trusts, and watershed organizations are involved in various aspects of implementation. Consequently, it is important to track the location, management status, and maintenance record of all practices to gauge the cumulative progress toward implementation in a coordinated fashion, namely with a multipurpose database that can be linked to a GIS.
EFFECTIVENESS MONITORING AND EVALUATION
Water quality monitoring is integral to measuring the success of any source water protection program. Monitoring data can be used to refine programs as necessary, by assessing the effectiveness of specific controls and identifying needed adjustments. In addition, monitoring is essential to earlier stages of watershed management, such as the watershed inventory and contaminant assessment steps. This section broadly discusses the multiple purposes of monitoring for source water protection, including water quality monitoring, monitoring of health outcomes among the consumers of drinking water, and monitoring of social and economic parameters to determine the success of watershed management. Subsequent data analysis and program evaluation are highlighted because these are frequently the weakest steps in a watershed management strategy.
Physical, Chemical, and Biological Monitoring of Water Supplies
Monitoring for source water protection is implemented to address four general objectives: compliance with environmental regulations, systems operations, performance of BMPs, and modeling activities. Compliance monitoring evaluates specific physical, chemical, and biological parameters for compliance with local, state, and federal regulations, including the SDWA. The methodologies used for monitoring most parameters are rigorously specified and regulated. An example of compliance monitoring might be daily sampling of turbidity, pathogens, and disinfection byproducts at the water intake. Operational monitoring is conducted on a broad set of parameters needed to effectively assess the ongoing and successional quality of water and reservoir dynamics and to determine the sources of pollution that influence water quality. This type of sampling is typically conducted at a broad spatial scale (e.g., in the reservoirs and upstream tributaries), and it is directed at specific water quality issues such as eutrophication. Performance monitoring is needed to evaluate the effectiveness of watershed management practices and policies and to isolate design factors that influence the variability of pollutant removal. Performance monitoring often involves intensive sampling of flow and pollutant mass as it passes through a particular management practice (such as a stormwater pond, OSTDS, or riparian buffer). Finally, monitoring data can be used to support modeling of projected changes in water quality under different conditions that are due to land use change or watershed management actions. Modeling-support monitoring involves both intensive and extensive sampling of a reservoir and its watershed to define parameter values, set initial conditions, and physically characterize the watershed. These four types of monitoring are not mutually exclusive, and often data collected for one activity can be used in other contexts. Physical, chemical, and biological parameters to target in monitoring for source water protection are described below.
Pathogenic Microorganisms of Direct Potential Risk to Human Health
Viruses, bacteria, and protozoans originating from animal and human sources are often found in drinking water supplies prior to treatment, albeit at low concentrations. As noted in Chapter 3, the protozoans Cryptosporidium and Giardia are especially troublesome because of their resistance to disinfection with chlorine. Pathogen monitoring programs must be capable of measuring dilute concentrations of these organisms, determining their viability, and assessing the impact of treatment processes on their survivability in drinking water.
As described in Chapter 3, eutrophication refers to an increase in the rate of organic matter production in surface waters as a result of nutrient (e.g., phosphorus) addition and subsequent microbial growth. This increased organic matter can cause reductions in dissolved oxygen, alter taste and create odors in drinking water, and it can cause destruction of fish and aquatic plant habitat. Comprehensive monitoring that encompasses a wide variety of eutrophication parameters (including phosphorus, chlorophyll a, organic carbon compounds, and dissolved oxygen among others) is critical to evaluating the success of source water protection programs.
Natural organic compounds in a drinking water supply can react with disinfection agents (e.g., chlorine) and form potentially carcinogenic chemical compounds (disinfection byproducts or DBPs) in water distribution systems. An effective monitoring program must be able to determine the sources and quantities of dissolved organic compounds and their potential for forming DBPs) while taking into account seasonal variations and the operational flexibility of the water supply system.
High turbidity levels are indicative of sediment transport generated by heavy or abnormal precipitation events in the drainage basins of source waters. Erosion either from land runoff or stream banks and the associated sediment transport can be detrimental to water supply systems and aquatic habitats for many reasons. Turbidity has been shown to interfere with the disinfection process (Symons and Hoff, 1976). In addition, sediments can introduce particulate-phase pollutants into water supply reservoirs. Assessing and understanding these effects are important goals of an effective monitoring program.
The presence of toxic chemicals in a drinking water supply is a more site-specific problem than the four pollutant categories discussed above. Those chemicals known to have adverse impacts on drinking water quality include metals and metalloids, synthetic organic chemicals, volatile organic chemicals, and pesticides. Persistent bioaccumulating toxic chemicals, such as polychlorinated biphenyls (PCBs) and mercury, are particularly troublesome and are a potential concern for the long-term safety of water supplies. The concentrations of PCBs can be substantial in the Great Lakes area as well as some reaches of the Hudson and Housatonic rivers. Monitoring of toxic chemicals should take into account these site-specific considerations.
Monitoring of Public Health
In addition to monitoring water quality, a comprehensive watershed management strategy includes monitoring of public health to confirm that no waterborne disease outbreaks or unacceptable levels of endemic illness are associated with the water supply. In some communities, this type of monitoring is mandatory for compliance purposes. There are several approaches for disease surveillance that can be implemented.
Public health surveillance activities typically consist of reporting cases of specific infections to local and national public health agencies, such as county health departments or the Centers for Disease Control and Prevention (CDC). Information on disease rates can serve several purposes as part of a watershed management program. Its main role is to provide baseline data on disease trends over time in a target population. Baseline disease rates can then be compared with new information to determine whether specific disease rates are generally increasing or declining, to elucidate seasonal trends in specific disease rates, and to delineate high-risk populations or geographic areas.
Passive surveillance refers to the voluntary reporting of cases of ''notifiable" diseases to public health authorities by a health care provider or laboratory. The list of reportable diseases varies from state to state. Waterborne infections that are usually reported include salmonellosis, shigellosis, hepatitis A virus infection, typhoid fever, cholera, E. coli O 157:H7 (48 states, including New York), cryptosporidiosis (44 states, including New York), and giardiasis (44 states, including New York) (www.cste.org). Passive surveillance systems are often very insensitive—especially for mild conditions where an ill person may not seek medical care—and thus they detect only a small fraction of the true incidence of cases.
Active surveillance is a system where the source of case information (health care provider or laboratory) is called on a regular basis to determine if any cases of a specific condition have been observed. Although this process is considerably more sensitive than passive surveillance, the sensitivity of both active and passive disease surveillance is limited at multiple steps in the process.
Finally, enhanced surveillance systems are those where "special additional efforts are made to encourage disease reporting" (Frost et al., 1996). Examples include surveillance for gastrointestinal disease in sentinel nursing homes, monitoring absenteeism in schools or among hospital employees, monitoring Health Maintenance Organization nurse hotline calls about gastrointestinal illness, and monitoring sales of antidiarrheal medications. These activities require far more intensity of effort and resources than passive surveillance but can provide real-time monitoring of the health of the target population.
Depending on the time for development of symptoms and on when a case is diagnosed relative to exposure, some active or enhanced surveillance systems may be able to alert health authorities about immediate, sharp increases in disease rates and enable detection of disease outbreaks. However, in most instances, surveillance for disease is too slow to detect waterborne disease outbreaks.
Epidemiological Studies of Waterborne Disease
Surveillance systems collect information on illness rates in the community but cannot determine risk associated with drinking water because illness reported to state surveillance systems may be associated with contaminated food or other common transmission routes. Determining whether an observed pattern of disease is associated with drinking water requires epidemiologic studies specifically designed to link health outcomes to specific exposures.
Most epidemiologic investigations of drinking water and health have been conducted following an outbreak of waterborne disease. Outbreak investigations have provided valuable information on risk factors and etiologic agents associated with waterborne disease. At least 740 recognized waterborne outbreaks occurred in the United States between 1971 and 1994, and enteric protozoa were the most frequently identified cause of waterborne outbreaks (20 percent) and illnesses requiring hospitalization (78 percent) (Craun et al., 1998).
Epidemiologic studies are also designed to examine endemic (baseline level) waterborne disease or other health risks that may be associated with low levels of microbial or chemical contaminants in drinking water. In situations where baseline levels of waterborne disease may be low, the study population must be large enough to detect any difference between the disease patterns in an "exposed" and an "unexposed" population. Although most epidemiologic study designs are too expensive, time-consuming, and long-term to be conducted on a regular basis, a well designed and conducted study can provide the ultimate test of the safety of a water supply.
Several epidemiologic study designs have been applied to the study of endemic waterborne disease (Box 4-2). Ecologic studies examine patterns of illness or infection (collected from surveillance systems) and concurrent data on water quality (such as turbidity) to determine if any correlation can be observed over space and time. These descriptive studies are relatively inexpensive and easy to perform because they usually take advantage of existing data on health outcomes and water quality. However, they are limited because they examine aggregate data from groups of populations and cannot take into account individual risk factors (such as contact with child daycare centers or overseas travel) or individual water exposure (such as use of bottled water or length of residence in the study community).
In case-control studies, the exposure histories of individuals with the disease of interest ("cases") are compared to the exposures of individuals without the disease ("controls"). Individual study subjects are queried about their residence history, water consumption habits, and risk factors for gastrointestinal disease. The analysis of these results allows the association between exposure and a single health outcome to be evaluated while controlling for individual risk factors. Case-control studies cannot prove that exposure caused the adverse health outcome because they do not provide evidence that the exposure preceded the disease. However, they are useful for examining risk factors for specific health outcomes and require fewer participants than cohort studies (described below).
Cohort studies also collect information on individual exposure, risk factors, and health outcomes. The illness rates in a group of people who are exposed to a water supply of interest are compared to the illness rates in a group of people exposed to a different water supply (such as water receiving a different type of treatment, bottled water, or water receiving additional in-home treatment). Because this design identifies the study population and measures exposure before the development of disease, it can be used to determine the temporal relationship between exposure and disease. These studies are typically the most expensive and time-consuming, especially if a long follow-up period is used.
For all of these epidemiologic approaches, accurately measuring actual exposure to microbial pathogens or chemical contaminants in drinking water and choosing appropriate health outcomes from the wide array of possible water-associated health effects is extremely challenging. Approaches and considerations for exposure assessment and outcome measurement are reviewed in detail elsewhere (NRC, 1998).
Monitoring of Social and Economic Parameters
Although it is not explicitly addressed in most programs, monitoring of social and economic factors can play an important role in measuring the success of a watershed management strategy. The chosen parameters must be tailored to the goals of the specific watershed management plan. Social metrics of interest
A variety of ecologic studies have been conducted to demonstrate a correlation between health outcomes and water quality parameters. Batik et al. (1980) attempted to find a relationship between endemic rates of hepatitis A infection reported in 75 counties and municipal source water quality and/or level of water treatment for all water supplies in the country. However, no statistically significant associations were observed. A longitudinal study of French alpine villages that used untreated groundwater for their drinking water supplies observed a weak relationship between rates of acute gastrointestinal disease and the presence of fecal streptococci indicator bacteria in the public water system over a 15-month study period (Zmirou et al., 1987). Illness data were collected through active surveillance by physicians, pharmacists, and schoolteachers, while weekly water samples collected from frequently used taps in the distribution system of each village were analyzed for several bacterial indicator organisms. Schwartz et al. (1997) attempted to link turbidity in the Philadelphia water supply with hospital emergency room visits for gastrointestinal symptoms. However, a number of methodological problems with this investigation make the study results questionable.
The role of filtration in affecting disease rates has been investigated in two contrasting studies. The first study examined Cryptosporidium antibodies in 86 blood samples collected from children as part of the lead-testing program in Massachusetts (Griffiths, 1999). Samples from children living in towns with an unprotected, filtered surface water supply were more likely to have Cryptosporidium antibodies (74 percent) and higher antibody levels (mean optical density = 0.250) than samples from children living in towns with an unfiltered, protected surface water supply (40 percent seropositive and mean optical density = 0.138). The authors concluded that there is significantly more exposure to Cryptosporidium for children supplied by an unprotected, filtered surface water than for children served by an unfiltered, protected supply and that increased watershed protection and more stringent filtration methods are need to reduce waterborne exposure to Cryptosporidium.
The second study evaluated cryptosporidiosis among AIDS patients in Los Angeles County by comparing prevalence in two communities with different types of water treatment (Sorvillo et al., 1994). One community used standard flocculation and sand filtration; the other community had
only water clarification and chlorine disinfection (which was later modified, in December 1986, to include pre-ozonation, coagulation, flocculation, high-rate filtration through anthracite media, and chlorination). The water sources for both communities included surface waters from which oocysts had been recovered. From 1983 through 1986, AIDS patients in both communities had a similar prevalence of cryptosporidiosis (6.2 and 4.2 percent), although the rate in the community with filtered water was slightly higher. During the 4-year period after filtration was installed in the second community (1986–1990), cryptosporidiosis rates declined in both communities to 3.3 percent and 3.4 percent. The authors concluded that municipal drinking water was not an important risk factor for AIDS patients in Los Angeles County. However, the authors noted that the ecologic nature of the study did not allow examination of the quantity and sources of water consumed by individuals, and there was no information on the levels of contamination in the different catchment areas for the two communities.
Two case-control studies have examined the relation between water supply and endemic giardiasis (Chute et al., 1987; Dennis et al., 1993). In these studies, cases of giardiasis were identified from a clinic or from a state registry, and controls (individuals without giardiasis), matched for age and sex, were recruited from the same clinic, from acquaintances of the case, or by random digit dialing. Cases and controls were either interviewed by telephone or they filled out a mail survey about potential risk factors such as source of drinking water, child daycare utilization, animal contacts, foreign travel, camping, and swimming in a natural body of fresh water. Both studies found that giardiasis was significantly associated with the use of a shallow dug well or surface water as the house-hold water source.
Cohort studies in the form of randomized intervention trials were conducted in Canada to examine the risk of gastrointestinal illness associated with the consumption of conventionally treated municipal drinking
water that met current microbiological standards (Payment et al., 1991). The first study used 606 households, 299 of which were supplied with reverse-osmosis filters that provided additional in-home water treatment. Gastrointestinal symptoms were recorded in family health diaries. Water samples from the surface water source, treatment plant, distribution system, and study households were analyzed for several indicator bacteria and culturable viruses. Over a 15-month period, a 35 percent higher rate of gastrointestinal symptoms was observed in the 307 study households drinking municipal tap water without in-home treatment compared to the 299 study households supplied with reverse-osmosis filters. Symptomatology and serologic evidence suggested that much of this increased illness may be due to low levels of enteric viruses in the municipal water supply that originated from a river contaminated by human sewage.
Using a similar design, a second intervention study was conducted in the same community in Montreal with 1,400 families randomly allocated to four groups of 350. One group consumed conventionally treated municipal tap water that met current North American drinking water standards. The second group consumed tap water from a continuously purged tap. The third group consumed tap water that was bottled at the treatment plant, and the fourth group consumed purified bottled water (tap water treated by reverse osmosis or spring water). The health of the families was monitored for a 16-month period. The groups consuming tap water and continuously purged tap water experienced 14 percent and 19 percent more illness, respectively, than did the families consuming purified bottled water. Greater illness rates were observed in children 2–4 years of age. The authors concluded that 14 percent to 40 percent of the gastrointestinal illnesses reported were attributable to tap water meeting current water standards and that contamination in the water distribution system was partly responsible for these illnesses (Payment, 1997; Payment et al., 1997). Congress has recently mandated that the CDC and EPA provide a national estimate of waterborne disease occurrence by August 2001. To address this issue, these agencies will be conducting two epidemiologic studies (household intervention studies) similar to those conducted by Payment and others. One study will be conducted in a city receiving drinking water from a surface water source and the other in a city receiving drinking water from a groundwater source.
may include (1) population growth in the watershed, (2) awareness levels of watershed residents regarding watershed management, (3) compliance with inspection and maintenance schedules for BMPs or OSTDS, and (4) rates of residential pesticide application in the watershed. Information about such social parameters can be gained by conducting surveys of watershed residents on a regular basis. Economic parameters may include (1) employment rates and opportunities for watershed residents, (2) types of new development in the watershed (as tracked by building permits), and (3) acreage of land acquired during watershed management. Depending on the scope of the watershed management program, such factors as these are likely to change during program implementation and merit monitoring because of their importance to overall program success.
Evaluation of Monitoring Data and Information
Monitoring, by itself, has limited management value in source water protection efforts unless it is integrated within a larger framework of watershed evaluation. In many watersheds around the country, significant resources are being expended on comprehensive monitoring programs. Often, however, data are not being thoroughly evaluated to draw scientifically based conclusions about the effectiveness of watershed management plans. In other words, current implementation of watershed management plans may be data-rich and information-poor.
The evaluation of monitoring data is needed for three of the four general categories of monitoring outlined above: compliance, operation, and performance. Modeling is itself a tool that can be applied to evaluate data for a variety of objectives. Effective monitoring programs include provisions to collect the data needed to fully support anticipated modeling efforts.
Determining compliance with federal, state, and local environmental regulations is the most frequent use of monitoring data. Generally, compliance evaluations consist of comparing water quality monitoring data and information to specific numeric objectives and determining the frequency and magnitude of any exceedances. Compliance can be measured for source water quality, finished water quality, and human health effects. It can also include an assessment of implementation rates (e.g., number of structural stormwater practices constructed in compliance with design specifications, or number of construction sites inspected and found to comply with erosion control ordinances) versus objectives for planned watershed practices.
Monitoring data can be evaluated to assess the condition of water sources and treated supply operations. Trend analysis—one means of evaluating improvements in or degradation of water quality over time—can be applied to reservoirs, upstream tributaries, and treated water. Water quality at points throughout a watershed can also be evaluated to identify the relative significance of various land uses or activities within the watershed and related impacts on water quality. Evaluating the relationships among water quality variables can also be useful in identifying cause-and-effect linkages. These kinds of evaluation can help define problem areas within the watershed and focus future management efforts.
Often, simple graphical presentations of data can provide visual overviews for the ongoing tracking of operations. Such graphical data displays are also useful in communicating to the public.
One of the more valuable, yet often overlooked, applications of water quality monitoring data is to evaluate performance. On a broad scale, the overall performance of a watershed plan can be measured by looking at key indicator variables in selected locations over time (see discussion below on risk analysis). On a smaller scale, data can be used to evaluate the effectiveness of specific structural or nonstructural practices. Information about the effectiveness of practices being implemented throughout the watershed is needed to determine the actual value of these practices in benefiting water quality. This information can then be used to support future decision-making. A particularly valuable activity is to evaluate the effectiveness of newer technologies or approaches on a small, pilot scale before widespread implementation.
Formal risk analysis provides an example of how monitoring data might be evaluated to assess the adequacy of a water supply from a public health point of view. Risk is defined as the possibility of suffering harm from a hazard, and risk analysis provides the tools by which the magnitude and likelihood of such consequences are evaluated. If available, pathogen or toxic chemical monitoring data can be used to estimate the risk of "infection" given a certain daily consumption of drinking water. Trends in the risk estimate over time can be used to evaluate the watershed management program over time. These and other types of evaluations of monitoring data are suggested throughout this report.
Successful watershed planning requires careful attention to the nature of public participation (NRC, 1999). Furthermore, watershed management must get
both the "right participation" and the "participation right" (Stern and Fineberg, 1996). That is, the relevant stakeholders must be represented throughout the planning process, and their viewpoints and concerns must be adequately understood in a timely fashion. A stakeholder is defined as an agency, group, organization, or person who has an interest in a process, has decision-making responsibility or authority over that process, or is affected by the outcome of that process.
Involvement of relevant stakeholders is complicated by the almost certain lack of correspondence between political jurisdictions and watershed boundaries. This raises the question of how a community of interest within a watershed context is defined. When a watershed covers a large area, geographically dispersed and socially diverse groups must be brought together to solve a common problem. In some cases, an institutional arrangement to facilitate such community formation may not be available.
The mix of stakeholders may differ depending on specific watershed problems, and the community of interest must be defined on a case-by-case basis. This is a daunting task, given the usually contingent and fortuitous nature of community formation. However, watershed plans can only be effectively implemented if such community definition and formation take place. There are several examples of major federally funded river system projects that have completed numerous scientific studies but have resulted in little change in the way land is managed because of the lack of stakeholder involvement. Particularly complex and context-specific issues like nonpoint source pollution, aquatic and terrestrial habitat preservation, and protection of wetlands are usually resolved only at the local level involving all the relevant stakeholders in planning and decision making (NRC, 1999).
Inclusion of the complete range of stakeholders in the watershed management process does not assure that all interests will be served. There are always tradeoffs and conflicts between competing economic, social, and environmental interests. Methods of conflict resolution can serve as useful tools in helping stakeholders to achieve an acceptable balance among tradeoffs. However, the successful resolution of conflicts requires that stakeholders develop a shared knowledge base regarding the watershed and that they work toward common goals. For this reason, expanding environmental awareness and improving stakeholder education are critical to the success of a watershed management program.
Education may broadly be thought of to include not only formal school curricula, but also media coverage and government dissemination of information. At a minimum, meaningful citizen participation is based on wide and open access to information (Popovic, 1993). Of particular importance are spatially distributed databases (such as GIS) and user-friendly models that can be used to understand both the present structure of the watershed and future configurations. Providing maps of watersheds and subbasins is a necessary step in disseminating information to stakeholders. Stakeholder education helps to create a more complete understanding of the consequences of local actions. For example, public
support for certain protection strategies requires an understanding of the "downstream" environmental consequences of local actions. This knowledge may force individuals to recognize public goods that conflict with their particular interests.
Because it requires recognition of a full range of key interests and tradeoffs, the building of trust, and identification of effective and equitable solutions, involving stakeholders in watershed management is an ongoing and lengthy process. There are a variety of tools to encourage stakeholder involvement, but none are completely effective. Typical methods include public hearings, citizen advisory committees and task forces, workshops, alternative dispute resolution, citizens' juries, citizens' panels, and public opinion surveys. The best approach given the current state of knowledge is to open various channels for involvement and to create and maintain an openness to a wide range of methods for integrating stakeholder inputs.
This chapter has defined the elements of a watershed management strategy for source water protection. As is apparent in Figure 4-1, watershed management is iterative or cyclical, designed to respond to natural, legal, social, political, and economic changes. Watershed management is a continuous process, with progress measured in discrete steps. Implementation of protection strategies is eventually undertaken, and monitoring is used to assess plan success. Watershed management must be adaptive and flexible in order to incorporate new information as it becomes available. Data inputs that allow assumptions and strategies to be refined keep the watershed management process going.
Ideally, a watershed management plan integrates complete knowledge of the watershed with known methods for water quality protection into a holistic strategy. In some watersheds, the main problems have been identified, and prevention and mitigation methods are available. In most other instances, however, information is limited. Watershed and water resource managers must use best professional judgment to implement protection strategies, monitor their effectiveness, and then make adjustments accordingly. The result of this approach is that watershed management, instead of being a single, comprehensive plan, is instead a series of smaller, individually oriented, plans. Although this expedites the implementation of management practices, it can also work at cross-purposes, since the solution to one problem can create new problems or exacerbate an already existing problem.
Experience has shown that although many communities have used watershed plans as a key tool to protect their local water resources, their success is highly variable (Schueler, 1995). (See Box 4-1 for a description of a successful, comprehensive watershed management strategy.) In general, local watershed plans are developed based on mapping, monitoring, and modeling studies and are adopted through a management process that involves key stakeholders (EPA,
1994). Ideally, the planning process results in the implementation of a future land use pattern and of appropriate management practices that are capable of meeting the water resource objectives developed for the watershed. In reality, however, many local watershed plans have not realized this promise (Schueler, 1995). The remainder of the report focuses on the watershed management strategy used for the New York City watershed, highlighting important successes of the program and making recommendations for improvement.
American Water Works Association Research Foundation (AWWARF). 1991. Effective Watershed Management for Surface Water Supplies. Arlington, VA: AWWARF.
American Water Works Association (AWWA). 1997. Source Water Protection Statement of Principles. AWWA MainStream.
Batik, O., G. F. Craun, R. W. Tuthill, and D.F.Kraemer. 1980. An epidemiologic study of the relationship between hepatitis A and water supply characteristics and treatment. Am. J. Public Health 70: 167–168.
Brown, W., and T. Schueler. 1997. National Pollutant Removal Performance Database for Stormwater BMPs. Center for Watershed Protection, prepared for Chesapeake Research Consortium, Inc.
Chute, C. G., R. P. Smith, and J. A. Baron. 1987. Risk factors for endemic giardiasis. Am. J. Public Health 77(5):585–587.
Clements, J. T., C. S. Creager, A. R. Beach, J. B. Butcher, M. D. Marcus, and T. R. Schueler. 1996. Framework for a Watershed Management Program, Project 93-IRM-4. Water Environment Research Foundation.
Colorado Department of Public Health and Environment. 1998. Colorado Source Water Assessment and Protection—Draft Program, Colorado Water Quality Control Division.
Craun, G. F., S. A. Hubbs, F. Frost, R. L. Calderon, and S. H. Via. 1998. Waterborne outbreaks of cryptosporidiosis. Journal of the American Waterworks Association 90(9):81–91.
Dennis, D. T., R. P. Smith, J. J. Welch, C. G. Chute, B. Anderson, J. L. Herndon, and C. F. vonReyn. 1993. Endemic giardiasis in New Hampshire: a case-control study of environmental risks. J. Infect. Dis. 167(6):1391–1395.
Environmental Protection Agency (EPA). 1989. Wellhead Protection Programs—Tools for Local Governments. EPA 440/6-89-002. Washington, DC: EPA.
EPA. 1992. Compendium of Watershed-Scale Models for TMDL Development. EPA 841-R-92-002, Office of Water. Washington, DC: EPA.
EPA. 1993. Guidance Specifying Management Measures for Sources of Nonpoint Pollution in Coastal Waters. EPA 840-B-92-002. Office of Water. Washington, DC: EPA.
EPA. 1994. The Watershed Protection Approach: Statewide Basin Management. Office of Water. WH-55. Washington, DC: EPA.
EPA. 1996. Better Assessment Science Integrating Point and Nonpoint Sources—BASINS Version 1.0 User's Manual . EPA-823-R-96-001. Washington, DC: EPA.
EPA. 1997a. Notice of Data Availability for the Enhanced Surface Water Treatment Rule. Washington, DC: Office of Water, EPA.
EPA. 1997b. State Source Water Assessment and Protection Programs Guidance—Draft Guidance. EPA 816-R-97-007. Office of Water. Washington, DC: EPA.
Frost, F. J., G. F. Craun, and R. L. Calderon. 1996. Waterborne disease surveillance. Journal of the American Waterworks Association 88(9)66–75.
Griffiths, J. K. 1999. Serological testing for cryptosporidiosis. Presentation at U.S. EPA Office of Ground Water and Drinking Water Stage 2 Microbial/Disinfection Byproducts Health Effects Workshop. February 10–12. Washington, DC.
Harold, C. H. H. 1934. Res. Chem. Bact. Biol. Exam. London Waters 29.
Meyer, J. K. 1997. Stream health: Incorporating the human dimension to advance stream ecology. Journal of the North American Benthological Society 16(2):439–447.
National Research Council (NRC). 1998. Issues in Potable Reuse: The Viability of Augmenting Drinking Water Supplies With Reclaimed Water. Washington, D.C.: National Academy Press.
NRC. 1999. New Strategies for America's Watersheds. Washington, DC: National Academy Press.
Paulson, C., and J. Vlier. 1997. Implementation of Source Protection Plans—Where is the Authority?, American Water Works Association Water Resources Conference—Water Resources Management: Preparing for the 21st Century. August 1997.
Payment P., L. Richardson, J. Siemiatycki, R. Dewar, M. Edwardes, and E. Franco. 1991. A randomized trial to evaluate the risk of gastrointestinal disease due to consumption of drinking water meeting current microbiological standards. Am. J. Public Health 81(6):703–708.
Payment P., J. Siemiatycki, L. Richardson, G. Renaud, E. Franco, and M. Prevost. 1997. A prospective epidemiological study of gastrointestinal health effects due to the consumption of drinking water. Int. J. Env. Health Res. 7:5–31.
Payment, P. 1997. Epidemiology of endemic gastrointestinal and respiratory diseases: incidence, fraction attributable to tap water and costs to society. Water Sci. Tech. 35:7–10
Popovic, N. A. F. 1993. The right to participate in decisions that affect the environment. Pace Environmental Law Review 19(2):683–709.
Salt Lake City Department of Public Utilities. 1998. Salt Lake City Watershed Management Plan. Final draft. Salt Lake City Department of Public Utilities.
Santa Clara Valley Water District. 1995. Comprehensive Reservoir Watershed Management—Phase 1 Project Report.
Schueler, T. 1995. Crafting better urban watershed protection plans. Watershed Protection Techniques 2(2):329–337.
Schueler, T. R., P. A. Kumble, and M. A. Heraty. 1992. A Current Assessment of Urban Best Management Practices–Techniques for Reducing Non-point Source Pollution in the Coastal Zone. Anacostia Restoration Team, prepared for U.S. EPA Office of Wetlands, Oceans, and Watersheds.
Schwartz, J., R. Levin, and K. Hodge. 1997. Drinking water turbidity and pediatric hospital use for gastrointestinal illness in Philadelphia. Epidemiology 8(6):615–620.
Sorvillo, F., L. E. Lieb, B. Nahlen, J. Miller, L. Mascola and L. R. Ash. 1994. Municipal drinking water and cryptosporidiosis among persons with AIDS in Los Angeles County. Epidemiol. Infect. 113:313–320.
Symons, J. M., and J. C. Hoff. 1976. Rationale for turbidity maximum contaminant level. Paper 2A-4a In American Water Works Association Technology Conference Proceedings, Water Quality Technology Conference, December 8–9, 1975, Atlanta, GA, American Water Works Association, Denver, CO.
Upper South Platte Watershed Protection Association. 1998. Memorandum of Understanding.
Water Environment Federation and American Society of Civil Engineers (WEF and ASCE). 1998. Urban Runoff Quality Management. WEF Manual of Practice No. 23 and ASCE Manual and Report on Engineering Practice No. 87.
Wischmeier, W. H., and D. D. Smith. 1978. Predicting rainfall-erosion losses. USDA Agriculture Handbook No.537. Washington, DC: USDA.
Zmirou, D., J. P. Ferley, J. F. Collin, M. Charrel and J. Berlin. 1987. A follow-up study of gastrointestinal diseases related to bacteriologically substandard drinking water. Am. J. Public Health 77(5):18–20.