Since its beginning, the U.S. Geological Survey (USGS) has been one of the primary federal agencies responsible for assessing the quantity and quality of the nation’s surface water and groundwater. In the early 1980s USGS performed and published an assessment of the nation’s water, titled The National Water Summary 1983—Hydrologic Events and Issues (USGS, 1984). After the completion of this document and related congressional testimony in the mid-1980s, USGS scientists concluded that their ability to say something meaningful about the quality of the nation’s waters was limited. Indeed, the USGS resources to assess national water quality were the National Stream Quality Accounting Network1 (NASQAN) and the Hydrologic Benchmark Network,2 which, while nationwide, were sparse and were conducting routine monitoring rather than data analysis. Furthermore, NASQAN and the Hydrologic Benchmark Network reflected water-quality sampling approaches from the early 1970s and 1960s, respectively, and thus did not provide data appropriate to address national water-quality questions of the mid-1980s.
Stimulated by the aforementioned events, the USGS contemplated and envisioned a national water-quality assessment program. Key pieces of this original vision included sampling hydrologeologically meaningful units of study or study units, using multiple scales of investigation to achieve a national picture by piecing together information from the study units, integrated teams of scientists performing the water-quality assessment, a
punctuated rotational sampling design, and assessment using established methods (Box 2-1).
Shortly after the NRC’s Water Science and Technology Board (WSTB) endorsed the original concept of the National Water-Quality Assessment (NAWQA) program (Chapter 1), it convened a colloquium in 1986 to articulate the necessary elements for a national water-quality assessment program (NRC, 1987). Colloquium participants endorsed the program concept and also raised new issues for consideration such as whether and how to interface with state regulators, which contaminants would be selected for monitoring, and the need to explore surface water and groundwater interactions. For example, the original study unit concept consisted of 123 separate surface water and groundwater units: 69 surface water-dominated and
The Original Vision for the NAWQA Program
The USGS vision for NAWQA included selecting study units, or hydrologically meaningful pieces of geography (Winter, 2001), in which to monitor water quality. The study units were building blocks for multiple scales of water quality investigation; they served not only as the base level but also as tools for “scaling up” to the bigger, national picture. Consistency between study units would allow the program to make comparable statements about the nation’s water quality.
Data collection and data analysis for the water quality assessment in each study unit were to be done by a team working together in an integrated group. This team of scientists was to make measurements, understand what these measurements meant, and make a statement about water quality in a given study unit. It was thought that sampling and assessment should follow a punctuated, rotational system of study with intense data collection for approximately 3 years followed by a period of analysis and publication, a time of minimal monitoring, and a return to the area to repeat the cycle.
NAWQA was envisioned to be a network for data collection defined by geology, hydrology, and land use, rather than a grid or a random sampling strategy. In this way, NAWQA could capture snapshots of both the entire system and “indicator” sites. The design had a strong prejudice toward collecting data in places where USGS had high-quality streamflow data records, in the belief that surface water-quality data are meaningless without considering flow and long-term history. Finally, use of known tools and understanding of processes to monitor the nation’s water quality were critical components of the original vision. NAWQA would not deploy untested methods and approaches for analyzing water quality unless on a limited scale. Rather, research and development of methods in other USGS programs would feed the program’s activities and assist the program in achieving the goal of assessing the nation’s water quality.
SOURCE: R. M. Hirsch, personal communication, May 13, 2009.
54 groundwater dominated (NRC, 1990). However, as the pilot program progressed, it became apparent to both the National Research Council (NRC) committee and USGS that the separate approach had the potential for missing important surface water-groundwater linkages that could have profound effects on the water quality of both systems. Consequently, the decision was made to consolidate groundwater and surface water study units, although most of the study units were either groundwater or surface water dominated.
USGS was authorized by Congress to establish a pilot program in 1986 with seven pilot study units representing a diversity of hydrologic environments and water-quality conditions, four of which were surface water dominated (the upper Illinois River basin in Illinois, Wisconsin, and Indiana; the Kentucky River basin in Kentucky; the lower Kansas River basin in Kansas and Nebraska; and the Yakima River basin in Washington) and three of which were groundwater dominated (the Delmarva Peninsula in Delaware, Maryland, and Virginia; the Carson basin in Western Nevada and Eastern California; and the Central Oklahoma aquifer in central Oklahoma) (NRC, 1990). USGS requested the NRC to undertake a 2-year evaluation of the pilot studies in 1987, and the NRC responded with A Review of the USGS National Water Quality Assessment Pilot Program (NRC, 1990). This NRC committee was invited to assist in the evolution and refinement of the NAWQA design as it moved toward full-scale implementation, deliberating on several NAWQA planning documents, issuing an interim report, and visiting the seven pilot study units. The NRC committee was supportive of the NAWQA effort (Box 2-2).
The success of the pilot effort led to NAWQA’s full-scale implementation in 1991 with the program goals of status, trends over time, and understanding as cornerstones of the program mission—cornerstones that have not changed through the evolution of the program. At the time of its
Perspective from NRC (1990)
“The [NRC] committee is convinced that there is a genuine need for a long-term, large spatial scale national assessment of water quality in the United States. Human health and environmental health are inextricably linked to our nation’s water quality…. The [NRC] committee is convinced that a national scale, long term water quality assessment is in the best interest of the country.”
SOURCE: NRC, 1990.
conception, NAWQA was the largest water resources program ever undertaken by USGS (R. J. Gilliom and R. M. Hirsch, personal communication, May 13, 2009).
In the first decade of water-quality monitoring (Cycle 1, 1991-2001) NAWQA set out to (1) accumulate high-quality, multidisciplinary, water-quality data and (2) generate a national synthesis of those data focusing on analysis of the highest-priority issues that cuts across the geography and answers the question, “How is the nation’s water quality changing?” The program demonstrated considerable progress toward a national water-quality assessment in Cycle 1. For thoroughness and to place this report in context, the committee notes key components of Cycle 1 here. (For a detailed review of Cycle 1 see NRC .)
The Study Unit Concept
The Cycle 1 study units accounted for 60 to 70 percent of the nation’s water use and population served by public water supplies and covered about one-half of the land area of the United States. A broad suite of physical, chemical, and biological constituents was selected based on relevance to water-quality issues and existing analytical methods including measurements of:
- dissolved oxygen,
- specific conductance,
- major ions,
- trace elements,
- organic carbon,
- pesticides, and
- volatile organic compounds (VOCs) (NRC, 2002).
Also, descriptions of biological communities were made based on different taxonomic groups and habitat conditions (NRC, 2002). A suite of surface water reference sites, a sampling site selected for relatively undisturbed conditions, was built into the surface water network design. At the end of Cycle 1, monitoring at 51 study units plus a study of the High Plains Aquifer in the central United States were completed. (The geographic scope of
the original design was 59 study units, which was adjusted to 51 to account for fiscal restrictions.) The High Plains Aquifer study was a pilot study for a regional approach to a groundwater assessment in the southern High Plains and was added near the end of Cycle 1.
In three groups over time, the study units were phased in during Cycle 1: study units 1-20 in 1991, study units 21-36 in 1994, and study units 37-51 in 1997 (Figure 2-1). At the onset, each study unit had a 2-year startup phase with time for planning and analysis of existing data, which was a major effort. At the same time, each study unit was developing liaison committees with local stakeholders, which became critical to guide how each study unit analysis was carried out and how the results were used to enhance water management. Within each study unit, an integrated group of scientists addressed the three primary objectives by (1) making measurements, (2) evaluating these measurements to understand water quality, and (3) making statements about what is learned and known about a particular study unit. After the 2 year startup, each study unit entered a 3 year intensive data-collection stage. This was followed by a period of data analysis and completion of major reports and then low-level monitoring
and assessment activities. Following a short period of retrospective analysis, each study unit would ramp back up and enter the intensive data-collection phase again—10 years after the previous data-collection phase (Figure 2-2).
This fixed site design with periodic rotational sampling allowed NAWQA to collect data at regular snapshots in time and document trends. Sampling a total of 505 stream sites and more than 6,000 groundwater wells, each study unit assessment resulted in many individual publications. At the end of 2001, more than 1,000 NAWQA publications were available (NRC, 2002). Also, the study units effectively bridged the environmental system because of a tailored sampling strategy in each study unit (ground-water and/or surface water; the water column and/or bed sediment; pesticides and/or nutrients) and a diverse team of scientists working on each assessment. The similar design of each study unit investigation and the use of standard methods made it possible to compare results between different study units, thus enabling multiple scales of investigation or regional and national assessments. These regional and national assessments, referred to as “national syntheses,” aggregated water-quality information and also allowed for analysis of important national issues such as, for example, non-point source pollution.
NAWQA phased in national synthesis assessments during Cycle 1, conducted by national synthesis teams. These included pesticides and nutrients in 1991, VOCs in 1994, and trace elements and ecology in 1997. Criteria for the selection of these topics considered a combination of understanding stakeholder priorities, capturing appropriate scale (i.e., topics should affect a large area or many small areas), representing persistent and recurring issues, importance to the study units that were in place, and complementing other national synthesis topics. NRC (2002) commended NAWQA for its groundbreaking work in these areas during Cycle 1.
NAWQA activities were developed with an “environmental framework” or a broader context through which the data were related to the bigger, environmental picture. This framework, composed of “common natural and human-related factors, such as geology and land use,” was used “to compare and contrast findings on water quality within and among study units in relation to causative factors and, ultimately, to develop inferences about water quality in areas that have not been sampled” (Gilliom et al., 1995). The environmental framework was reflected in the entire program design from sampling type to the interdisciplinary staffing structure. Application of the environmental framework assisted the program in, for example, choosing a drainage basin to study or a set of indicator sites. The environmental framework concept was and is today a touchstone for program efforts.
The second cycle of water-quality monitoring (Cycle 2) began in 2002 and extends to the end of fiscal year (FY) 2012, slightly past the duration of this committee’s review. Per the original design, NAWQA implemented a shift toward trends and understanding as the program moved out of Cycle 1. NAWQA integrated a number of new components as a result of evaluations from the Cycle 2 National Implementation Team (NIT), input from NAWQA personnel who were the primary drivers of the original design, and recommendations from the 2002 NRC report.3 NAWQA investigated select new contaminants and addressed many complexities involved with their environmental occurrence such as seasonal variations, degradation products, and chemical mixtures. These new activities were pursued
3 Approximately 80 percent of the 2002 NRC recommendations were implemented by NAWQA, and those that were not were omitted largely because of funding restrictions.
through program components such as Topical Studies and the Source Water Quality Assessment, discussed in the following pages. However, because of limited funding NAWQA was unable to pursue the following recommendations from the NRC report (2002): sample lakes and reservoirs that are important sources of water supply; enhance sediment monitoring, enhance interpretation, and make sediment a topic of a national synthesis team; and add pharmaceuticals, high production volume chemicals, and waterborne pathogens and microbial indicator organisms to the list of contaminants monitored in Cycle 2. The program also continued to assess the current water quality of the nation through standardized data collection, in concert with the goal of assessing long-term water-quality trends. Planned activities were grouped into 12 themes:4
- drinking water sources
- trends in status
- response to agricultural management
- response to urbanization
- sources of contaminants
- transport to and within groundwater
- transport to and within streams
- groundwater and surface water interactions
- effects on aquatic biota
Each theme correlated to NAWQA’s goal of status (themes 1-3), trends (themes 4-6), and understanding (themes 7-12).
In Cycle 1, NAWQA focused 80 percent of program resources on the status effort, continuing to establish the nation’s baseline water-quality condition. This was reduced to 20 percent of available resources in Cycle 2, although NAWQA did enhance the status activities with the Source Water Quality Assessments, an examination of the drinking water in communities across the United States, corresponding to status theme 2 (Delzer and Hamilton, 2007). The program placed an increased emphasis on trends (40 percent of program resources) and understanding (40 percent of program resources) through planned topical studies with a source, fate, and transport perspective (Figure 2-3).
51 monitored in Cycle 1 (Figure 2-4). NAWQA conducted a detailed analysis to determine which study units should be discontinued or consolidated and which were the most representative study units. Discontinued study units include those in Hawaii (the Oahu Study Unit), Alaska (the Cook Inlet Basin Study Unit), and the Lower Susquehanna basin in Pennsylvania (the Lower Susquehanna River Basin Study Unit). For example, the decision was made to discontinue the Hawaii study unit because of low population density relative to water use in comparison with other study units. Low population density or low water use criteria drove the discontinuance of most of the other study units as well.
As Cycle 2 progressed, perhaps the most notable design change began in 2004. The program transitioned away from the study unit focus and moved to a larger-scale regional design for status and trends assessment because of limited resources. The regional design retained a core of status and trends monitoring still conducted within the study units, but de-emphasized the role of more detailed study unit investigations and their individual teams and liaison committees. Status and trends data analysis and modeling, as well as program products, were shifted to teams organized by 8 Major River Basins (MRBs) and 19 Principal Aquifers (PAs) (Figures 2-5 and 2-6).
The MRB and PA regions are similar in concept to the role of study units as the building blocks of Cycle 1, but on a larger scale that collectively includes the conterminous United States, albeit at lower resolution. Cor-
responding to the study unit redesign, monitoring for specific conductance and temperature ceased, and pesticide monitoring at reference sites was discontinued. Also, the role of study unit liaison committees was reduced, which in turn reduced the degree of local stakeholder input to NAWQA (see Chapter 5 for further discussion).
NAWQA expanded efforts toward modeling in Cycle 2, to allow the program to extrapolate water-quality conditions across the country in areas not sampled by the program. This began in 2002 with an assessment of nutrient conditions in six large regions across the country using the SPAtially Referenced Regressions on Watershed Attributes (SPARROW) model (Smith et al., 2003). Later, mid-Cycle 2, the shift from study units to MRBs and PAs was considered an opportune time to begin developing planned regional-scale water-quality models. For example, a regional-scale SPARROW model was developed for the southeastern United States (Hoos and McMahon, 2009).
NAWQA increased efforts to communicate and disseminate its products and information. NAWQA moved from dissemination through paper reports in Cycle 1 to a multimedia in Cycle 2. Communication strategies were created for each major report, and more web-based dissemination and decision-support tools were initiated to reach a variety of audiences. Components of the enhanced communication effort included5:
5 NAWQA leadership, personal communication, May 9, 2009.
- improvements to the NAWQA website,
- creation of a publication search engine,
- multi-level rollouts of high-visibility findings,
- detailed communication plans for visible reports,
- improved data warehouse with data mapping capability,
- web-based decision-support tools, and podcasts.
Status and Trends Networks
As NAWQA moved from Cycle 1 to Cycle 2 in the midst of planned and unplanned program design changes, the status and trends sampling networks also changed. As described above, these changes emphasized regional assessment and resulted in more regional-scale analysis in Cycle 2.
Surface Water Status and Trends Network
In Cycle 1 505 surface water sites were sampled in 3-year, intensive, water-quality sampling periods per the original design. Of the original 505 stream sites monitored in Cycle 1, 145 were selected for annual trends monitoring at the start of Cycle 2 as specified in the Cycle 2 NIT’s report, Study Unit Design Guidelines for Cycle II of the National Water Quality Assessment Program6 (Gilliom et al., 2000). However, by 2005 available funding could only support monitoring of 84 sites annually, which lasted about 2 years until 2007. Since 2007, NAWQA has maintained 113 sites at the expense of an annual sampling strategy; a rotational design was employed where the majority of the sites were sampled 1 year out of every 4 years (Table 2-1 and Figure 2-7). Twelve of these sites, the larger river integrator sites or sites on large rivers that drain significant agricultural and urban areas, are sampled yearly. All Cycle 2 sampling sites were selected from Cycle 1 sites for NAWQA to preserve and maintain a long record of consistent data that is useful for trend analysis (G. Rowe, personal communication, May 17, 2010).
Aquatic Ecology Status and Trends Network
The aquatic ecology sampling network was cut back even more than the surface water network in Cycle 2, to 58 sites, rotated and sampled every other year (Table 2-1). NAWQA divided the country into two sections, western and eastern, with sampling rotating back and forth along with detailed investigations that continue today. NAWQA’s philosophy on selecting these sites used environmental framework as a touchstone, with two key components: (1) choose representative sites and (2) scale the sampling design to accommodate the size of the river. During sampling, a detailed habitat assessment is performed—algae, invertebrates, and fish samples are taken at each site along with a riparian assessment—assessing the biological status of each site. All sites are co-located with sites where water chemistry, bed sediment, and streamflow sampling occur. This pro-
6 The NIT’s report describes the design and implementation strategy for Cycle 2 investigations. The report Opportunities to Improve the U.S. Geological Survey Water Quality Assessment Program reviewed this report (NRC, 2002).
TABLE 2-1 The Evolution of NAWQA Program Status and Trends Networks
|Cycle 1||Cycle 2|
|Number of surface
water sampling sites
|Number of aquatic
|416||125||75||58 (6 sites are
networksb and wells
aThe ecology sites are included in the total number of sampling sites.
bA groundwater network is a group of sampling wells.
duces a “reach assessment” where NAWQA probes what organisms are exposed to in a given watershed. To enhance NAWQA’s ability to use these data to provide a national assessment of ecological conditions given limited sampling, the program collaborated with the U.S. Environmental Protection Agency’s (EPA’s) Environmental Monitoring and Assessment Program (EMAP)7 to paint a picture of water quality in the western United States based on indicator organisms (Carlisle and Hawkins 2008). (For additional information, see Chapter 5.)
Groundwater Status and Trends Network
In Cycle 1 NAWQA had approximately 272 groundwater networks or clusters of sampling wells, for a total of 6,307 wells sampled throughout the study units for groundwater status and trends (Figure 2-8, top). In Cycle 2, NAWQA viewed activities on the basis of Principal Aquifers representing a more regional view than Cycle 1, i.e., Principal Aquifer Assessments. The program sampled 137 groundwater networks for a total of 3,698 wells (Table 2-1, Figure 2-8, bottom), evaluating conditions and trends in 19 regional aquifers per the NAWQA Cycle 2 Implementation Team’s original design (Gilliom et al., 2000). These 19 aquifers account for approximately 75 percent of the estimated withdrawals of groundwater for drinking water supply in the United States (Lapham et al., 2005).
Unlike the Surface Water Status and Trends Network and the Aquatic Ecology Status and Trends Network, the groundwater network design remained largely unchanged through the duration of Cycle 2. This design included decadal sampling of all wells, biennial sampling of 5 wells within each network, and seasonal sampling of wells selected for biennial sampling (Gilliom et al., 2000). This was, in part, due to the relatively modest scale of the original Cycle 2 Groundwater Status and Trends Network design but also to the slow hydrologic response time of groundwater, permitting more flexibility in correlating the timing of, for example, biennial sampling with budgetary realities (NAWQA leadership, personal communication, July 20, 2012).
Status and Trends Assessments and Activities
During Cycle 2, NAWQA mined the 10 years of monitoring data from Cycle 1, augmented by continued monitoring in Cycle 2, to determine and publish long-term assessments of trends in the nation’s water quality. The program synthesized data from the Surface Water Status and Trends Network, Aquatic Ecology Status and Trends Network, and Groundwa-
Reports on Water-Quality Trends in the Major River Basins
1. Trends in Nutrient Concentrations, Loads, and Yields in Streams in the Sacramento and Santa Ana Basins, California, 1975-2004 (Kratzer et al., 2011)
2. Nutrient and Suspended-Sediment Transport and Trends in the Columbia River and Puget Sound Basins, 1993-2003 (Wise et al., 2007)
3. Trends in Nutrient and Sediment Concentrations and Loads in Major River Basins of the South-Central United States, 1993-2004 (Rebich and Demcheck, 2007)
4. Nutrient and Suspended-Sediment Trends in the Missouri River Basin, 1993-2003 (Sprague et al., 2007)
5. Trends in Streamflow, and Nutrient and Suspended Concentrations and Loads in the Upper Mississippi, Ohio, Red, and Great Lakes River Basins, 1975-2004 (Lorenz et al., 2009)
6. Trends in Pesticide Concentrations in Corn-Belt Streams, 1996-2006 (Sullivan et al., 2009)
ter Status and Trends Network and released a variety of products. Many water-quality trends emerged at the local, regional, and national levels. For example, NAWQA showed that the investment by the Bureau of Reclamation in improving the water quality of the Colorado River resulted in a decrease in dissolved solids downstream of salinity control projects (Anning et al., 2007). A select group of status and trends publications and results are highlighted here; Chapter 3 and other areas of this report continue this discussion.
At the time of this report, six trends reports exist for the major river basins on nutrients, sediment, and pesticides (Box 2-3).8 NAWQA also has published a variety of information assessing the PAs.9 For example, the High Plains Aquifer Professional Paper summarizes the water quality in this aquifer and was NAWQA’s first systematic regional assessment of groundwater (McMahon et al., 2007). Also, NAWQA National Synthesis Teams synthesized the results from NAWQA investigations for priority water-quality issues (pesticides, nutrients, VOCs, ecology, and trace elements) and produced capstone reports on pesticides, nutrients, and VOCs in Cycle 2. The National Synthesis Assessments are discussed further in Chapter 3.
NAWQA’s Source Water Quality Assessments (SWQA) examined drinking water quality of community water systems across the United States by comparing compounds in raw ambient water collected at a supply well or surface-water intake prior to treatment (i.e., “source water”) to compounds in the finished water supplied to the community (Delzer and Hamilton, 2007). The assessment focused on 280 unregulated organic compounds with a focus on VOCs and pesticides. Carter et al. (2007) provide information on the design and analytical methods used in the SWQA. While a diverse group of compounds were present in source water, the majority of the compounds assessed were present only at low concentrations (<< 1 ppb). Compounds detected in source water were often in finished water, although compounds detected in finished water were below human-health benchmarks if one existed. Mixtures of compounds were commonly detected in both. Capstone products were released in 2008 and 2009 (Hopple et al., 2009; Kingsbury et al., 2008).
The understanding component of NAWQA was carried out in Cycle 2 through five hypothesis-driven topical studies. The conceptual approach of these studies was to understand contaminant source, fate and transport, and impacts on humans and aquatic ecosystems. NAWQA took a mass balance approach to the studies, understanding that a mass balance of water and a mass balance of constituents go hand in hand (i.e., scientists should understand how water is flowing through the system in order to eventually understand the effects of contaminants). NAWQA integrated the use of models into a few of the topical studies. With each topical study, NAWQA adhered to the concept of a national program with a focus on a national understanding of water-quality problems. In each of the five topical studies, NAWQA probed multiple locations, scales, and gradients (i.e., multiple climate, landscape settings, hydrology, crops, land use settings, and atmospheric deposition settings). The topical studies were nested within the study units of Cycle 1, using knowledge gained in Cycle 1:
- Topical Study 1: Agrochemical Sources, Transport, and Fate10
- Topical Study 2: Effects of Nutrient Enrichment in Stream Ecosystems11
- Topical Study 3: Mercury Cycling in Stream Ecosystems12
- Topical Study 4: Effects of Urbanization on Stream Ecosystems13
- Topical Study 5: Contaminant Transport and Public Supply Wells14
The topical studies produced a variety of interesting findings, published in methods papers, comprehensive journal article series, and USGS reports. Due, in part, to an underestimation of the amount of work associated with these efforts, some topical studies progressed further than others during Cycle 2. For example, the mercury study (Topical Study 3) documented methylmercury concentrations across the United States and observed that the highest levels of methylmercury in fish are found in the southeastern United States and in mined areas in the western United States (Scudder et al., 2009). (Methylmercury is the most toxic form of mercury in the environment and is readily taken up by aquatic organisms.) NAWQA noted that major urban centers are experiencing a significant increase in mercury deposition. Finally, because of biogeochemical properties of methylmercury, concentrations of the contaminant in streams are driven by wetland density and dissolved organic carbon concentration (Figure 2-9).
Monitoring … to Monitoring and Modeling … to the User
In Cycle 2, NAWQA moved from monitoring to monitoring and modeling water quality of the nation’s groundwater, surface water, and ecology at all scales (i.e., using deterministic models at smaller scales and statistical regression at large scales). The NAWQA Cycle 2 modeling approach is to use monitoring data and stream network to probe water quality from the regional and national to the local scales. Modeling efforts amplify the program goals through (1) extrapolation of water-quality conditions to unmonitored areas to facilitate a “national assessment” and (2) forecasting of conditions and simulation of the effects of changes in influencing factors (test scenarios). As Cycle 2 draws to a close, the modeling efforts are improving understanding of the factors (sources, transport, etc.) that influence water quality.
The goal of one of NAWQA’s first exercises in modeling was to predict groundwater vulnerability to nitrate contamination at the national scale. The program showed this vulnerability based on monitoring data, fertilizer data, and soil characteristics, which were integrated into a model called GWAVA (Ground-WAter Vulnerability Assessment). In the southeastern United States NAWQA reported lower concentrations of nitrogen where de-nitrification is promoted compared to the central plains (Nebraska), where the United States has high fertilizer use, gravel and sand, fast transport, and
lack of denitification (Nolan and Hitt, 2006). EPA uses this information to help prioritize monitoring and better assess its regulatory efforts.
During Cycle 2 NAWQA developed empirical models to probe hydrologic alteration nationwide as well as the connection between hydrologic alteration and the structure of macroinvertebrates and fish assemblages. NAWQA successfully modeled ecologically important flow metrics under a “natural” or “minimally disturbed” flow regime using geospatial data and a reference condition approach. This opened the possibility of quantification of hydrologic alteration across the United States (Figure 2-10). Using geospatial models and NAWQA data, Carlisle et al. (2011) demonstrated that diminished magnitude of flows was the best predictor of impairment of macroinvertebrate and fish assemblages nationally. NAWQA integrated macroinvertebrate data (collected by NAWQA and the EPA Wadeable Stream Assessment15) to expand the scope of a model assessment of biological condition in streams in the western United States (Carlisle and Hawkins, 2008). These studies are the foundational material for a USGS Circular summarizing findings on aquatic communities across the United States prepared by the Ecological National Synthesis Project, planned for 2012.
The SPARROW16 model is NAWQA’s most popular and visible regression model.17 The SPARROW model is a watershed based model designed to predict patterns in water quality, concentration, and amount of constituents, across spatial extents ranging from entire regions of the United States to smaller watersheds. The model is perhaps best known for contributing to understanding of key parameters that affect hypoxia in the Gulf of Mexico by determining nutrient load to the Gulf and pinpointing which watersheds or which of the 31 state drainage basins are the greatest contributors. Specifically, the SPARROW effort highlighted that nine states18 making up one-third of the Mississippi River drainage area contribute 75 percent of the nitrogen and phosphorus to the Gulf (Alexander et al., 2008). This study also filled gaps in the understanding on the sources of phosphorus in the Gulf; phosphorus associated with animal manure contributes almost as much phosphorus as cultivated crops (37 versus 43 percent) (Alexander et al., 2008).
Currently NAWQA is in the process of developing fine-scale, regional water-quality models in each MRB. Nutrients are the focus of these modeling efforts, except in the arid southwest, where dissolved solids are of greater importance. To do this, NAWQA is using local ancillary data and refining the SPARROW model to reflect the unique environmental conditions and smaller scale of each MRB. At this time, models have been developed for six of the eight MRBs. Regional models for the remaining basins, California and the Southwest, are planned for the future. The preliminary findings from this effort show the promise of future regional SPARROW modeling of water-quality conditions in the United States. The October 2011 issue of the Journal of American Water Resources Association provides a featured collection of articles on the regional SPARROW effort.19
NAWQA is exploring uncertainty in all the modeling efforts, i.e., associating uncertainty with all the estimates the program produces. For example, Robertson et al. (2009) examined approximately 800 watersheds in the Mississippi River basin and assigned a ranking that indicated whether nutrient yields from the basin were among the highest delivering of nutrients contributing to hypoxia in the northern Gulf of Mexico (Figure 2-11, top). This involved a robust statistical procedure applied to the results from a previous application of SPARROW to identify the top 150 watersheds. Once identified, scientists incorporated information on confidence intervals
17 Development of SPARROW was initiated by the Branch Systems Analysis working on new and emerging technical issues and techniques used within the former Water Resources Division. The branch was dissolved in the late 1990s because of funding shortfalls, and the individuals developing SPARROW joined NAWQA and continued their work.
18 Illinois, Iowa, Indiana, Missouri, Arkansas, Kentucky, Tennessee, Ohio, and Mississippi.
of these model predictions estimating the probability that these watersheds are among those that have the highest nutrient yields to the Gulf (Figure 2-11, bottom). This was a SPARROW spin-off project and was EPA driven. This information has important management implications for the Midwest and is being used by EPA to target non-point source pollution in those watersheds.
NAWQA is offering the use of monitoring and modeling tools to the user, an effort that will extend into Cycle 3. Although these efforts are still in their infancy, they represent a significant step forward for NAWQA and the user community. For example, the Watershed Regression for Pesticides models, referred to as WARP models, predict specific concentration statistics for a given pesticide in the United States. These models establish linkages between pesticides measured at NAWQA surface water sampling sites to variety of factors (pesticide use, soil characteristics, hydrology, and
climate) that affect pesticides in streams. One of the first completed WARP models was for the pesticide atrazine (Larson and Gilliom, 2001), which was improved during Cycle 2 (Larson et al., 2004). Today, the atrazine WARP model and associated data are available for public use on the web.20 The user can visit a website and see estimates of atrazine concentrations in an area or basin along with the error and uncertainty associated with that estimate. NAWQA scientists are planning to bring other pesticide data to the web in a similar fashion.
Another example of bringing modeling and monitoring activities to the user, the SPARROW Decision Support System provides online access to SPARROW models that can be used to predict long-term average water-quality conditions and source contributions by stream reach and catchment and to evaluate management source-reduction scenarios (Booth et al., 2011).21 (For additional information see Box 4-1.) Also, USGS and EPA are working together to provide interested parties with a web service to assist in integrating large water-quality databases.22 Users can go into the USGS website and retrieve data from the National Water Information System, which includes water-quality data from NAWQA, in a common format and go to the state EPA data (STORET) and retrieve data formatted in the same way.
Using the FY2011 appropriations for USGS as the metric, NAWQA’s budget of $62.9 million was approximately one-third of the appropriation for water-related programs at USGS (the former Water Resources Discipline area). Although the allocation of the budget evolves with programmatic design, in FY2010 the majority of NAWQA’s budget was used for program activities (for example, status and trends networks) versus program management or support of broader USGS efforts (Figure 2-12). The appropriations in actual or nominal dollars for NAWQA have been flat since the late 1990s or declining when adjusted for inflation (Figure 2-13). This has been consistent with the overall budget and staffing trends of water-related programs at USGS over the past 16 years, which are flat or declining (NRC, 2009).
NAWQA is visible to the public via the data and interpretive delivery systems the program strives to make publicly available, and the program has a record of scientific achievement since its inception (NRC, 1990, 2002, 2009, 2010, 2011a; USGS, 2010). NAWQA has produced approximately
1,900 reports during its 20-year history, a publication every 4.2 days on average, a value which, while not an indicator of quality, provides a sense of the quantity of work produced over the history of the program. (M. Larsen, personal communication, May 13, 2009). If released products are the metric (those already released and to be released), NAWQA has mined approximately one-third of the Cycle 1 data (NAWQA leadership, personal
communication, May 9, 2009), a value that, although not an indicator of quality, provides a sense of the quantity of work produced over the history of the program. A Customer Satisfaction Survey, conducted in 2010,23 indicates that the majority of NAWQA users are either satisfied or very satisfied with NAWQA information (Figure 2-14).
The Statement of Task charges the committee to conduct an assessment of NAWQA’s accomplishments. In response, the committee notes 10 representative accomplishments of NAWQA in Chapter 3 to answer the Statement of Task.
23 The 2010 NAWQA Customer Satisfaction Survey, referenced several times in this report, was conducted by the USGS Office of Budget, Planning, and Integration. It was conducted in July and August of 2010 and consisted of a random sample of 500 persons from the NAWQA stakeholder database. The response rate to the survey was 37 percent.