Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 121
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program 5 NAWQA Cycle II Goals—Understanding INTRODUCTION “Understanding” is the last of the three primary goals set for the National Water Quality Assessment (NAWQA) Program (with the status and trends goals described in Chapters 3 and 4, respectively). The NAWQA Cycle II Implementation Team (NIT) guidance document (Gilliom et al., 2000; see also Appendix A) states that this goal “is to improve explanation and understanding of: the sources of contaminants, their transport through the hydrologic system, the effects of contaminants and physical alterations on stream biota and ecosystems, . . . and implications for the quality of drinking water.” To make progress in the realm of understanding the major factors that affect water quality, model development and application are essential. Understanding can be gained through the linkage of field studies with the analytical use of models, where observations are compared to a conceptual relationship expressed mathematically. Success of such a model in explaining observations is regarded as a measure of understanding the primary factors or mechanisms involved. Conversely, model development and application require understanding. The studies that provide an understanding of contaminant sources and their transport can be viewed as the raw materials for the design and development of water quality models. The previously described status and trends networks can provide such information, and therefore the proposed application of models in Cycle II rests firmly on the knowledge gained from Cycle I. Models are developed to provide predictions of water quality conditions both spatially and temporally (i.e., through geographic extrapolation or prediction of future conditions). Detailed understanding of contaminant sources and their trans-
OCR for page 122
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program port to water resources and an ability to predict future conditions are key to the development of efficient management and policies to protect the beneficial uses of the nation’s water resources, including drinking water and viable ecosystems. As NAWQA progresses into Cycle II with an increased emphasis on its understanding goal, the importance of model application, as recommended by previous National Research Council (NRC, 1990, 1994) committees, should not be underestimated. Understanding and prediction, embodied in water quality models, are the cornerstones of water resources management for the future. This chapter discusses four topics related to the goal of understanding cause and effect as related to water quality. First, the important role that models play in scientific understanding is explored. Understanding evolves from the thorough evaluation of data, and this is certainly one of the primary functions of model application. Models represent conceptual and mathematical relationships between the observations that we interpret as cause and effect. They are formulated at a wide range of temporal and spatial scales to represent a variety of phenomena— from rapid chemical reactions to long-term changes in the global environment. Each scale represents a different aggregation of controlling variables, so models are often not directly transferable between scales. This results in a hierarchy of models according to scale and highlights the importance of choosing models appropriately. They are also formulated in a variety of ways, including those based on mass-balance, statistical regression, and process-based (mechanistic) models. Regardless of the means of model formulation, there are several sources of uncertainty, including measurement uncertainty in the observations that serve as input and verification, model structure uncertainty, and parameter uncertainty. The quantification of uncertainty is important in providing perspective for interpretation of model results. Despite uncertainty, models play an important role in understanding causal factors that affect water quality, and model application is one way to illuminate the degree of understanding that exists. Different aspects of these topics are reviewed later in this chapter to highlight associated strengths and pitfalls in model application. Second, the practical aspects of the overall proposed Cycle II implementation strategy are discussed. As a starting point, Cycle I of NAWQA naturally serves as the foundation for Cycle II. For example, key components of the surface water and groundwater monitoring assessments as well as the pilot-scale monitoring in lake and reservoir sediments conducted in Cycle I are carried over and expanded in Cycle II to form the National Trend Network for Streams, National Trend Network for Contaminants in Sediment, and National Trend Network for Ground Water (discussed in Chapter 4). Cycle I activities will also help form the planned Cycle II spatial studies of effects of land-use change on stream and groundwater quality. These trend assessments (both spatial and temporal) provide both the initial information needed to design the “targeted studies” proposed for Cycle II (see more below) and a context that gives a sense of their national priority (see in discussion of trends in Chapter 4). The two new components
OCR for page 123
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program proposed for Cycle II that specifically reflect an increased emphasis on understanding and model application are: targeted water quality studies (that will variously address the six themes and 17 objectives of the understanding goal), and hydrologic systems analysis (HSA) to be guided by a newly formed Hydrologic Systems Team (HST). As discussed in detail later, the plan is for “targeted studies” to focus on a limited set of the most important water quality topics and link these studies with other parts of the Cycle II design (Gilliom et al., 2000). Thus, each specific targeted study will be designed and executed by various topical teams composed of one or more planned Cycle II study units, and they will be assisted by a single, nationally responsible HST. These so-called targeted studies will focus on the major factors that govern water quality. The contaminants studied in Cycle I (e.g., pesticides, nutrients, volatile organic compounds) and the new drinking water source status assessments planned for Cycle II (see Chapter 3) are expected to provide the foundation for targeted study design wherever possible. Third, the six themes and their corresponding objectives developed to describe the understanding goal (Gilliom et al., 2000; see also Appendix A) in Cycle II are assessed. These six themes were developed by categorization of the wide variety of scientific studies to be conducted by Cycle II study units to determine where commonalties or synergies between studies exist on a regional or national scale. Further perspective on the problem of setting goals and priorities on a national scale can be gained by relating the themes to the conceptual “source-transport-receptor” (STR) model described in the NIT guidance. The STR is a simple conceptual framework that can be used to organize a wide range of studies from across the nation to show both how they relate to each other and where information may be lacking. Although it is not explicitly stated in the latest NIT report, five of the six themes proposed for the understanding goal can be categorized according to the STR model as follows: Model Component NIT Understanding Theme Source Sources of contaminants Transport Contaminant movement from land surface to groundwater Contaminant movement from land surface to surface water Groundwater-surface water interactions Receptor Effects on stream ecosystems
OCR for page 124
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program The last understanding theme proposed for Cycle II, “extrapolation and forecasting,” relates to the use of models for synthesis and prediction of water quality conditions for unsampled geographic areas (extrapolate) and future conditions (forecast). Indeed, it is the objective of Cycle II to “link the status and trend studies with an understanding of the natural and human factors that affect water quality” (Gilliom et al., 2000). Defining the relationships of the studies to each other suggests how regional- and national-scale understanding can be approached and is the first step in linking diverse information to use in setting future directions and in identifying gaps in understanding that may hinder extrapolation and forecasting. Thus, it is important that the targeted studies are interpreted in the context of these themes (or other conceptual frameworks) so that they may be evaluated in terms of geographic applicability and national priority. The discussion of extrapolation and forecasting and the many portions of this chapter that lead into it directly address the committee’s statement-of-task issues (2) (extrapolation) and (4) (aggregation of information at regional and national scales) (see Chapters 1 and 8 for further information). The fourth and final section includes the conclusions and recommendations for this chapter. The general conclusions of each section are followed by specific recommendations presented in the same order as the chapter sections. This chapter, focusing on the understanding goal for Cycle II and the related themes and objectives put forward by NAWQA, directly addresses the committee’s state-ment-of-task issue (1), on methods for the improved understanding of causative factors affecting water quality conditions. ROLE OF MODELS IN UNDERSTANDING CAUSE AND EFFECT Introduction Models may serve several functions with respect to understanding the causes and effects of water quality conditions. For example, they can be used as tools for diagnosis and explanation of the underlying mechanisms associated with the fate and transport of pollutants in the environment. They can serve as frameworks to integrate observations that vary in time and space and to predict the spatial and temporal distribution of future events. Models also are useful to focus and organize one’s thinking about an environmental system. Good models should be consistent with current scientific understanding, and insofar as they are representative of reality, models provide a framework for organizing, communicating, and applying knowledge. Model application, as recommended by the previous NRC committees, is gaining in importance as NAWQA progresses into Cycle II with an increased emphasis on the goal of understanding the major factors that affect observed water quality conditions (status) and trends. Understanding contaminant sources and contaminant transport to water resources is important for the development of
OCR for page 125
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program effective management strategies to protect drinking water quality and the condition of stream ecosystems. Targeted studies and proposed model applications, which provide higher resolution (in time and space) than synoptic surveys, are important for identification and differentiation of natural and anthropogenic sources of contaminants in groundwater and surface water. Models are becoming increasingly important for water quality management, a notable example is the model used for the U.S. Environmental Protection Agency (EPA) Total Maximum Daily Load (TMDL) Program (NRC, 2001). For such applications, models may have a variety of uses that extend beyond their initial purpose(s). Models exist in many forms. The models expected to play a major role in Cycle II of NAWQA are categorized into statistical models and analyses, geographic information system (GIS) analyses, process-based models, and hybrid statistical-GIS-process based models. Mathematical models are often characterized as either empirical or mechanistic (process oriented), even though most useful models have elements of both. In the purest sense, an empirical model is based on a statistical fit to data; the only substantive input is in the selection of predictor and response variables. Similarly, in pure form, a mechanistic model is a mathematical characterization of scientific understanding of the critical processes in the natural system; the only data input is in the selection of model constants, initial and boundary conditions. Mass Balance The NAWQA Planning Team suggested that mass balance of constituents of concern could be utilized in a number of ways, especially related to explanatory science, a major activity proposed for Cycle II (Mallard et al., 1999). Simple mass-balance models will be utilized to help quantify the sources of contaminants to streams and aquifers. Mass balance (material balance) is frequently the beginning of quantitative assessment of the sources of nonpoint source (NPS) pollutants and their fate in the environment (e.g., Nolan, 1998; Vollenweider, 1975). Figure 5-1 shows a simplistic approach to developing a mass balance. A control volume is defined, and rates of mass inputs and mass outputs cutting across the surface of the control volume are estimated. Next, transport within the control volume and across its borders are characterized, and sources and sinks of the constituent of concern within the control volume are defined. Lastly, the rate of change in mass per unit time in the control volume is obtained. For dissolved or particulate waterborne constituents, mass is defined as the product of water volume and constituent concentration. Thus, a mass balance of the water is also essential for waterborne constituents. The spatial and temporal scales of a mass balance may vary over several orders of magnitude, depending on the objectives of the study in question. Control volumes may range from entire watersheds as in river basin models to infinitesimal slices as in numerical transport models. Other model characteristics also
OCR for page 126
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program FIGURE 5-1 Components of a simple mass-balance system and the continuity equation. vary. When the entire watershed is not subdivided spatially, it is assumed to be spatially homogenous and the model is characterized as “lumped parameter.” In contrast, when the watershed is subdivided, the model is characterized as “distributed parameter.” Essentially, there are subcontrol volumes nested in a control volume for distributed parameter modeling. Distributed mass-balance models may be further subdivided into process-level mass-balance models. The pollutant may be considered conservative or reactive and assigned appropriate rates. Transport may be assumed to be convective (advective) or a combination of convective and diffusive (dispersive) flow, and the model may be deterministic or stochastic. The choice of constructing a particular mass balance is dictated in part by the study objectives and the availability of data. There is a continuum of mass-balance efforts ranging from conceptual mass balance of aggregated data to comprehensive numerical models that operate in small time and space intervals. The requirements for constructing even a simplified, conceptual mass balance may entail substantial data and information collection. Take, for instance, determining sources of nitrogen in a stream where the land use is predominantly agricultural. The sources may include nitrogen inflow from the upstream reach, agricultural surface runoff, precipitation, discharge of collected subsurface drainage, groundwater discharge contributions into the stream, and so forth. The nitrogen species may include soluble (e.g., nitrate, ammonia, organic nitrogen) as well as particulate forms (e.g., sediment-bound nitrogen, plankton-bound nitrogen, other organic particulates). The nitrogen loading into the stream will have to be evaluated in terms of both concentration (mass per unit water volume) and mass (concentration × water volume). Thus, the nitrogen emission into a stream can become quite complex.
OCR for page 127
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program The examples taken from U.S. Geological Survey (USGS) studies in Box 5-1 illustrate the value of conducting mass balance for NPS pollutants. They not only provide significant insights into, and understanding of the problem situation, but also guide the design for further data collection, such as selection of fixed-site monitoring stations, and aid in the interpretation of the data collected. Moreover, the mass-balance approach may be used to track historical changes and provide predictions into the future with changes in management alternatives, land use, or environmental-ecological conditions, and so on. There are also other simple conceptual frameworks that can be used to view mass balance data that give insight into the behavior of systems. These include the concepts of “retention” and “relative residence time.” Retention is defined as the fraction of inflow mass that is retained in a control volume (such as a reservoir) and can be represented by the equation: R = (inflow mass – outflow mass) ÷ inflow mass (Dillon and Rigler, 1974). This is an expression of net losses, or sedimentation of conservative (nonreactive) substances. Retention can be used to compare systems to each other and to determine its relationship with other factors. Residence time of a substance can be thought of as replacement rate and may be defined as T = control mass ÷ inflow mass per unit time. For example, if a reservoir holds 100 million gallons (MG) of water and the inflow rate is 50 MG per year, then the residence time of water Tw is (100 MG ÷ 50 MG per year), or 2 years. The residence time of a nutrient or pollutant can similarly be calculated. As discussed in Chapter 2, providing such simple measures for lakes and reservoirs where NAWQA data exist would be of benefit and likely stimulate further needed work. In conclusion, conducting mass balances on constituents of concern is a worthy effort for Cycle II of NAWQA. Mass balances identify the sources and sinks of pollutants and their fate. They are a central tool for obtaining a better understanding of a problem situation. The committee recommends that at least a conceptual mass balance be developed for nonpoint source pollutants and constituents of concern for TMDLs in all Cycle II study units. The reactivity of a given NPS pollutant may be generic, but the extent of biological and chemical reactions of the pollutant is highly dependent upon site-specific conditions. Moreover, having such a knowledge base of mass-balance data is essential for evaluating the effectiveness of alternative best management practics to reduce pollutant loads. Depending on the availability of funding and expertise, more comprehensive mass-balance (e.g., hydrologic or hydraulic water quality) models might be applied to the more critical pollutants of selected study units. Statistical Models of Water Quality As noted above, a statistical (or empirical) water quality model reflects greater attention to fitting the pattern in a data set than to describing mechanisms. Perhaps the simplest example of a statistical water quality model is the relation-
OCR for page 128
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program BOX 5-1 Examples to Illustrate the Utility of Models Used by USGS The question of why mass balances and other model applications should be conducted is addressed below. Five examples of computations by NAWQA or other USGS study teams are briefly summarized to underscore the significance of their findings or implications. Example 1. Alexander et al. (2000) studied the effect of stream size on the delivery of nitrogen to the Gulf of Mexico. This study examined the flux of nitrogen from the Mississippi River to shallow waters of the Louisiana shelf in the Gulf of Mexico causing eutrophication and seasonal hypoxia. The average first-order loss rate of nitrogen shows rapid decline with channel size from 0.45 to 0.0005 per day. The proximity of nitrogen sources discharged into large streams and rivers is a major determinant of downstream transport of nitrogen to coastal waters. Example 2. Puckett et al. (1999) estimate nitrate contamination in an agro-ecosystem on an outwash aquifer using a nitrogen mass-balance budget. The Otter Tail outwash area (west central Minnesota) is intensively cropped. The nitrogen mass balance reveals that croplands contributed 89 percent of the excess total nitrogen, and fertilizers were the primary source of nitrogen leaching into the glacial outwash. If it were not for denitrification, the average nitrate leaching losses would be 19.3 instead of 10.8 kg of nitrogen per hectare-year. Example 3. McMahon and Woodside (1997) examined the nutrient mass balance for the Albemarle-Pamlico drainage basin in North Carolina and Virginia. The mass balance on total nitrogen and total phosphorus indicates the relative importance of agricultural NPS, the equal importance of atmospheric sources of nitrogen and phosphorus as compared to crop fertilizer, and the large amounts of nitrogen and phosphorus that are unaccountable or residual in the mass balance. The residuals reflect the uncertainty and error associated with estimating the various total nitrogen and total phosphorus mass-balance compartments. Example 4. Bachman and Phillips (1996) estimated the baseflow nitrogen load into the Chesapeake Bay. Nitrogen concentrations in well-drained soils were higher than in poorly drained soils. About 40 percent of the total nitrogen into the bay is from baseflow, indicating the importance of groundwater-stream interactions. Example 5. Smith et al. (1997) conducted a regional interpretation of a national water quality monitoring system. SPARROW was used to reduce problems of data interpretation caused by sparse sampling, network bias, and basin heterogeneity. The constructed regression models of total nitrogen and total phosphorus for monitoring 414 National Stream Quality Accounting Network streams throughout the conterminous United States provided insight into important sources and processes affecting nutrients in watersheds.
OCR for page 129
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program ship between concentration and streamflow, referred to as a rating curve model. Such models describe the often weak relationship between the concentration of a water quality constituent and river discharge. Rating curve models are used widely for extrapolating water quality loads when only a few measurements are available and for performing trend assessments. The widespread use of such statistical models is reviewed by Cohn (1995), who also discusses research developments and statistical issues involved in estimating loads from the concentration-discharge relationship. Since such rating curve models are used so widely, it is important that research be performed to improve these models. The data sets developed by NAWQA provide an excellent resource with which to test and improve rating curve-based water quality models. As is often the case, improvements in one’s ability to model the processes that give rise to water quality changes are likely to come from improvements in one’s ability to understand those processes. Even a purely statistically based water quality model can be improved by introducing a physical basis. O’Connor (1976) used a deterministic approach to derive expressions for the spatial and temporal distribution of conservative dissolved solids in rivers in such a way that they could be expressed as rating curves. Similarly, Duffy and Cusumano (1998) document the physical situations that give rise to hysteretic (looped) behavior of concentration-discharge relationships. The USGS has historically made many contributions to the literature on statistical water quality (rating curve) models (see Cohn, 1995, for a review) with a focus on statistical innovations. Future improvements in rating curve and other statistically based water quality models are likely to come from the integration of statistical and physical process-oriented models. NAWQA should devote greater attention to research that seeks to improve the important concentration discharge model based on physical reasoning and physical process interpretations of the concentration discharge relationship of the type introduced by O’Connor (1976) and refined by Duffy and Cusumano (1998). Statistical water quality models range from the very simple concentration-discharge models described above to the more complex regional multivariate statistical models such as SPARROW (Spatially Referenced Regression on Watershed Attributes). SPARROW is an effort to provide resource managers with spatially detailed information describing the location and magnitude of nutrient sources and watershed factors that affect the delivery of important chemical and biological constituents to receiving waters. Early applications of SPARROW to the Chesapeake Bay (Preston and Brakebill, 1999) and elsewhere have proven quite promising. The results of SPARROW modeling efforts led to an illustration of the spatial distribution of pollutant loads within the basin; such information is instrumental in watershed management programs that seek to target nutrient reduction areas. The committee recommends that NAWQA continue research relating to the improvement of SPARROW and to the application of this model to other watersheds in Cycle II. Continued comparisons should be per-
OCR for page 130
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program formed between the output of SPARROW and analogous, yet more complex, process-oriented watershed-based water quality models. Also, as discussed in Chapter 4, SPARROW can be explored further for regionalization and extrapolation of NAWQA data. Process-Oriented Models of Water Quality Process-based (or mechanistic) watershed models are mathematical representations of one’s current understanding of hydrologic and water quality processes. A number of comprehensive, physically based watershed models now exist that are intended to simulate most watershed processes, ranging from hydrological state variables such as groundwater, streamflow, and evapotranspiration to water quality and sediment transport. A common feature of these models is the large number of parameters required relative to the available data. As a consequence, parameter selection becomes an art, and the models suffer from a general lack of rigorous testing. This suggests that the large process models are likely to be more useful for quantitatively representing processes and interactions in research studies than for forecasting the outcomes of proposed management actions. In addition, overparameterization means that the models lack an error term, so the research applications of these models is limited to informal analysis, because formal hypothesis testing is not truly possible. One notable gap in process modeling identified in a recent NRC study (NRC, 2001) of the scientific basis for EPA’s TMDL program is the lack of simple process models. Simulation models can, in principle, be crafted on a continuous scale of process detail, and there is nothing inherently correct, or better, about the scale of existing process models. The fact that current models are overparameterized suggests that simpler mechanistic expressions identifiable from the available data (in other words, those that support parameter estimation from the available data) should be considered as a viable alternative. One recent example of a model designed from this perspective is described in Borsuk et al. (2001); by avoiding overparameterization, Borsuk et al. were able to optimize the model fit to available data and also calculate a prediction error term. Accordingly, the committee recommends that process modeling within NAWQA avoid overparameterized models whenever possible. To address problems of model selection, application, and analysis, a set of modular modeling tools, termed the Modular Modeling System (MMS; Leavesley, 1997) is being developed by the National Research Program of the USGS. The MMS approach attempts to enable a user to selectively couple the most appropriate process algorithms from applicable models to create an “optimal” model for the desired application. Where existing algorithms are not appropriate, new algorithms can be developed and easily added to the system. This modular approach to model development and application provides a flexible
OCR for page 131
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program method for identifying the most appropriate modeling approaches given a specific set of user needs and constraints. Nearly all of the programmatic objectives relating to the development of the MMS satisfy the NAWQA Cycle II objectives relating to the greater use of models for improving our understanding of the causes and effects of water quality. Integration of water quality-based modules within the MMS is a natural extension to both the MMS and the existing watershed-based water quality models. Furthermore, it may be quite possible to integrate water quality models that are not watershed based into the MMS modeling framework. The committee recommends that NAWQA attempt to merge developments in watershed based water quality models with the current Modular Modeling System, while striving to avoid overparameterized models. Uncertainty Must Be Considered Water quality modeling is inherently uncertain. There is often significant uncertainty associated with the input, output, model structure, and parameters of water quality models. Uncertainty exists regardless of the model structure and regardless of whether the models have a statistical or process-oriented foundation. It is essential that all modeling studies account for the various sources of uncertainty; otherwise the results of such studies can be misunderstood. Statistical methods such as the simple rating curve approach and SPARROW can provide detailed uncertainty statements associated with estimates of long-term loads in the form of confidence intervals for predictions and/or standard errors associated with model parameters. As noted above, process-oriented modeling studies often lack sufficient data for parameter estimation; as a consequence, there is little basis for, and experience in, thorough error analysis with these models. The committee recommends that the HST attempt to quantify uncertainty associated with all aspects of its water quality modeling evaluations and that the HST develop a position paper on uncertainty. PROPOSED IMPLEMENTATION APPROACH FOR THE UNDERSTANDING GOAL OF CYCLE II Several practical aspects of conducting scientific research have to be considered in an evaluation of the potential for successful implementation. In the case of NAWQA, these aspects include preliminary procedures such as defining objectives and selecting targeted studies to meet these objectives. Then there is a requirement for coordination of the hydrological and water quality models chosen for application. Sufficient staff, expertise, and financial support are all part of what is needed for successful implementation of a program, especially at the national scale. The degree of attention to each of these aspects will determine the success of Cycle II implementation.
OCR for page 151
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program streams reduces its ability to contribute to the nationwide debate in all regions. Clearly, there are nutrient enrichment problems in urban ecosystems as well, and NAWQA should assess these as well, to the extent possible. Habitat degradation is often argued to be the most significant cause of ecological impairment. Thus, NAWQA should reconsider including studies on the impact of alterations in hydrologic regime and sediment transport in its laterphase Cycle II plans. Although there may not be a specific objective relating to exotic species, data already collected and being analyzed by NAWQA will be valuable for examining questions about their impacts. NAWQA should find ways to encourage this synthesis, perhaps by developing cooperative arrangements with the U.S. Fish and Wildlife Service’s Invasive Species Program. The USGS’s NAWQA program is in an excellent position to make a meaningful contribution to the debate on which biological indices are most meaningful based on results from Cycle I studies in which it has conducted bioassessments based on fish, benthic invertebrate, and algal assemblages. This should be a top priority of the Ecological Synthesis Team. Further, the information developed on which indices are most informative should be employed to suggest cost-effective ways to use biological monitoring in Cycle II. In addition to evaluating different indices, it is critical that the Ecological Synthesis Team explore quantitative relationships and potential threshold responses between biotic indices and other measures of water quality. Extrapolation and Forecasting Geographic extrapolation of analyses from NAWQA study units to unmeasured areas and forecasting future water conditions (e.g., based on land management changes) is an essential feature of NAWQA. In the Cycle I national synthesis reports (see also Chapter 6), scientific inferences tend to be expressed in terms of “percent of the samples,” which in effect restricts the geographic coverage to NAWQA watersheds. Strictly speaking, this is incompatible with the initial NAWQA program goal “to describe the status and trends in the quality of the nation’s ground- and surface-water resources,” but it is nonetheless a prudent strategy given the limited basis for extrapolation to unsampled areas (representativeness issues that are fundamental to extrapolation considerations are also discussed in Chapters 2 and 4). It is noteworthy that the Cycle II goals have been scaled back to describing “water-quality conditions for a large part of the nation’s water resources” (emphasis added), while three Cycle II understanding objectives (U15 to U17; see below) explicitly address extrapolation and forecasting. These are both improvements to NAWQA; the first acknowledges the limitations in coverage of the NAWQA study units, and the second targets a scientific basis for inferences beyond the study unit site samples.
OCR for page 152
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program Objective U15: Evaluation of empirical extrapolation and forecasting models: Develop, evaluate, and improve empirical models for spatial extrapolation and forecasting using statistically based methods such as regression analysis. Objective U16: Evaluation of deterministic extrapolation and forecasting models: Systematically evaluate and test selected existing simulation models for their potential value in extrapolation and forecasting of water quality in streams and groundwater. Objective U17: Application of extrapolation and forecasting models: Apply the most appropriate empirical and deterministic models to specific extrapolation and forecasting objectives. The three understanding objectives for Cycle II address extrapolation to unmeasured geographic areas and future forecasting. Based on the brief statement and description of objectives in the NIT guidance (Gilliom et al., 2000; see Appendix A), empirical and mechanistic (simulation) models are expected to play a key role in this extrapolation and forecasting theme. Beyond that, no additional details are presented in the Cycle II implementation guidelines to describe how this will be accomplished, except that consideration will be given to “knowledge of land use and contaminant sources, natural characteristics of the land and hydrologic system, and our understanding of governing processes.” Despite the absence of details on extrapolation and forecasting planned for Cycle II, several papers and reports have been prepared by USGS-NAWQA scientists that provide an indication of related techniques that may be applied in Cycle II. For example, Nolan et al. (1997) mapped the United States based on nitrate input risk groups to statistically assess (using boxplots and analysis-of-variance techniques) the relationship between risk group and groundwater nitrate concentration (based on NAWQA measurements). Black et al. (2000), Rupert (1998), and Tesoriero and Voss (1997) all used logistic regression to develop statistical models to predict the probability of contaminant (nitrate and pesticides) concentration exceedances in surface and groundwater from land use and various land features. These articles indicate a consistent strategy to develop statistical relationships between NAWQA water quality data and predictor variables that are widely measured and reported. This is a prudent strategy. However, NAWQA scientists have recognized the importance of technical support for modeling and hydrologic system analysis with plans to establish the HST (Gilliom et al., 2000). In a similar manner, NAWQA should establish a statistical model support team (perhaps as a subgroup within HST) to provide guidance on the selection and application of these increasingly important modeling tools. This should be an easy task for USGS to support, since some of the best hydrologic statisticians in the United States already are employed by the USGS. As noted in Chapter 4, USGS has long been a leader in “regionalization” and extrapolation, of streamflow data in particular. The statistical model support group should build on this experience to
OCR for page 153
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program adapt methods for water quality assessments. Among the issues that this proposed group could advise on are (1) conditions permitting extrapolation of statistical model inferences to unsampled areas, (2) alternative regression and multivariate methods such as smoothing and robust estimation, (3) consequences of extensive model specification searching, and (4) statistical support for causal conclusions and application of emerging methods in causality assessment (e.g., Spirtes et al., 2001). These specifications parallel the prior recommendations for a focus on parsimonious and adaptive models; including Bayesian and Kalman filtering analysis. Recommendation NAWQA should establish a statistical model support team (perhaps as a subgroup within the HST) to provide guidance on the selection and application of the increasingly important modeling tools. This group can build on the USGS’s historical strength and experience in regionalization and extrapolation of streamflow. Also the group should explore and advise on various issues, including more parsimonious and adaptive models, new techniques (e.g., Bayesian), and previously discussed issues such as uncertainty assessment. CONCLUSIONS AND RECOMMENDATIONS The application of water quality models is an essential activity to achieve the understanding goal of Cycle II NAWQA. The status and trends networks provide the basis for modeling contaminant sources and their transport, and therefore, the proposed application of models in Cycle II will rely heavily on the knowledge gained from Cycle I. Modeling of contaminant sources and their transport to water resources, and hence an ability to predict future conditions, are key to the development of efficient management and policies to protect the beneficial uses of the nation’s water resources, including drinking water and viable ecosystems. Therefore, the application of water quality models is the cornerstone of water resource management for the future, and this was clearly recognized by the NRC as early as 1990 when model application was recommended for NAWQA. Several types of models are proposed for application in Cycle II. The committee presents an overview of models, including those categorized as conceptual, mass balance, statistical regression, process based (mechanistic), and hybrids of these. The systematic arrangement of these models in a hierarchy of spatial and temporal scales can be linked to, and will advance the development of, the Modular Modeling System. When more complex models are applied, the focus should be on those that can be parameterized with data. Once models are applied, the quantification of uncertainty is important to provide perspective for interpretation of model results; therefore, written guidance to standardize the documentation of uncertainty should be developed.
OCR for page 154
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program Successful implementation of model application will require detailed planning to accommodate the practical aspects of the Cycle II implementation strategy. The practical aspects to be dealt with include a clear statement of objectives, coordination of hydrological and water quality models, designation of appropriate staff resources, and optimization of financial support. The six themes with 17 objectives that describe the understanding goal for Cycle II were developed (in the preliminary planning) by categorization of the wide variety of scientific studies proposed by the Cycle II study units. This is the first step needed to determine where commonalties or synergies exist between studies on a regional or national scale. From this, topics are to be selected for targeted studies to meet the understanding goal, but the selection process has not yet been clearly defined. A greater distillation of this collection of ideas is needed since it does not appear that the objectives have been sufficiently refined and focused. The final section of the NIT guidance (Gilliom et al., 2000) presents an example for targeted study development by an Agricultural Chemical Source and Transport Team. This is a strong example that included a set of more clearly defined hypotheses for study. NAWQA should build on this example to further define and clarify study development. It also may be useful to relate the proposed studies and their components to the conceptual source-transport-receptor model presented in the NIT guidance. The STR could be used to organize a wide range of studies from across the nation to show how they relate to each other and where information may be lacking. It is important that the targeted studies are interpreted in the context of NAWQA’s themes and related objectives for Cycle II so they may be evaluated in terms of geographic applicability and national priority. At this important juncture in the development of NAWQA, the committee concludes that the USGS has several major opportunities to advance scientific understanding of factors that affect water quality conditions. As noted throughout this report, the USGS as an organization is uniquely qualified to pursue this national monitoring work. No other federal agency has the organization and infrastructure to collect, analyze, and report on the water resources of the nation with the long-term consistency critical to the success of this program. However, the committee is concerned whether sufficient staff, resources, and expertise are allocated to ensure that modeling implementation can be accomplished and that targeted studies can be adequately developed. Although the NIT report (Gilliom et al., 2000) recognizes the resource problem, its resolution will ultimately govern what can be accomplished in Cycle II. Given these general conclusions from the review of NAWQA plans for the understanding goal in Cycle II, there are many specific recommendations that the committee offers; these are listed below in order of the chapter sections.
OCR for page 155
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program Recommendations for Modeling Approach Conducting mass balances on constituents of concern is a worthy effort in Cycle II. It is strongly recommended that at least a conceptual mass balance be developed for nonpoint source pollutants studied in each study unit. The committee recommends, as discussed in Chapters 2 and 3, that although lakes and reservoirs are not a prominent focus of NAWQA, there are opportunities to use a mass-balance approach to examine and compare their behavior. NAWQA should use the data collected to summarize retention and residence time for key contaminants. This will contribute to understanding system behavior and provide data to stimulate further studies of controlling factors of water quality. The USGS and NAWQA should continue research on improving rating-curve and other statistically based water quality models. Future improvements are likely to come from the integration of statistical and physical process-oriented models. Research should focus on physical reasoning and physical process interpretations of the concentration-discharge relationship of the type introduced by O’Connor (1976) and Duffy and Cusumano (1998). Early applications of SPARROW (e.g., to the Chesapeake Bay; Preston and Brakebill, 1999) have proven quite promising. The committee encourages NAWQA to continue research relating to the improvement and application of SPARROW. NAWQA-USGS should focus on simple, parsimonious process models (i.e., models that are not overparameterized) where parameter estimation and mechanistic expressions can relate to available data. Further, consistent with the recent NRC study of the scientific basis of EPA’s TMDL program (NRC, 2001), NAWQA should encourage that such models be usable in an adaptive framework. This includes techniques such as Bayesian analysis, Kalman filtering, and data assimilation that allow model forecasts to be updated and improved over time with monitoring data. All future NAWQA studies should attempt to evaluate the uncertainty associated with all aspects of their water quality modeling evaluations, and the Hydrologic Systems Team should develop a guidance document on this topic. Recommendations for Implementation The 17 objectives for the understanding goal, as presented in the NIT report, are complex and address too many issues to have a good focus. Further, the process of selecting topics for targeted studies to meet the understanding goal in Cycle II has not yet been clearly defined. The example of the ASTT, however, was well done, presenting clearly defined hypotheses and a focus lacking in other areas. The committee recommends that NAWQA build on this example to develop and refine its targeted study approach.
OCR for page 156
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program NAWQA must resolve staffing, expertise, and resource issues if the objectives of the understanding goal are to be successfully addressed in Cycle II. It is unclear what portion of the objectives can be met at the level of resources suggested in planning documents and whether sufficient staff expertise is available for model application and implementation. Recommendations on Understanding Themes and Objectives Sources of Contaminants NAWQA should consider evaluation of the variability of sources and contaminant occurrence as a first step for the objectives related to “understanding contaminant sources”; this should take place prior to subsequent space-for-time and targeted study design. NAWQA should also consider collaboration with other agencies that may have pertinent data to evaluate statistical properties of contaminant variability. Sampling design should include interpolation and indexing considerations that may be necessary to develop as complete a mass balance as possible. For the identification of sources, when feasible, monitoring programs should employ highly specific or supplementary analyses that permit distinction between different sources of the same contaminant (e.g., coliphage or ribotyping of microbiological specimens, isotopic analyses of some chemicals). Some sampling strategies may require higher resolution to assess some contaminant sources. Sample data should be stratified between stormflow and baseflow for the development of relationships between land use and stream concentrations for many contaminants. Contaminant Fate and Transport The HST should ensure some uniformity (and/or compatibility) in the use of models and software packages in Cycle II to enable national comparisons and aggregation of data and results. As previously discussed, to test research hypotheses the HST must also address error terms and uncertainty in model selection and application. The current NAWQA program is not geared to the assessment of ephemeral streamflows that can be an important contaminant source in arid areas in particular. NAWQA should develop strategies to sample ephemeral streamflows, especially in perceived high-risk areas, such as those in which contamination could threaten perennial streams, lakes or reservoirs, or shallow groundwater systems.
OCR for page 157
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program Groundwater-Surface Water Interactions The hyporheic zone, or groundwater-surface water ecotone, is an important component in understanding surface water quality and near-surface groundwater quality. All appropriate Cycle II study units should endeavor to (1) design a process-based approach to characterize GW-SW interactions and their effects on water quality and (2) use process-based models that can include GW-SW interaction components to delineate the spatial and temporal variations in GW-SW interchange and the concomitant water quality changes. This may mean seeking out collaborators and cooperators to find the needed expertise and study sites, and this should be encouraged as well. Effects on Stream Ecosystems and Assessments of Biological Integrity It would be inadvisable for NAWQA to embark upon a major ecotoxicology program that would overlap with the USGS’s Biological Resources Division program. Thus, it is critical that Objective U12 be clarified to better articulate proposed studies so that potential collaborative relationships with other programs can be more clearly defined. Focusing NAWQA’s attention only on nutrient enrichment in agricultural streams reduces its ability to contribute to the nationwide debate in all regions. Clearly, there are nutrient enrichment problems in urban ecosystems as well, and NAWQA should assess these to the extent possible. Habitat degradation is often argued to be the most significant cause of ecological impairment. Thus, NAWQA should reconsider including studies on the impact of alterations in hydrologic regime and sediment transport in its later-phase Cycle II plans. Although there may not be a specific objective relating to exotic species, data already collected and being analyzed by NAWQA will be valuable for examining questions about their impacts. NAWQA should find ways to encourage this synthesis, perhaps by developing cooperative arrangements with the U.S. Fish and Wildlife Service’s Invasive Species Program, for example. The USGS’s NAWQA program is in excellent position to make a meaningful contribution to the debate on which biological indices are most meaningful based on results from Cycle I studies in which it has conducted bioassessments based on fish, benthic invertebrate, and algal assemblages. This should be a top priority of the Ecological Synthesis Team. Further, the information developed on which indices are most informative should be employed to suggest cost-effective ways to use biological monitoring in Cycle II. In addition to evaluating different indices, it is critical that the Ecological Synthesis Team explore quantitative relationships and potential threshold responses between biotic indices and other measures of water quality.
OCR for page 158
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program Extrapolation and Forecasting NAWQA should establish a statistical model support team (perhaps as a subgroup within the HST) to provide guidance on the selection and application of the increasingly important modeling tools. This group can build on the USGS’s historical strength and experience in regionalization and extrapolation of streamflow. Also the group should explore and advise on various issues, including more parsimonious and adaptive models, new techniques (e.g., Bayesian), and previously discussed issues such as uncertainty assessment. REFERENCES Alexander, R. B., R. A. Smith, and G. E. Schwarz. 2000. Effects of stream channel size on the delivery of nitrogen to the Gulf of Mexico. Nature 403:758-761. Bachman, L. J., and P. J. Phillips. 1996. Hydrologic landscapes on the Delmarva Peninsula, Part 2, Estimates of base-flow nitrogen load to Chesapeake Bay. Journal of American Water Resources Association 32(4):779-791. Bencala, K. E., and R. A. Walters. 1983. Simulation of solute transport in mountain pool-and-riffle stream: A transient storage model. Water Resources Research 19:718-724. Beven, K. J. 1997. Distributed Hydrological Modelling: Applications of the TOPMODEL Concept. New York: Wiley & Sons. Black, R. W., A. L. Haggland, and F. D. Voss. 2000. Predicting the probability of detecting organochlorine pesticides and polychlorinated biphenyls in stream systems on the basis of land use in the Pacific Northwest, USA. Environmental Toxicology and Chemistry 19:1044-1054. Bohlke, J. K., and J. M. Denver. 1995. Combined use of groundwater dating, chemical and isotopic analyses to resolve the history and fate of nitrate contamination in two agricultural watersheds, Atlantic coastal plain, Maryland. Water Resources Research 31(9):2319-2339. Borsuk, M. E., D. Higdon, C. A. Stow, and K. H. Reckhow. 2001. A Bayesian hierarchical model to predict benthic oxygen demand from organic matter loading in estuaries and coastal zones. Ecological Modelling 143:165-181. Boyer, E. W., G. M. Hornberger, K. E. Bencala, and D. M. McKnight. 1996. Overview of a simple model describing variation of dissolved organic carbon in an upland catchment. Ecological Modelling 86:183-188. Cohn, T. A. 1995. Recent advances in statistical methods for the estimation of sediment and nutrient transport in rivers. In Vogel, R. M. (ed.) U.S. National Report 1991-1994—Contributions in Hydrology. Washington, D.C.: American Geophysical Union. Available online at http://www.agu.org/revgeophys/cohn01/cohn01.html. Cuffney, T. F., M. R. Meador, S. D. Porter, and M. E. Gurtz. 1997. Distribution of fish, benthic invertebrate and algal communities in relation to physical and chemical conditions, Yakima River basin, Washington, 1990. U.S. Geological Survey Water-Resources Investigations Report 96-4280. Raleigh, N.C.: U.S. Geological Survey. Dillon, P. J., and F. H. Rigler. 1974. A test of a simple nutrient budget model predicting the phosphorus concentration in lake water. Journal of the Fisheries Research Board of Canada 31:1771-1778. Duffy, C. J., and J. Cusumano. 1998. A low-dimensional model for concentration-discharge dynamics in groundwater stream systems. Water Resources Research 34(9):2235-2247. EPA (U. S. Environmental Protection Agency). 1997. New York City Watershed: Filtration Avoidance Determination. Mid-Course Review. Available online at http://www.epa.gov/r02earth/water/nycshed/fadrev2.htm.
OCR for page 159
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program Gibert, J., M. J. Dole-Olivier, P. Marmonier, and P. Vervier. 1990. Surface water/groundwater ecotones. Pp. 199-225 in Naiman, R. J., and H. Decamps (eds.) Ecology and Management of Aquatic-Terrestrial Ecotones. Park Ridge, N.J.: The Parthenon Publishing Group. Gilliom, R. J. 2000a. Hydrologic systems analysis and modeling in Cycle II of NAWQA. Draft memo, submitted to the NRC committee, October 3. Gilliom, R. J. 2000b. Preliminary draft of themes and objectives for Cycle II investigations. Draft memo, submitted to NRC committee, April 19. Gilliom, R. J., K. Bencala, W. Bryant, C. A. Couch, N. M. Dubrovsky, L. Franke, D. Helsel, I. James, W. W. Lapham, D. Mueller, J. Stoner, M. A. Sylvester, W. G. Wilber, D. M. Wolock, and J. Zogorski. 2000. Study-Unit Design Guidelines for Cycle II of the National Water Quality Assessment (NAWQA). U.S. Geological Survey NAWQA Cycle II Implementation Team. Draft for internal review (11/22/2000). Sacramento, Calif.: U.S. Geological Survey. Harwood, A. K. 1995. The Urban Stormwater Contribution of Dissolved Trace Metals from the North Floodway Channel, Albuquerque, NM, to the Rio Grande . Professional project report, Master of Water Resources Administration Degree, University of New Mexico. Hawkins, C. P., R. H. Norris, J. N. Hogue, and J. W. Feminella. 2000. Development and evaluation of predictive models for measuring the biological integrity of streams. Ecological Applications 10:1456-1477. Hinkle, S., J. Duff, F. Triska, A. Laenen, E. Gates, K. Bencala, D. Wentz, and S. Silva. 2001. Linking hyporheic flow and nitrogen cycling near a large river in the Willamette Basin, Oregon. Journal of Hydrology 244(3-4):157-180. Hirsch, R. M., W. M. Alley, and W. G. Wilber. 1988. Concepts for a National Water-Quality Assessment Program. U.S. Geological Survey Circular 1021. Denver, Colo.: U.S. Geological Survey. Hornberger, G. M., K. E. Bencala and D. M. McKnight. 1994. Hydrological controls on dissolved organic carbon during snowmelt in the Snake River near Montezuma, Colorado. Biogeochemistry 25:147-165. Karr, J. R. 1981. Assessment of biotic integrity using fish communities. Fisheries 6:21-27. Karr, J. R., and E. W. Chu. 1999. Restoring Life in Running Waters. Washington, D.C.: Island Press. Kerans, B. L., and J. R. Karr. 1994. A benthic index of biotic integrity (B-IBI) for rivers of the Tennessee Valley. Ecological Applications 4:768-785. Leavesley, G. H. 1997. The Modular Modeling System (MMS)—A modeling framework for multidisciplinary research and operational applications. Pp. 27-31 in Freeman, G. E., and A. G. Frazier (eds.) Proceedings of the Scientific Assessment and Strategy Team Workshop on Hydrology, Ecology, and Hydraulics. Volume 5 of Kelmelis, J. A. (ed.) Science for Floodplain Management into the 21st Century. Washington, D.C.: U.S. Government Printing Office. Lenat, D. R. 1993. A biotic index for the southeastern United States: Derivation and list of tolerance values, with criteria for assigning water quality ratings. Journal of the North American Benthological Society 12:279-290. Mallard, G. E., J. T. Armbruster, R. E. Broshears, E. J. Evenson, S. N. Luoma, P. J. Phillips, and K. R. Prince. 1999. Recommendations for Cycle II of the National Water Quality Assessment (NAWQA) Program. U.S. Geological Survey NAWQA Planning Team. U.S. Geological Survey Open-File Report 99-470. Reston, Va.: U.S. Geological Survey. McMahon, P. B., and J. K. Bohlke. 1996. Denitrification and mixing in a stream-aquifer system: Effects on nitrate loading to surface water. Journal of Hydrology 186:105-128. McMahon, G., and M. D. Woodside. 1997. Nutrient mass balance for the Albemarle-Pamlico Drainage Basin, North Carolina and Virginia. Journal of American Water Resources Association 33(3):73-589. Naiman, R. J., J. J. Magnuson, D. M. McKnight, and J. A. Stanford. 1995. The Freshwater Imperative. Washington, D.C.: Island Press.
OCR for page 160
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program Nolan, B. T. 1998. Modeling approaches for assessing the risk of nonpoint-source contamination of ground water. U.S. Geological Survey Open-File Report 98-531. Denver, Colo.: U.S. Geological Survey. Nolan, B. T., B. C. Ruddy, K. J. Hitt, and D. R. Helsel. 1997. Risk of nitrate in groundwaters of the United States–A national perspective. Environmental Science and Technology 31:2229-2236. NRC (National Research Council). 1990. A Review of the USGS National Water Quality Assessment Pilot Program. Washington, D.C.: National Academy Press. NRC. 1994. National Water Quality Assessment Program: The Challenge of National Synthesis. Washington, D.C.: National Academy Press. NRC. 2001. Assessing the TMDL Approach to Water Quality Management. Washington, D.C.: National Academy Press. O’Connor, D. J. 1976. The concentration of dissolved solids and river flow. Water Resources Research 12(2):279-294. Poff, N. L., J. D. Allan, M. B. Bain, J. R. Karr, K. L. Prestegaard, B. D. Richter, R. E. Sparks, and J. C. Stromberg. 1997. The natural flow regime. Bioscience 47:769-784. Preston, S. D., and J. W. Brakebill. 1999. Application of Spatially Referenced Regression Modeling for the Evaluation of Total Nitrogen Loading in the Chesapeake Bay Watershed. U.S. Geological Survey Water-Resources Investigations Report 99-4054. Reston, Va.: U.S. Geological Survey. Puckett, L. J. 1999. Estimation of nitrate contamination of an agro-ecosystem outwash aquifer using a nitrogen mass-balance budget. Journal of Environmental Quality 28(6):2015-2025. Reckhow, K. H., M. N. Beaulac, and J. T. Simpson. 1980. Modeling Phosphorus Loading and Lake Response Under Uncertainty: A Manual and Compilation of Export Coefficients. EPA-440/5-80-011. Washington, D.C.: U.S. Environmental Protection Agency. Richter, B. D., J. V. Baumgartner, J. Powell, and D. P. Braun. 1996. A method for assessing hydrologic alteration within ecosystems. Conservation Biology 10:1163-1174. Rupert, M. G. 1998. Probability of Detecting Atrazine and Elevated Concentrations of Nitrate (NO2+NO3-N) in Ground Water in the Idaho Part of the Upper Snake River Basin. U.S. Geological Survey Water-Resources Investigations Report 98-4203. Boise, Idaho: U.S. Geological Survey. Smith, R. A., G. E. Schwarz, and R. B. Alexander. 1997. Regional interpretation of water-quality monitoring data. Water Resources Research 33(12): 2781-2798. Spirtes, P., C. Glymour, and R. Scheines. 2001. Causation, prediction, and search. In Adaptive Computation and Machine Learning (book series). 2nd edition. Cambridge, Mass.: MIT Press. Squillace, P. J., E. M. Thurman, and E. T. Furlong. 1993. Groundwater as a nonpoint source of atrazine and diethylatrazine in a river during base flow conditions. Water Resources Research 29(6):1719-1729. Stalnaker, C., B. L. Lamb, J. Henricksen, K. Bovee, and J. Batholow. 1995. The Instream Flow Incremental Methodology: A Primer for IFIM. Biological Report No. 29. Fort Collins, Colo.: U.S. Department of Interior, National Biological Service. Tesoriero, A. J., and F. D. Voss. 1997. Predicting the probability of elevated nitrate concentrations in the Puget Sound Basin: Implications for aquifer susceptibility and vulnerability. Ground Water 35:1029-1039. Triska, F. J., V. C. Kennedy, R. J. Avanzino, G. Zellweger, and K. E. Bencala. 1989. Retention and transport of nutrients in the third-order stream in northwestern California: Hyporheic processes. Ecology 70:1893-1905. Vervier, P., J. Gilbert, P. Marmonier, and M. J. Dole-Olivier. 1992. A perspective on the permeability of the surface freshwater-groundwater ecotone. Journal of the North American Benthological Society 11:93-102. Vollenweider, R. A.. 1975. Input-output models with special reference to the phosphorus loading concept in limnology. Schweitz. Z. Hydrol. 37(1):53-84.
OCR for page 161
Opportunities to Improve the U.S. Geological Survey National Water Quality Assessment Program Vollenweider, R. A.. 1976. Advances in defining critical loading levels for phosphorus in lake eutrophication. Memorie dell’Istituto Italiano di Idrobiologia 33:53-83. Winter, T. C., J. W. Harvey, O. L. Franke, and W. M. Alley. 1998. Ground Water and Surface Water: A Single Resource. U.S. Geological Survey Circular 1139. Denver, Colo.: U.S. Geological Survey. Wright, J. F. 1995. Development and use of a system for predicting the macroinvertebrate fauna in flowing waters. Australian Journal of Ecology 20:181-197.
Representative terms from entire chapter: