Monitoring and Data Management for USGS River Science
The USGS and other agencies maintain the river monitoring and data management infrastructure that supports existing activities in river science and management. However, this infrastructure is woefully inadequate to address the immense and growing problems related to the deterioration of the nation’s rivers. The lack of sufficient data for river systems is one of the biggest obstacles to providing the science-based information needed to effectively manage the nation’s rivers.
Although Chapter 4 identifies five science priority areas where the USGS can contribute to a national effort on river science, how the USGS is able to address these priorities is predicated on new river monitoring and data management efforts that fill science data gaps in critical and neglected areas. Chapter 5 describes activities that must build on multidisciplinary databases created from national monitoring networks and intensive local sampling, as well as other ancillary geophysical datasets, for (1) a survey and synthesis of river features and (2) the development of models of river processes. For science priorities in focused topic areas, Chapter 5 identifies data gaps in (1) ecological monitoring baselines for adaptive management and the establishment of environmental flows, (2) water and sediment flux monitoring for understanding changes in sediment transport and river geomorphology in response to dam operations, land-use changes or river restoration activities, and (3) monitoring the exchange of water and chemicals between groundwater sources and rivers and their effects on characteristics of floodplain and riparian areas. Therefore, new USGS initiatives to enhance and strengthen data collection, archiving, and dissemination are an essential element of a USGS river science initiative.
Expanded river monitoring and data collection activities at the USGS also
need to provide fundamental long-term baseline monitoring in anticipation of the nation’s need for river science information. Although monitoring is usually designed to address specific problems, the value of a consistent national baseline monitoring approach to address emerging problems is frequently overlooked and undervalued. Understanding the impacts of agricultural and urban land-use changes on rivers and the influence of climatic variations on biogeochemical and water cycles are just two examples of the unforeseen usages of long-term streamflow observations from USGS streamgages. Another example is the integrated study of hypoxia in the Gulf of Mexico (Goolsby et al., 1999), which has been instrumental in promoting a scientific understanding of the sources and fluxes of nutrients responsible for this problem. The report builds on baseline streamflow and nutrient monitoring by the USGS. Unfortunately, the degradation of the nation’s baseline monitoring of rivers over time (NRC, 2004d) threatens our ability to assess emerging problems in river science.
In the first part of this chapter, we describe opportunities for the USGS to build on its data collection infrastructure and expand its monitoring of hydrologic, geomorphic, chemical, biological, and ecological processes in river and floodplain ecosystems for national and regional science synthesis. We focus on enhancements in streamflow, biological, and sediment monitoring and on the establishment of a reach-scale monitoring approach. This section also describes some considerations for the general design principles of a modern river monitoring system, highlighting the importance of partnering monitoring efforts with other organizations and incorporating measurement technologies.
The value of these enhancements to river monitoring activities to river science depends on easy access to data, and the ability to efficiently utilize diverse measurements and data products from multiple disciplines, by the community of scientists and decision makers. In the second part of this chapter we discuss the data management challenges for river science, and recommend an informatics component for integrated data archiving, dissemination, and management.
One of the fundamental implementation challenges for a nationally relevant river science program is to leverage data resources to avoid duplication and target data collection activities to support the portfolio of data needs and uses. Although the focus of this chapter is on USGS activities in monitoring and data management for river science, coordination and cooperation among the federal resource management agencies and their nonfederal partners is imperative because of the scope, scale, and intensity of data needed to support river science. Plans for interagency collaboration need to be an integral part of any USGS river science monitoring and data archiving activity. No single federal agency can collect, quality assure, manage, and disseminate all data and observations relevant for river science. Yet all federal agencies, nonfederal partners, and stakeholders with an interest in river science and resource management will benefit from access and availability of accurate, reliable, and well-documented data.
INTEGRATED DATA COLLECTION AND RIVER MONITORING
Scientific data related to riverine systems, and the knowledge generated using these data, are needed to support decision making for many of today’s policy and management challenges related to rivers (see Chapter 2). Whether the problem is to establish policies to prevent river degradation from upstream urban land-use changes, to prioritize investments in river restoration, or to decide whether to remove a dam, scientific data informs the decision-making process. But, there are critical gaps in river monitoring. Making river policy and management decisions with limited or insufficient data leads to costly and potentially irreversible mistakes. Additionally, in some cases, the absence of data fuels controversy, delaying or preventing policy makers from making decisions. Where decisions are made but their impacts on rivers not closely monitored, science and the public never learn the valuable lessons from successes and failures that would improve future policy and management decisions.
Monitoring of our nation’s rivers is the foundation for the USGS’s contribution to river science. By expanding monitoring activities on rivers, the USGS can develop a modern, coordinated 21st-century river monitoring system. The USGS streamgaging network, with its quality-controlled and publicly accessible data archive, is the fundamental building block for such a system. Still, a river monitoring system to support integrative river science and river management must encompass not only streamgaging but also the chemical, biological, and sedimentological characterization of river flows. It would merge other components, such as a national sediment monitoring program and biological reach indicators, as integral elements of a coordinated system. It would include the mapping of the physical and ecological conditions of rivers and riparian areas. Finally, it would establish protocols for data collection and dissemination, and investigate new river monitoring technologies.
Recommendation: The USGS should expand its monitoring activities on rivers to better incorporate river physical, chemical, and biological conditions within its existing river and streamflow monitoring programs. Its goals should include development of a 21st-century river monitoring system for data collection, transmission, and dissemination.
Data Collection Needs for River Science
Data gaps exist in all priority research areas in river science (see Chapter 4). Monitoring baseline conditions is essential to establish environmental flows that support ecological functions. Synthesizing baseline information provides guidance for river restoration and adaptive management for degraded rivers, while scientifically designed monitoring of the outcomes of these activities is needed
to test the underlying hypotheses guiding decision making. These baselines also help in understanding riparian areas that are also affected by changes in flow, sediment, and nutrient compositions; developing a predictive understanding of riparian ecosystems requires data on the physical system, the vegetation, and the interactions of surface and groundwater in riparian corridors.
Relatively speaking, streamflow data are generally available for larger river systems, but coincident observations of water quality, sediment transport, biological indicators, and riparian ecosystems, are usually lacking. For smaller river systems, there are few integrated river datasets. Therefore, continuously monitoring key observations that address these issues for extended periods in larger streams is, at a minimum, a prerequisite for synthesizing this information to generate estimates at smaller unmonitored sites.
For integrative scientific investigation of river and riverine processes, the biggest challenge in data collection is planning and implementing a cost-effective, yet scientifically sound, mix of hydrologic, geomorphic, chemical, biological, and ecological monitoring. For example, addressing certain science questions in river restoration may well require data on river and riparian ecology, hydrologic conditions, sediment concentrations and fluxes, watershed conditions, and economic variables (Box 5-1 and Appendix B). To be effective, integrated monitoring networks need to be designed based on scientific principles, utilize documented data protocols, and produce quality-assured/quality-controlled (QA/QC) datasets.
Such measurements often cannot be done at a single site. The relevant physical processes often act at different characteristic scales. For river geomorphology, some characteristic scales of length include the channel width, which responds to flow and sediment transport regimes; the meander wavelength, which is related to geologic and geomorphic processes within the river valley; and the river network link length, which is related to landscape evolution and river network development processes. Pool spacing, which has important influences on stream habitat, is related to the width of rivers, meander wavelengths, and the length of stream segments between junctions relative to drainage density and average hillslope length (Leopold et al., 1964; Gregory and Walling, 1973; Rodriguez-Iturbe and Rinaldo, 1997).
The connection between processes and characteristic river scales is also evident in the River Continuum Concept (Vannote et al., 1980), which relates stream size and river order to the structural and functional attributes of river biology. Matching the spatial and temporal scales of physical measurements with the interactions between hydrologic, chemical, biological, and ecological processes is a significant challenge. Still, meeting this challenge is necessary to understand the complex interactions, and to develop predictive models that couple them.
Valuable Data Resources for Answering River Restoration Science Questions
Representative Research Questions
How can modifying hydrology and geomorphology restore riverine ecology?
How can the ecological success of restoration be assessed?
What might be the cumulative impacts on restoration of multiple segments of rivers?
What are hydrologic or chemical ecosystem thresholds at multiple scales?
Riparian and River Ecological Data
Baseline ecological data
Algal distribution identification
Micromet (fluxes and isotopes)
Surface flow velocity distribution
Chemical composition of waters
Age dating and natural tracers
Precipitation and meteorological data
Sediment deposition/erosion pattern
Sediment size distribution
Watershed Physical Data
Soils types and quality
See Appendix B for a more comprehensive set of research questions and the types of data required to address them.
The USGS Role in Monitoring for River Science
The USGS has a federal mandate to serve the nation by providing reliable scientific data and information. Its monitoring activities are distinguished for their scientific rigor and quality control. The USGS has also been a leader in data distribution and in the development of instrumentation for river monitoring.
The core river monitoring capabilities of the USGS already serve as our nation’s primary data resource in many areas of river science. The most visible example is the streamgaging network (see Chapter 3), with over 7000 active sites. At selected sites, the USGS also continuously records water quality (including pH, specific conductance, temperature, and dissolved oxygen) and suspended sediment concentrations. These data are archived and disseminated online through the National Water Information System (NWIS) (http://waterdata.usgs.gov/nwis/). Through satellite telemetry systems, NWIS provides real-time access to streamflow data at most sites, and water-quality data at some.
USGS expertise in data collection and river monitoring is incorporated in the design and implementation of its research programs. For example, the National Water-Quality Assessment (NAWQA) Program (see Chapter 3) has an experimental design that schedules intensive sampling periods to monitor trends in water quality and to help identify the reasons for these trends. The USGS uses standard data collection protocols to provide a nationally consistent water-quality dataset. One example of a current program-specific monitoring effort is the Long Term Resource Monitoring Program (LTRMP) for the upper Mississippi River (see Chapter 3). In this program, the USGS samples fish, macroinvertebrates, vegetation, and water quality from six field stations, and combines these data with other data (e.g., bathymetry, fish passage information at locks and dams, water levels, and discharge), to provide information to assess the impacts of river management decisions on biological resources, and to support development of future alternatives.
Still, the USGS must build on its existing capabilities to more comprehensively enhance its river monitoring to address the nation’s need for science information on rivers. In the following subsections, we describe recommended enhancements in river monitoring for a USGS river science initiative, and describe some considerations for the design of a modern river monitoring system.
Enhancements in Streamflow Monitoring
Recommendation: The USGS should investigate cost-effective opportunities to augment site information and, in some cases, increase the sampling at targeted National Streamflow Information Program streamgages to make the gage data more useful for river science initiatives. There should be a renewed effort to collect, archive, and disseminate opportunistic data for hydrologic extremes (floods and droughts).
The USGS streamgaging network is one of the most important resources available for river science and management. The National Streamflow Information Program (NSIP) needs to be fully implemented and maintained for long-term archival data to assess trends driven by anthropogenic landscape changes and future climate change. There are some efforts that the USGS can take within this infrastructure of data acquisition that would require only modest additional resources but would maximize the usefulness of NSIP streamgages for river science. These were highlighted in the NRC report on NSIP (NRC, 2004d), but they deserve to be summarized here.
First, the USGS needs to archive and disseminate previously and recently collected cross-section and unit value (hydrograph) data along with the streamflow data, which are readily accessible as daily averages and flood peaks. Both fine resolution streamflow time series and cross-section data are essential to document channel changes, evaluate nonstationary hydrograph characteristics, assess the duration and frequency of floodplain inundation, and infer hydroecological relationships. Rescuing cross section and unit value data at both carefully selected index gages and at gages where the cumulative impacts of natural and anthropogenic changes can be critically analyzed would fill crucial data gaps.
The USGS should investigate cost-effective opportunities to augment site information and, in some cases, increase the sampling at targeted NSIP streamgages to make the gage data more useful for river science initiatives. For example, expanding the monitoring of river temperature and water quality (including pathogens and organics) at some NSIP sites would be invaluable with respect to regional and national synthesis. The USGS might also consider developing a plan to collect stream gradient and bed material size at more gaging station locations, perhaps in conjunction with other organizations.
Finally, there should be a renewed effort to collect, archive, and disseminate opportunistic data for hydrologic extremes (floods and droughts). The series of 12 reports after the 1993 upper Mississippi River flood (e.g., Parrett et al., 1993) serves as a prototype to emulate; it demonstrates how opportunistic sampling and timely dissemination of unbiased information can contribute to the scientific and public debate that follows such events. Similar efforts over the course of widespread severe droughts are needed to study their effects on river ecology. For making opportunistic measurements, the USGS should consider the role of new measurement technologies for expanding the gaging network (e.g., to collect additional crest-stage and slope-area data), as noted in NRC (2004d) and alluded to in the “Measurement Technologies” section below.
Sentinel watersheds could serve as stream morphology and sediment reference sites across fundamental landscape divisions. Still, a key question is whether the stretches of rivers where the sentinel streamgages are now located are representative of the watershed and ecoregion of which they are a part. Typically, streamgages are sited on stretches best suited for flow measurement (e.g., access
and clear channel geometry). Using an integrative approach, the siting of sentinel streamgages should be reevaluated to best measure stream processes.
Enhancements in Biological Monitoring
Recommendation: Expanding the collection of biological and ecological data at streamgaging sites is needed to develop integrated biophysical datasets for river science. However, fundamental questions remain on how to implement a monitoring program to support national and regional synthesis. The USGS should continue its efforts to define relevant biological monitoring activities for national implementation, while still expanding biological and ecological monitoring in a targeted fashion to address clearly defined regional data needs.
NSIP provides a strong framework for assessing river hydrology and integrating flow with other measured variables for river regime characterization. However, there are no comparable biological data routinely collected at NSIP gages to characterize the ecological or biological status of our nation’s rivers on the whole. Instead, biological data are collected as part of programs like NAWQA and LTRMP, with sampling designed to track regionally significant biological and ecological indicators.
A key question is whether there are river biological and ecological indicators that need to be surveyed nationwide, or whether most indicators are meaningful only in a regional or local context. There is a clear need to identify such indicators to assess ecological trends in river systems; many agencies and research organizations have research programs designed to develop ecological indicators for both scientific and management applications. For instance, the EPA’s newly completed Wadeable Streams Assessment, which was a snapshot of the ecological condition of small streams throughout the United States using 500 randomly selected sites, used benthic macroinvertebrates as biological indicators, phosphorus, nitrogen, salinity, and acidity as chemical indicators, and streambed sediments, instream fish habitat, riparian vegetative cover, and riparian disturbance as physical condition indicators (http://www.epa.gov/owow/streamsurvey/).
Developing ecological indicators to track changes in rivers and provide early warning on river impairment is a significant component of the EPA’s Environmental Monitoring and Assessment Program (EMAP) as well. At the USGS, the Biological Resources Discipline is wrestling with the question of national indicator monitoring needs as part of its Status and Trends Program (see Chapter 3). These activities, within and external to the USGS, may well lead to a future consensus on a national approach to the monitoring of indicators.
In the meantime, the USGS should continue its support of initiatives such as
the National Biological Information Infrastructure (NBII), and similar biological data gathering in cooperation with state and other federal agencies, as monitoring data assembled by NBII are valuable for a national synthesis of biologic status and trends (USGS, 1998). The USGS National Water-Use Information Program (NWUIP) is an analogy for an initiative that uses multiple sets of data of mixed quality. NWUIP involves multiple partnerships with state agencies, from which water use data of highly variable quality is synthesized at the national level (NRC, 2002b). In particular, NWUIP leverages its district structure to facilitate the sharing of data from state and local water agencies.
On a regional basis, the LTRMP is a prototype for doing science-quality monitoring to meet an experiment design that incorporates river biological components. A current weakness of LTRMP is its limited integration of monitoring to test distinct hypotheses and to develop models and predictive capabilities. Still, the data monitoring and exploration effort exemplified by this program should be emulated elsewhere, if done in an integrated scientific way that includes elements of hydrology, geomorphology, and water quality. Expanding biological monitoring activities in a targeted fashion, using an LTRMP prototype, is an approach that the USGS can readily implement to collect integrated biological and ecological datasets needed for research on USGS river science priorities and site-specific river management problems.
Enhancements in Sediment Monitoring
Recommendation: Leveraging the infrastructure of the streamgaging network, the USGS should greatly expand sediment monitoring of the nation’s rivers. To meet the growing needs for sediment data, the USGS should be a leader in developing a comprehensive national sediment monitoring program.
The decline in sediment monitoring in recent decades from approximately 300 sites in 1980 (Glysson, 1989) coupled with the growing needs for sediment information to address an expanding number of river management issues act together to generate a major data gap in river science understanding (see also Chapter 4, “Sediment Transport and Geomorphology”). This gap is exacerbated by the economic consequences of accelerated erosion and sediment transport; sediment-related damages for North America are estimated to be on the order of $16 billion annually (Osterkamp et al., 1998).
The Subcommittee on Sedimentation (SOS) of the Advisory Committee on Water Information (ACWI) has called for a national monitoring program for sediment investigation (Osterkamp et al., 2004). The proposed program includes a core network with continuous fluvial-sediment monitoring at existing flow and water-quality gages. Other components would address improving supplementary sampling at existing sites, measurement techniques, data synthesis and assess-
ment, and the rescuing of archives of previously collected sediment data. The SOS reported that if the program leads to a 1 percent reduction in sediment-related damages through better river management, the damage reduction would be about 40 times the program cost.
The USGS should take a lead in developing a national sediment monitoring program. Leveraging the infrastructure of the streamgaging network, the USGS should enhance the sediment monitoring at selected NSIP gages. To maximize the utility of the sites for river science, coincident measurement of temperature and other water-quality variables should be made. New measurement technologies hold promise in advancing this effort by both reducing the costs of sediment monitoring and enhancing the information gained. For instance, the use of hydroacoustics to measure three-dimensional flow structure would provide data to better interpret sediment flux measurements. Analysis of the chemical composition of river sediments would be valuable for identifying source areas for fluvial sediments.
Establishment of Reach-Scale Monitoring
Recommendation: An index reach monitoring approach would help address many data needs for USGS river science priorities. To integrate monitoring of physical, chemical, and biological conditions for river science investigations, the USGS should begin efforts to design and implement sampling plans on reach scales.
At the reach scale, the river and floodplain ecosystem is driven by its physical settings and hydrologic conditions, including the stream slope, hydrologic regime, groundwater interactions, and land use. Few monitoring networks adequately measure both the drivers and ecosystem responses to support scientific understanding and model development at the reach scale.
The USGS should consider establishing index biological reaches throughout the nation to support its science priority areas for river science (see Chapter 4). Index reaches would serve as integrated measures of reach-specific responses to changing environmental conditions and support the careful evaluation of responses of biota over time. Index biological reaches would be locations where coupled measurements of river flows, groundwater levels and fluxes, and water quality are made.
As part of a continuous monitoring effort, the USGS also needs to map riparian cover of index reaches on at least an annual or semiannual basis. Remotely sensed data, combined with in situ observations by the USGS and other agencies, could be used to develop riparian cover maps. Reach-level riparian ecological services, such as flood control, bank storage of water, interception of pollutants, and habitat diversity, can be inferred from relatively coarse-grained riparian cover information, hydrologic data, and water-quality data.
The colocation of biological and hydrologic monitoring would help assess how much water is needed to sustain ecological goods and services and help establish the environmental flows needed to meet ecosystem needs. The data collected at reaches would also help in understanding the role of groundwater exchange, the hyporheic zone, and the chemical composition of these waters, on river ecosystems. In turn, index reach monitoring could provide a means for evaluating the effects of river restoration or adaptive management experiments. The reach-scale activities described in Chapter 4 for the “Surveying and Synthesizing” science priority area would be an important first step in designing a network of index reaches that sample across diverse physical and ecological settings.
It should be noted that “reach” as a river concept has a variety of definitions and invokes very different pictures in the minds of river scientists. Within the USGS, at least three programs operate in reaches—the Grand Canyon, the Upper Mississippi (LTRMP sampling reaches), and NAWQA. However, these programs operate at different spatial and temporal scales. The USGS would need to consider which current programs are operating at an appropriate scale for this concept to determine what programs might be good to build off of.
River Monitoring—Design Principles
Recommendation: For the design and implementation of a coordinated river monitoring system, the USGS should develop specific monitoring goals and objectives for building on its existing infrastructure. The USGS should prioritize these activities based on variables with broad science and management applications.
Unlike previous NRC reports for specific USGS programs like NAWQA (NRC, 1990), NWUIP (NRC, 2002c), and NSIP (NRC, 2004d), detailed recommendations on the sampling design and implementation of a river monitoring system cannot be made for a USGS river science initiative; such an initiative encompasses much more than a single USGS program. Hence, the sampling design and implementation of its river monitoring components would depend on the scope and institutional organization of the science priority areas, involve elements from all the science disciplines of the USGS, and conceivably build upon and integrate monitoring activities of many existing programs. Still, general principles that should guide enhanced monitoring activities are worth mentioning.
As with the sampling design for NSIP and NAWQA, a USGS river monitoring system needs a sound scientific framework for implementation. For example, the USGS has outlined five components for NSIP, covering topics ranging from data collection strategies to information dissemination and measurement technologies. For its base streamgaging network, the USGS has defined specific
monitoring goals, ranging from measuring flows at major river outlets for water budget accounting to monitoring sentinel watersheds to evaluate long-term trends. A similar design framework is needed to establish a meaningful USGS river monitoring system.
A modern river monitoring system would necessarily contain certain elements. There would be a core network of river monitoring sites closely coupled to the NSIP network. Selected sites within the core network would maintain intensive integrated reach-scale measurements of hydrologic, geomorphic, chemical, and biologic variables. The system would be supported by the mapping of the physical characteristics of rivers and the riparian vegetation at reach-scale monitoring sites. The design of the river monitoring network would meet specific monitoring goals, such as determining sediment transport and sediment budgets for major basins, and assessing trends in river conditions. The data collected by the river monitoring system would be quality controlled and archived using data and geospatial information standards (see the following section on “Integrated Data Archiving, Dissemination, and Management”) that facilitate public access and use.
Expanding monitoring at sites within the network should be designed primarily to support USGS river science priorities for improved understanding and characterization of river processes. A long-term commitment to monitoring at selected sites is required to develop necessary biophysical data archives. The sampling must also consider the data needs for estimation of critical river conditions at unmonitored sites, as is exemplified by the USGS activities in regional interpretation of flow and water-quality data (e.g., SPARROW).
Clearly, cost constraints mean that choices have to be made on what variables to measure and where. The selection of hydrologic, geomorphic, chemical, and biological variables to monitor at sampling sites should be guided by both the river monitoring system design goals and the river science drivers within the region. By way of illustration, Appendix B provides a brief survey of pressing river science questions, and describes the data inventories that are needed to address these questions. Many of the needed datasets can be assembled from existing inventories of the USGS and other agencies. Others may be met through monitoring activities designed to support specific river management decisions (by the USGS or other agencies). Still, data gaps will remain in many situations. By identifying key river science questions on a regional basis, the USGS can target its limited resources to expand sampling at NSIP sites and index biological reaches to provide long-term monitoring of variables that address relevant river science data needs.
In some cases, it may be possible to leverage or combine monitoring efforts of existing programs (e.g., National Stream Quality Accounting Network, NAWQA, LTRMP, National Research Program) in more efficient ways. For example, the Water Quality in the Yukon River Basin Project provides an example of integrating science and monitoring programs. Data obtained from fixed
site and synoptic sampling of water quality and lake sediments is being used to address the multiple scientific issues in the basin. These include (1) carbon cycling: the origin, quantity, chemical characteristics, fate and transport of carbon at specific locations along the river and at high and low flow; and (2) mercury stores and cycling: concentrations and chemical forms of mercury in water, sediment, and biota; transport phases of inorganic and methylmercury; methylation rates in sediments; and release rates of gaseous mercury. The program, still in its early stages, was initiated with funding primarily from NASQAN; other USGS contributors include the Earth Surface Dynamics, Mineral Resources, and Biomonitoring of Environmental Status and Trends Programs.
Analogous to the data collected with the streamgaging network, which are used for a range of purposes that were never conceived by the individual gage sponsors, so too can river monitoring data be used for a range of river science questions. A perusal of the river science questions and data needs in Appendix B shows that many kinds of monitoring data, including baseline ecological data, indicator species, water flow and timing, and sediment fluxes and grain size, can be heavily leveraged for many different kinds of river science questions. Whenever possible, the USGS should prioritize its expanded data collection and mapping activities on those variables with broad science and management applications.
River Monitoring—Partnering in Monitoring Efforts
Recommendation: The implementation of a USGS river monitoring system should be informed by the data and science information gaps that limit effective policy and management decision making of other organizations, including mission-oriented government agencies at federal, state, and local levels, nongovernmental organizations, and academic research institutions. Partnering with these groups to design and implement scientific data monitoring in support of site-specific management and research objectives must be a component of USGS river monitoring.
A unique challenge and opportunity for the USGS in developing a coordinated river monitoring system is that many other federal and local agencies have a diverse range of river monitoring components to support their management activities. For instance, the EMAP effort of the EPA collects physical, chemical, and biological data on rivers to assess existing conditions at monitored and unmonitored sites (e.g., the Mid-Atlantic Integrated Assessment [MAIA]). EMAP has developed a considerable body of knowledge on indicators that the USGS can profit from. EMAP’s probabilistic sampling design is in stark contrast with the USGS’s fixed location sampling network, but this difference in approaches can be turned into a strength as results from the programs are compared
and their resulting predictions tested. Thus, the USGS river monitoring system should seek to complement, rather than duplicate or replace, monitoring efforts of other agencies.
Increasingly, river science monitoring problems will involve partnerships between the USGS and other agencies involving site-specific problems, with specific actions and expected outcomes. Clearly, USGS expertise in the design of data collection networks for experiments, and its reputation for impartiality, can contribute to such multiagency efforts by providing baseline data and interpretation of management outcomes. As it does with many of its individual programs, USGS coordination of this work within the larger framework of a national river monitoring system would enhance both activities.
Likewise, a USGS river monitoring system would be enhanced by partnerships with recently proposed environmental observatories. These include the National Ecological Observatory Network (NEON) and the joint proposal for hydrologic observatories and advanced environmental measurement technologies by the Consortium of Universities for the Advancement of Hydrologic Science, Incorporated (CUAHSI) and the Collaborative Large-scale Engineering Analysis Network for Environmental Research (CLEANER). The USGS should seek to collaborate with these groups, and integrate contributions from these efforts within its river monitoring sampling design.
River Monitoring—Measurement Technologies
Recommendation: The USGS must remain at the forefront of river monitoring technologies. The development of new cost-effective instrumentation and measurement techniques for monitoring physical and biological variables is an essential component of a river monitoring system.
To develop a modern river monitoring system, the USGS also needs to be at the forefront of river measurement technologies. Advances in sensor and network technologies, wireless communications, and remote sensing are likely to change the way environmental monitoring is done in the future. Extensive research on embedded networked sensing for environmental applications is underway in both laboratories and field stations. For example, researchers at the James Reserve in Southern California have been testing a diversity of wireless and robotic networked instruments since 2002. These include aquatic sensors; acoustic animal sensors; instruments for collecting root, soil, and fungi data; and live webcams. Similar efforts have begun at the Santa Margarita Ecological Reserve, also in Southern California, and field experiments are underway in snow-dominated environments as well.
Thus, opportunities will exist to gather more data and increase the variables that can be monitored, all at lower costs than today. Sensors will be able to
communicate with one another in order to increase or decrease monitoring frequency as appropriate, thereby lowering costs and increasing the quality of the information flow. The USGS must continue to investigate and invest in new river monitoring technologies to develop a cost-effective sampling design for river monitoring and to enable integration of these data with those from sensor networks at various scales.
The USGS is already heavily involved in evaluating new technologies for NSIP. Increasingly, acoustic methods are being used to map the three-dimensional flow velocity structure of rivers, rather than simply estimating area-integrated variables such as discharge. As part of the USGS HYDRO-21 program, noncontact measurement technologies, including Doppler radar techniques, are being tested for direct and continuous noncontact measurements of river flows (Costa et al., 2000). Novel techniques, combining concepts in river hydraulics and fluid mechanics, are being employed to derive river stage-discharge ratings from first principles, rather than empirically through repeated (and costly) direct discharge measurements (Kean and Smith, 2005). These efforts need to be broadened to include all elements of river monitoring, and a continual program of measurement technology assessment must be implemented as an integral part of a 21st-century river monitoring system.
A greater emphasis on the use of remote sensing techniques for river monitoring is also needed by the USGS. Terrestrial, airborne, and satellite remote sensing technologies have significant potential to advance river science by providing observations related to a river’s temperature, water quality, and ecosystem functions, and the vegetation type and physical characteristics of riparian areas, with better spatial coverage than attainable with in situ measurements. One of the most promising technologies is airborne light detection and ranging (lidar). As part of the Federal Emergency Management Agency’s (FEMA) Map Modernization program for flood hazard mapping, the use of lidar-based measurements to create very high resolution digital elevation models of floodplain areas is becoming commonplace. Within the USGS, side-scanning terrestrial lidar is being used to map coastal bluffs; similar techniques could also provide a cost-effective approach for obtaining detailed information on river channels and their changes (e.g., due to streambank erosion).
Space-based remote sensing products for land surface monitoring, developed by NASA and other agencies (including the USGS); can also support river science activities. For example, the Enhanced Thematic Mapper Plus (ETM+) from the Landsat 7 satellite provides high resolution (15-60 m) multispectral data to characterize land surface conditions. The Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) and the Moderate Resolution Imaging Spectroradiometer (MODIS) on NASA’s Terra satellite provide high (15-90 m) and medium (250-1000 m) resolution information on vegetation and land cover. These and other satellite and airborne remote sensing products are readily available to USGS researchers from the USGS Earth Resources Observa-
tion Systems (EROS) Data Center (see Chapter 3). EROS data archives, obtained in part in collaboration with other federal agencies, contain not only aerial and satellite remote sensing imagery but also maps, land cover data, and other derived products.
For river science, the USGS needs to be actively involved in developing ways to combine terrestrial, airborne, and satellite remote sensing information with in situ measurements at index reaches and study sites, to develop indirect means for mapping the physical and biological characteristics of floodplain and riparian areas along the river corridor. These efforts will require collaboration and cooperation with other federal agencies (e.g., FEMA and NASA), and should leverage existing interagency relationships and capabilities of the EROS Data Center.
INTEGRATED DATA ARCHIVING, DISSEMINATION, AND MANAGEMENT
River science is an integrative science, requiring synthesis of measurements from diverse sources, collected over disparate time and space scales. In addition to streamflow data and point measurements of water quality, the suite of observations that support river science investigations include two-dimensional data and observations describing stream channel geometry, and time-varying data on bed forms, channel sediments, and the land uses and vegetative cover of riparian corridors and upstream drainage areas. Three-dimensional data describing flow velocity fields are also now available from innovative acoustic Doppler technologies and even four-dimensional measurements (i.e., time-varying three-dimensional fields) are both technologically and economically practical data forms with great potential value for river science.
The interdisciplinary needs for river science data—common to many agencies, institutions, and stakeholders—have the characteristics of a public good, and merit public support for not only data collection, but also archiving, maintenance, and dissemination. USGS information is frequently used to support and inform decision making. Decision support systems, as connectors between science and decision making, can be improved if the data are compatible with these systems (NRC, 1999b). Therefore, to maximize the value of the appropriate public investment in data, the design and implementation of enhanced data collection activities to support river science should carefully examine the opportunities and emerging technologies to standardize data collection protocols, carefully document quality assurance and quality control procedures and metadata, and exploit technologies for data sharing and virtual data warehousing.
In light of this variety of river science data and their multiple uses, this next section addresses how USGS databases that span multiple disciplines need to be modified to better store, manage, and disseminate river science data.
Recommendation: The USGS should include in its river science initiative an informatics component that includes developing a common data model for river science information that can be used to archive the diverse river science metadata and data. This data model should be developed in coordination with and capable of supporting other federal agency river science data needs. The data model should accommodate data from multiple sources, including nonfederal sources. Such a program would facilitate the integration and synthesis of river science data to address the diverse range of river science questions discussed in previous chapters.
Data Management Needs for River Science
For many river science questions, collecting data to address critical data gaps is just the first step; effective data management systems for the archival and dissemination of interdisciplinary datasets are also needed to facilitate the integration and synthesis that advance our understanding of river processes. River restoration, as described below, is a good example of how a data management system can greatly enhance the efficiency with which researchers and managers can understand the river system,
Government agencies and various stakeholders now accept river restoration as an essential complement to conservation and natural resource management. However, despite legal mandates, massive expenditures, and the burgeoning industry of aquatic and riparian restoration, many restoration activities have failed. Reasons include lack of (1) a solid conceptual model of river ecosystems; (2) a clearly articulated understanding of ecosystem processes; (3) recognition of the multiple, interacting temporal and spatial scales of river response; and (4) long-term monitoring of success or failure in meeting project objectives following completion. These problems suggest that the scientific practice of river restoration requires an understanding of natural systems at or beyond our current knowledge, and this presents a significant challenge to river scientists.
The National River Restoration Science Synthesis (NRRSS; http://nrrss.nbii.gov) Restoring Rivers.org project (http://nrrss.nbii.gov/) is an effort to facilitate integration and synthesis across multiple disciplines using diverse databases and models. The NRRSS maintains a database of more than 50,000 restoration projects from around the country. This database holds information on why projects were undertaken and how they were implemented and provides the opportunity for synthesis to understand the effectiveness of river restoration practices. This database provides much useful information for the evaluation of the effectiveness of river restoration efforts based on sound river science, and serves as an example of the synthesis that can be achieved through the use of integrated data management systems. Participants in the NRRSS have called for major
efforts to develop national standards for reporting restoration and a national tracking system to ensure better coordination of restoration efforts and to facilitate effectiveness evaluations. They identified USGS as an appropriate candidate for spearheading this effort (Palmer and Allan, 2006).
The USGS Role in Data Management for River Science
From a wide range of studies related to river processes, considerable data resources already exist at the USGS. Much of the data is archived and disseminated through data management systems, including the NWIS, NAWQA program/Data Warehouse, NBII, The National Map, EROS-EDC (Earth Resource Observation Systems-Earth Data Center), and others. However, some of the data collected are in the individual holdings of scientists and thus data users have difficulty locating datasets that serve their needs. Even more importantly, data are archived in ways that make sense from the perspective of the study for which the data were collected, but that may not be conducive to integrated data analysis with other data types.
For river science, it is in the national interest for data holdings to be standardized and archived with sufficient ancillary information (metadata) that allows them to provide traceable heritage from raw measurements to useable information and allows the data to be unambiguously interpreted and used. Technological capabilities and measurement methods are advancing rapidly, yet synthesis of trends over time mandates standard, consistent, fully documented measurement protocols and the maintenance of data in systems that are accessible using the most current technology. Given the diversity of USGS river science information, this is a daunting information science (informatics) challenge. Nevertheless this challenge needs to be addressed if the capability for national synthesis is to be sustained.
A large number of federal agencies are involved in various aspects of river science data collection. Cooperation and coordination among these agencies and nonfederal partners is needed for the management and sharing of data in a clear, public, national way. Cooperative federal data initiatives, such as the Office of the Federal Coordinator for Meteorological Data, and the Subcommittee on Water Availability and Quality (SWAQ) of the Committee on Environment and Natural Resources of the National Science and Technology Council (NTSC) (the primary coordinating and planning group for water related science and technology in the federal overnment), offer lessons and experience to guide river science data needs. Emerging multiagency efforts such as NBII, The National Map Initiative, and NHDPlus serve as examples that can guide river science data management and coordination.
In principle, any of the federal agencies involved with river science could take the lead in coordinating data management and dissemination. There are, however, three compelling reasons why the USGS should be a leader in this
area. First, the USGS has considerable expertise and experience in data management and dissemination, supporting systems such as NSIP, NAWQA, NBII, and The National Map. Second, the USGS, as a science agency, has a reputation for impartiality and quality. And third, the primary focus of the USGS is on earth science, which in its broad sense includes river science.
Existing Data Management Systems Supporting River Science Activities
A first step in designing standard data models that accommodate integrated archiving and dissemination of river system data is to consider how well existing data models and data management systems (of the USGS and other institutions, including private and commercial database management systems) support river science activities. What follows is an evaluation of these data systems, where we highlight both their strengths and limitations. Insights from this review motivate recommended design principles for an improved approach.
NWIS—NWIS (http://waterdata.usgs.gov/nwis/), the USGS Water Resources Discipline’s flagship data system, has served daily USGS streamflow to the public and provided an internal USGS system for the distributed archiving and management of streamflow data for many years. It also contains NASQAN data. In many respects with NWIS, the USGS has led the way in making river science data publicly available. However, NWIS was designed at a time when computers and network bandwidth were much more limiting than today and when commercial database systems and web service capability were in its infancy or not existent. As a result, many of the features of NWIS appear primitive by modern standards, such as the limitation of data distribution via NWISWeb to 24 hour daily values rather than unit values recorded more frequently at stream gages, an implementation limitation adopted due to computer system limitations at the time computerized storage of streamflow data was developed.
NAWQA data warehouse—The NAWQA data warehouse (http://water.usgs.gov/nawqa/data/) has been developed to facilitate national and regional analysis of data from NAWQA study units. The data warehouse contains links to the following data:
Chemical concentrations in water, sediment, and aquatic organism tissues and related quality control data (from NWIS);
Stream habitat and community data on fish, algae, and benthic invertebrates;
Site, well, and basin information associated with thousands of descriptive variables derived from spatial analysis like land use, soils, population density, etc.; and
Daily stream flow and temperature information for repeated sampling sites (from NWIS).
In contrast with NWIS, the more recently developed NAWQA data warehouse was designed to make maximum use of the most current off-the-shelf software. Development was by USGS staff with expert consultants hired from the private sector. The warehouse uses relational databases and web services to extract information from data sources and transform and stage it for loading into the NAWQA databases, which provide read-only open database access that supports querying to facilitate integrative data analysis and synthesis.
NBII—NBII (http://www.nbii.gov) is a broad, collaborative program that provides increased access to data and information on the nation’s biological resources. NBII links diverse, high-quality biological databases, information products, and analytical tools maintained by NBII partners and other contributors in government agencies, academic institutions, nongovernment organizations, and private industry. NBII uses a distributed, node-based architecture, with considerable diversity of information presented across nodes. This fulfills the needs of making data available, but the diversity of formats appear to hamper the integration and synthesis of information across the nodes. NBII was instrumental in helping to build the database structure for the NRRSS (http://www.restoringrivers.org).
Regional USGS databases—As noted in Chapter 3, a number of the USGS’s river programs have their own databases. For example, LTRMP on the Upper Mississippi River System has the largest integrated database (2.2 million records) on the nation’s largest river; it provides query tools and graphical browsers for a variety of datasets, and is used to produce decision support and synthesis reports. Where developed, such regional databases should be closely examined and compared for broader applicability.
The National Map—The National Map (http://nationalmap.gov) is the USGS’s interactive map service intended to provide a consistent framework for serving the nation’s mapping needs. The content that has been integrated into the national map is impressive; however, it is primarily geographic.
NHDPlus—NHDPlus (http://www.horizon-ystems.com/NHDPlus/index.htm) is an integrated suite of application-ready geospatial datasets that incorporate many of the best features of the National Hydrography Dataset (NHD), the National Elevation Dataset (NED), the National Land Cover Dataset (NLCD), and the Watershed Boundary Dataset (WBD). The NHDPlus consists of nine components:
Greatly improved 1:100K National Hydrography Dataset;
A set of value-added attributes to enhance stream network navigation, analysis, and display;
An elevation-based catchment for each flowline in the stream network;
Headwater node areas;
Cumulative drainage area characteristics;
Flow direction, flow accumulation, and elevation grids;
Flowline min/max elevations and slopes; and
Flow volume and velocity estimates for each flowline in the stream network.
The NHDPlus is being developed by Horizon Systems, a contractor working for the EPA, with considerable USGS participation in the development team. Much of the stream reach information available in NHDPlus is particularly relevant for river science.
Arc Hydro—Arc Hydro was developed as a generic geographic data model for water resources and comprises hydrography, watersheds, the stream network, channels, and hydrologic time series. Recently this framework has been extended to groundwater. Arc Hydro has been implemented as a geodatabase schema and toolset using ESRI’s ArcGIS geographic information system software. Arc Hydro provides a structured model for the representation of geographic surface and groundwater features that facilitates integrated analysis based on these features. The Arc Hydro structure provides a basis for hydrologic information systems and the nascent hydrologic information science.
Web-service-oriented data delivery—Much Internet-based data delivery involves browsers and portals. However, the Internet delivery paradigm based on web services is growing in use and capability. Web-service-based systems are being developed by some of the National Science Foundation (NSF) cyber infrastructure initiatives such as CUAHSI, CLEANER, Geosciences Observatory Network (GEON), and NEON. Web-service-oriented data architecture uses Internet-based object access protocols to enable access to data directly from application programs. This serves to make data available in the analysis environment of a scientist’s choice, such as Matlab, Excel, or ArcGIS. The web services direct access model avoids the need to use a browser to access and download the data required for analysis from multiple sources. Web services integrate data delivery with analysis and provide a mechanism for integration across data holdings from different sources. To the extent possible, the USGS is encouraged to develop a data model that supports web service access to river science data.
Design Principles for USGS River Science Data Management Models
The examples presented above provide a survey of some of the database management systems available that are relevant to river science. Each has its strengths and weaknesses. No one of these could completely provide for the needs of river science. In developing river science data models to fulfill the needs of a USGS river science initiative, we suggest these systems be reviewed to determine which aspect of each could be adopted.
A common data model would provide an intellectual framework under which river science data holdings are catalogued and accessible. To develop such a model, a strategic plan put together by informatics experts from the USGS, other
agencies, and academics needs to be developed. The notion of “dataspaces” (Franklin et al., 2005) extends information management beyond one integrated database into distributed information elements managed by different participants. The following characteristics of dataspaces merit evaluation as part of the common river data model framework:
Management of a dataspace system does not assume complete control over the data in the dataspace.
The dataspace encompasses all of the data and information in the organization regardless of its format or location.
The key service provided in the dataspace is the cataloging of participant data elements. This cataloging provides basic information, including data quality that supports other dataspace services like searching and querying.
Not every participant data source is necessarily interfaced to support all of the integrative functions available. The inclusion of these data sources is prioritized and approached as needed.
In designing a data management system for the archiving and dissemination of river science data, the data model should be constructed in accordance with the standards of the National Geospatial Data Infrastructure and should be based on sound, robust, and scalable relational database and geographic information science design principles that can be implemented using advanced commercial database technology. Implementation should include coordination with other federal agencies and nonfederal partners involved in river science and should incorporate analytic capability for scientists to efficiently query and use the data in the course of their research. Ideally the system should make river science data analysis easier so it becomes the preferred platform for data analysis and sharing, rather than a chore for scientists to have to load their data into at the end of a project. The data model implementation should facilitate the Internet dissemination of river science data approved for public use. To achieve this coordination and acceptance, we suggest that the system not be a centralized, single entity but rather a distributed confederation of participants, each responsible for and knowledgeable regarding their own element. The involvement of people fulfilling multiple roles (users, data providers, and analysts) working with the system is as important as the technology.
Whatever data models arise to support river science, one can confidently predict these needs will continue to evolve in unanticipated ways with the emergence of new problems, sensors, and assimilation techniques. The modern history of national data collection efforts offers vivid examples of data that has been “stranded” by technological innovations in database management. One example is the vast store of archived paper records from the national streamgage network that are ripe for data rescue (i.e., the conversion from paper to digital formats). Another example is the modernization of the EPA STORET system for manag-
ing water-quality data from a broad varying community of users. Challenges in incorporating data collected using historical protocols has resulted in the awkward use of Legacy STORET and Modernized STORET. The data management system to support an integrative river science should therefore be designed and structured to be adaptable and scalable in order to avoid stranding the next generation of river science data, as the information technologies continue their rapid pace of innovation and development.
The ultimate goal should be a common data model for river science information that fulfills the needs of the USGS, other federal agencies, nonfederal partners, and the public. This needs to include a system that is queryable, easy to use, and provides effective visualization and analysis capability. An open architecture data model is suggested to clearly and unambiguously document the data being presented. Another goal is the infusion of expertise into USGS river science staff, to the point that data management based on advanced relational database systems becomes part of the culture and business practices of the organization. Just as the USGS has historically piloted many innovative physical measurement methods, the USGS should invest in evaluating and testing new data management techniques and then training and disseminating the findings to individuals throughout the organization.
To provide long-term baseline science information on our nation’s rivers, and to support research in USGS river science priority areas, new river monitoring and data management activities are essential for a USGS river science initiative. The USGS has historically provided unbiased fundamental river science data used to characterize river processes. Maintaining and expanding this capability is critical to the national need, particularly because long-term datasets on river hydrology, chemistry, and biota are essential for addressing future river problems that broadly affect the nation’s economy and well-being. Some of these problems cannot be anticipated in detail now, but baseline information will be needed to investigate causes of river changes at all scales. Expanded monitoring is also needed to address the USGS river science priority areas identified in Chapter 4. Currently, the science data gaps that exist are an impediment to progress in these areas, and future science advances are predicated on new river monitoring and data management efforts.
Expanded data monitoring for river science should focus on developing integrated datasets on the hydrologic, geomorphic, chemical, biological, and ecological conditions of rivers. This objective could be achieved in part through the establishment of reach-scale monitoring sites. To plan and implement a 21st-century system for river monitoring, the USGS will need to define specific monitoring goals, and develop a scientific framework for sampling to meet these goals. Advancing measurement technologies to support river monitoring
is important to this effort, as is partnering and cooperation with other federal and local agencies, since many groups collect and archive data relevant to river science.
Archiving, disseminating, and managing integrated datasets for river science is a challenging problem. However, advances in information technologies are transforming science by making complex data archives easier to access and analyze. An informatics component is needed to develop a common data model for efficiently archiving and distributing datasets and metadata. Conceptualizing such a data model, and putting it into practice, will require an ongoing, long-term effort on the part of the USGS as an institution, as well as its individual disciplines. Coordination and partnership with other federal agencies and nonfederal partners is important. The coordination necessary to achieve the goal of a common data model for river science may serve to stimulate integration between fragmented river science activities across the federal agencies, the nongovernmental sector, and within the USGS, and provide a basis for a coordinated interdisciplinary management approach, as considered in the following chapter.