Integrating Multiscale Observations of U.S. Waters (2008)

This report offers a broad review and vision of integrated observing for the hydrologic and related sciences. As was articulated in Chapter 1 and the case studies in Chapter 4, population growth and global climate change will increasingly strain the use of fresh water for human activities and require greater efficiencies in our utilization of this essential resource. The difficulties in quantifying many of the stores and pathways of water, and the related energy and biogeochemical fluxes through the environment add layers of uncertainty to the problem. Since these fluxes vary over time and space scales, the integration of observations, data management, and process-based predictive models is essential to provide the information needed to expand our understanding of the linkages among the water, energy and biogeochemical cycles in both pristine and highly modified areas, and to provide useful information for water and environmental managers.

The vision is of the future where on-site (“in-situ”) sensors that measure properties such as temperature, soil moisture, and water quality at high spatial and temporal resolution are integrated to form an embedded network that is in turn connected to other networks of observations. These other networks may consist of traditional observation platforms such as ground-based precipitation gages, river discharge gages, or “grab samples” for water quality, or remotely sensed measurements from airborne or spaceborne sensors at larger spatial scales. The vision includes the ability to integrate these measurements across all relevant scales and with models to offer wide-ranging predictive capabilities, and to make both data and predictions widely available through “web portals”—that is, through Internet-based access where users of the predictions and meas-

urements such as water managers and educators could easily access the information.

Measuring physical, chemical, and biological properties at multiple scales will increase our process-based and predictive understanding and improve management. An excellent example of this is provided in Chapter 4 by the Neuse River Basin Study, because nitrogen cycling in that basin is controlled to a large extent by microscale processes in regions such as the hyporheic zone, yet understanding actual ecosystem dynamics and effects on estuarine systems requires the broad view provided by spaceborne remote sensing.

While this vision appears futuristic, the study found—and the report documents—that many elements of this vision currently exist. For example

• Sensors: Chapter 2 presented a comprehensive discussion of sensors at a wide variety of scales that are currently under development and being tested in experimental observatories or in research projects that focus on particular environmental variables. This chapter also identified approaches that are emerging in other disciplines that may lead to additional new environmental sensor development, and that will need to be modified for actual field deployment to address issues in water and environment. These advances in measurement technologies range in scale from nanosensors focusing on biological variables to airborne sensors that offer spatial context for point measurements to satellite sensors that offer regional-to-continental scale perspectives.

• Embedded sensor networks: Embedded sensor networks, consisting of spatially distributed sensor-containing platforms connected to and often controlled by computers, and with the sensors themselves often containing microprocessors, are also being demonstrated. For example, as presented in Chapter 2, wireless but interconnected instrument buoys at remote lakes in Wisconsin have provided frequent measurements of lake quality. Dense measurements of hydrologic and meteorological sensors at the Santa Margarita Ecological Reserve are providing real-time data to drive watershed models. And sensors embedded in sewer systems are helping to combat combined sewage overflow.

• Other communications and cyberinfrastructure: In Chapter 3, the NSF-supported CUAHSI project development of “web portal” applications was presented along with one current example using this technology, the Central America Flash Flood Guidance (CAFFG) System. The importance of cyberinfrastructure was highlighted in several of the case studies presented in Chapter 4.

• Modeling and data assimilation: Chapter 3 also reviewed the considerable advances in data assimilation, that is, the merging of data across scales from in-situ point measurements to coarser scale airborne or satellite observa-

tions for use in models. Data assimilation has long been used in applications such as operational weather prediction and in ocean modeling, and its application in land hydrology and environmental science is being demonstrated in research applications.

• Existing experimental watersheds, planned observatories, and international initiatives: The USDA and others have many monitored catchments that have been and can continue to be used for basic science related to watershed processes. While these are generally at a scale too small to integrate satellite observations, they can be incorporated into the environmental observatories planned by the National Science Foundation. At the international level, the intergovernmental Global Earth Observation System of Systems (GEOSS) has been conceived and has strong international support.

Despite the great promise of integrated hydrologic measurement, many challenges stand in the way of its full development and application. Many of these challenges are illustrated by the case studies offered in Chapter 4. These studies cover a diverse range of current hydrologic research and management efforts, encompass a variety of related sciences, and provide detailed discussion on the need for integrating a wide variety of measurements with models for improved process understanding and management. While these case studies are not meant to be exhaustive in discussing integrated observations, they are representative of the breadth of opportunities facing the community. The most striking theme from the case studies is that the gaps between the vision of what the researchers or managers want to achieve and their ability to realize that vision are real, but in many cases extremely narrow. However, there are major challenges, which are listed below, followed by recommendations for overcoming these and other challenges.

Despite the promising advances enumerated above, there are at least nine major challenges that need to be overcome before the vision for integrated observations described in this report can be brought closer to reality. These relate to

• Developing appropriately scaled sites for water science;

• Developing and field deploying land-based chemical and biological sensors;

• Inspiring a greater agency commitment to developing airborne sensors;

• Developing both new spaceborne sensors and a “research-to-operations” strategy for existing ones;

• Bridging the gap between sensor demonstration and integrated field demonstration;

• Integrating data and models for operational use;

• Adopting new integrated hydrologic measurement and modeling systems for water resources applications;

• Funding interdisciplinary science; and

• Addressing the fractured federal responsibility for hydrologic measurement, monitoring and modeling.

These challenges are described in more detail in the following paragraphs.

Most existing field-based watershed research programs focus on small catchments—from <1 to (rarely) several hundred square kilometers—and most of the larger ones are not extensively monitored. They are managed by the U.S. Department of Agriculture (USDA) Agricultural Research Service and Forest Service, the U.S. Geological Survey (Water, Energy, and Biogeochemical Budgets, or WEBB, program), the Environmental Protection Agency (such as the Chesapeake Bay Program), and the National Science Foundation (Long-Term Ecological Research, or LTER, program). A large quantity of historical data exists from these sites, but much of it is difficult to access; even within the USDA these data are not in a single database.

Critically, few of these field observatories envisioned a major role for remote sensing when they were conceived, in some cases because of their small footprint, in others because they were designed before modern remote sensing methods were developed, and in still others perhaps simply because of institutional barriers. They generally succeed well in supporting the small-scale catchment science that they are designed for, but need to be about two orders of magnitude larger to approach scales that are appropriate for land-atmosphere coupling.

The proposed hydrologic, ecological, and water-quality “observatories” such as those of the National Ecological Observatory Network (NEON) and the WATer and Environmental Research Systems Network (WATERS) are designed to be much larger—on the scale of 10⁴ to 10⁵ km. Direct estimates of varying quality are now available for variables such as snow extent, snow water equivalent, surface-water height, soil moisture, total water storage, and precipitation; indirect estimates for groundwater, evapotranspiration, and streamflow can also be made in some cases (see Chapter 2, Spaceborne Sensors section). Not all of these estimates (e.g., groundwater) can be made at the scale of the proposed observatories, but improved sensors are proposed for many of these variables for the future (NRC, 2007), and data assimilation and modeling techniques are advancing as well.

There are, therefore, both unprecedented opportunities and unprecedented challenges for incorporation of space-based and airborne observations in obser-

vatories—and not as an afterthought, but as an essential part of the experimental design.

As discussed in Chapter 2, physical sensors, such as those that measure air and water temperature and pressure, radiation, and wind speed and direction, have evolved over decades and are now mass produced and routinely packaged together in small instruments along with power and communication devices. However, sensor development for many important chemical and biological measurements is relatively immature. Chemical sensors are needed to measure a wide range of elements and inorganic and organic molecules, in all environmental media. Biological sensors can provide key information on the function and structure of biologically influenced ecosystems in real time (NSF, 2005). For widespread use in the field, chemical and biological sensors would need to be inexpensive; robust against environmental stresses, such as temperature extremes and biofouling; have stable calibrations or be capable of remote calibration; and have low energy requirements. Development of a wide range of field-robust chemical and biological sensors is one of the greatest challenges facing widespread deployment of sensor networks in the hydrologic sciences.

As discussed in Chapter 2, airborne and spaceborne sensors are needed in integrated observing systems to extend observations beyond the point measurement scale, with airborne measurements at a spatial scale that fills the gap between the in-situ plot-scale observations and the larger satellite-scale observations. As discussed in Chapter 2 and in National Research Council (2007), airborne remote sensing at the National Aeronautics and Space Administration (NASA) historically has been viewed as an intermediate step between initial sensor development and space deployment. Measurements from such airborne systems were collected during limited-duration field campaigns to help develop retrieval algorithms. Airborne sensors are also used to under-fly new satellite sensors for their initial validation. In these ways, NASA views its airborne science program as supporting its satellite sensor development program and not as a sensor program in its own right. This approach has unfortunately impeded the development of operational airborne observing platforms that could play a very important role in hydrologic observations.

From the above discussion, as illustrated by the case studies in Chapter 4, the critical challenge to address is the limited commitment of NASA and other federal agencies to airborne platforms as effective operational measurement systems, lead-

ing to a paucity of programs to develop smaller, less-expensive sensors that could be used on these platforms.

In the area of satellite-based remote sensing, NASA has made good progress in developing and deploying a wide range of sensors for hydrologic science that are used primarily for research. Most of these are described in Chapter 3. Addressing the future needs in this area was the charge of another NRC committee (the “decadal survey”; NRC, 2007), whose relevant results are summarized in Appendix C. The missions that study recommended—including missions to measure diurnal precipitation, soil moisture, water storage in lakes and wetlands, and snowpack water storage, especially in mountainous regions—are consistent with the vision, findings, and recommendations of this study.

Nonetheless, two related challenges are relevant to this report: (1) for NASA, how to transition research sensors (and their costs) to operational agencies, and (2) for other (non-research) users, determining how to utilize these observations for operational uses. In the case of weather-related sensing, Congress has provided guidance. For other agencies and users, this remains a challenge. As an example, the loss of high-resolution (<100m) thermal imaging on the Landsat TM satellite and the decision by NASA not to have such a capability on the replacement satellite has affected agencies such as the USDA and state agencies concerned with irrigation water management that use this information operationally.

Thus, the critical challenges to address here are (1) a resolution of the “research-to-operations” transition from NASA-developed “experimental” satellite observations to the broad variety of operational agencies and users that need routine (i.e., operational) observations, and (2) the lack of a corresponding monitoring strategy by federal and state agencies (for example, the Environmental Protection Agency [EPA], USDA, the National Oceanic and Atmospheric Administration [NOAA], and state water and natural resources agencies) that would incorporate airborne and/or satellite remote sensing measurements, where appropriate (various case studies in Chapter 4 offer examples). These apply to existing but underutilized data sets as well as to future data sets.

The last section of Chapter 2 (see Figure 2-10) presented the steps needed to advance sensors from experimental development through operational deployment. Also shown in the figure are various agencies that are currently involved in the different steps of the process or in complementary activities. While Figure 2-10

shows an idealized summary with overlapping agency activities, in reality there are significant interagency gaps between the steps of sensor development, sensor demonstration, integrated field demonstrations, and operational deployment of sensors. The greatest gap is between sensor demonstration and integrated field demonstration. Closing this gap would involve integrating the sensor networks and webs within hydrologic observatories and experimental demonstration sites, and interfacing the sensor networks with the broader development of cyberinfrastructure.

The importance of data-model integration is apparent in a number of the case studies. For the Mountain Hydrology study in Chapter 4, predictions of water availability are made from point measurements and model forecasts, leading to management decisions. In the Neuse River Basin Study, the management decisions are based on sparse water-quality measurements, largely limited to the main stem and high order tributaries. In these two case studies, both models and observations are used to guide management decisions; in each case a data assimilation system that merges models and observations would offer improved predictions for decisions. However, the systems required would likely be very different, as they would be for many other critical water resources problems. The challenge is to develop data-model integration methods that will be useful for broad families of applications, rather than just a few of the many possible applications.

In the United States, large water resources problems involve multiple stakeholders, including government agencies, business interests, and the public. Management is typically diffuse, although the benefits of coordinated management are well recognized. Standard measurement and modeling techniques and rules for water management are entrenched and often legally mandated. For example, as mentioned in the case study on Mountain Hydrology in the western United States, the principal methods for monitoring snow pack and predicting snowmelt volumes have changed little over the past 100 years. This has the benefits of producing a consistent data set to show trends over time, and of simplifying training and daily tasks of staff. However, it also leads to missed opportunities to improve the accuracy and precision of the data and resulting model predictions. Changing modes of measurement and management will be extremely difficult, even as the nation’s water resources challenges increase over time.

Interdisciplinary science is much more common than it was 20 years ago, and it is no longer uncommon to see scientists and engineers from a variety of fields working on related problems in experimental watersheds. However, the design and use of integrated hydrologic measurement systems in specific research applications adds an extra layer of complexity to the challenge. Thus, these new kinds of projects will require unprecedented interdisciplinary cooperation, including interactions among scientists, engineers, field researchers, modelers, and theorists. Each application will be somewhat unique, requiring the participation of the technology developers (electrical engineers, computer scientists, and modelers) and the physical, chemical, and biological scientists who apply the technology to hydrologic research. While many universities and research laboratories have the required diversity of expertise, marshalling this expertise on specific projects will likely require new programs or sources of funding.

The overarching barrier to the development and implementation of integrated hydrologic measurement systems is the lack of a single federal agency with primary responsibility for measuring, monitoring, and modeling the environmental factors and processes that control the hydrologic cycle. USGS is the agency with a mandate that comes closest to meeting this responsibility. It continuously monitors streamflows at over 7300 sites, as well as lake and groundwater levels at a smaller number of sites. It also collects data on stream water quality at about 2800 sites. Through its National Water-Quality Assessment (NAWQA) Program, USGS performs integrated assessments of a small number of watersheds. NOAA has the responsibility for monitoring precipitation and other meteorological variables and for monitoring and managing coastal waters. EPA has regularity authority for water quality and oversees the setting of water-quality standards on impaired stream reaches through the Total Maximum Daily Load (TMDL) Program. EPA also manages the Environmental Monitoring and Assessment Program (EMAP), which assesses national ecosystem health. Hydrologic modeling has been advanced by EPA, U.S. Army Corps of Engineers, the USGS, and NOAA—agencies with a wide variety of missions. Finally, NASA develops, deploys, and maintains satellite sensors that are used by others to assess the environmental factors, although its primary mission is space exploration.

Thus, it is easy to understand why the responsibility for measuring and monitoring the environmental factors and processes that control the hydrologic cycle is so fractured, given the evolution in our management of water resources

and of our understanding of the hydrologic cycle. But the dual threats of global climate change and population growth demand a focused strategy for providing information on the nation’s water resources and the environment.

As shown in this study, there is significant development in new innovative sensors, especially in the area of chemical and biological sensors. Sensor development is taking place for a number of hydrologic processes and these new sensors are being tested within embedded networks. But these developments are not well coordinated, especially in their testing through field deployment. Linking sensor development to integrated field demonstration and operational deployment by state and federal water agencies needs to be fostered. Involvement of the private sector can help to ensure that sensor innovations are carried through the development phase and become incorporated in products that are widely distributed and economically viable.

There is an important opportunity for partnerships between the National Science Foundation (NSF), NASA, other agencies, universities, and the private sector in the development of new sensor systems. Such partnerships would enable the development of sensors and sensor systems that address multidisciplinary observational needs, including the needs of operational agencies. Interagency laboratories have been created in the past and have achieved some success, as in the case of the national sedimentation laboratory in Vicksburg, Mississippi.

Recommendation 1-1: NSF, in partnership with NASA, NOAA, EPA, USGS, and possibly national health and security agencies, and with collaboration from the private sector, should develop one or more programs that address the need for multidisciplinary sensor development. An interagency sensor laboratory should be considered.

Concepts such as the Consortium of Universities for Advancement of Hydrologic Science’s proposed “Hydrologic Measurement Facility” (HMF) Center, which would house community instruments based on mature technologies, and HMF “nodes,” which would be university-based, have a 3-6 year life cycle, and support a specific emerging technology, would complement but not replace such programs. In particular, universities are the ideal environment for training future users of environmental sensor networks and the data that will stem from them.

As described above for sensor development, and earlier in the challenges facing the development and deployment of integrated hydrologic measurement systems, the pace of advancement is constrained by the fractured federal respon-

sibility for the development and deployment of integrated measuring, monitoring, and modeling systems, which seems to impede their utilization.

Recommendation 1-2: Serious consideration should be given to empowering an existing federal agency with the responsibility for integrated measurement, monitoring, and modeling of the hydrological, biogeochemical, and other ecosystem-related conditions and processes affecting our Nation’s water resources.

This agency would manage the design, development, and deployment of integrated hydrologic measurement systems. Candidate agencies may include NASA, NOAA, and USGS because they all have extensive sensor and observations systems and modeling, and have the potential to take such a leadership role. It is recognized that taking on such a leadership activity would require new responsibilities, probably new organizations within the responsible agency, and new funding to carry out the mandate. There are some areas, such as cyberinfrastructure and the development of web-portal user interfaces, where expertise needs to be developed as suggested in other recommendations. There are many forms that such a lead agency activity could take, ranging from interagency oversight that utilized the services of existing agencies and private entities, to taking on operational responsibility for monitoring. The committee did not explore these details in any depth.

Well-designed observatories, demonstration projects, test beds, and field campaigns have been and can continue to be effective for testing integrated observational-modeling systems to address research questions and operational needs. Many of the case studies presented in Chapter 4 clearly articulated the need for hydrologic observatories to demonstrate how the advances in embedded sensors can be integrated with traditional sensors, models, and remote sensing to provide new hydrologic process understanding. In fact, the successes achieved by the current experimental watersheds argue for more such projects.

Recommendation 2-1: Coordinated and jointly funded opportunities for observatories, demonstration projects, test beds, and field campaigns should be significantly increased.

The observatories provide mechanisms to improve knowledge of complex environmental systems and processes through improved sensor information and modeling. It is hoped that the observatories would have academic and government research scientists working together on problems that lead to enhanced management approaches by the agencies.

Generally, the funding of observatories, test beds, and field campaigns at the “integrated field demonstration” scale is challenging, because of the relative lack of funding at a scale greater than the individual research project and smaller than programs like NSF’s Major Research Equipment and Facilities Construction (MREFC) projects. This is true even considering that costs can often be kept reasonable by encompassing projects within existing networks when appropriate to the research question. Likewise, most grant horizons and federal funding cycles (1-5 years) are inappropriately short for interdisciplinary, integrative projects.

Recommendation 2-2: Agencies should consider offering new funding streams for projects at the scale of several million dollars per year for approximately 5-10 years to help close the gap between sensor demonstration and integrated field demonstration.

These projects could be “problem-centric” with a significant research question or management goal. An integrated hydrologic observations network, as envisioned in this report, would be designed and deployed as an integral part of the project, and used to address the project goals and objectives. This class of project would help fill the gap between sensor development and operational use of established sensors. It could also be a critical demonstration mechanism to provide feedback to the more fundamental sensor development activities and offer results to guide operational applications of integrated sensors and cyberinfrastructure.

The problem posed in the Chapter 4 case study on water and malaria in sub-Saharan Africa illustrates this class of project. Its research question requires integration of data from a variety of sensor systems measuring quite different types of variables (hydrologic, biologic, social, etc.) at differing scales and interpreted by several disciplines due to the interdisciplinary nature of the problem. The research team achieves this integration through the use of a “collaboratory,” that is, a web-based system where different researchers and users can come together to build a system of data, predictive models, and management projects. A collaboratory develops and evolves over time, adapting to new data and modeling systems and the needs of the users of the systems projects, such as seasonal forecasts of malaria intensity in the case study.

Another possible model would mix support for observatories and test beds with programs with shorter timescales and Integrative Graduate Education and Research Traineeships (IGERTs); a third would be an analogy to the NSF Engineering Research Centers, which have a 10-year time horizon. All of these funding models would have as a goal the field demonstration of integrated sensors (across variables and platforms) further integrated with cyberinfrastructure development bridging the critical gap of single sensor demonstration to integrated field demonstration.

As noted earlier, a review of, and recommendations concerning, the development of new airborne and spaceborne sensors by NASA, known as the “decadal survey” (NRC, 2007) has recently been completed. The recommendations of the survey for specific water-related missions are summarized in Appendix C. However, another important recommendation of the decadal survey was that “NASA should support Earth science research via suborbital platforms: airborne programs, which have suffered substantial diminution, should be restored…”

As shown in the Arctic case study in Chapter 4, certain remotely sensed measurements are more useful from airborne platforms than from spaceborne platforms, and the capability of low-cost airborne remote sensing needs to be further developed and integrated with other measurement systems.

In recommendation 2-1, the report discusses the interagency gaps from sensor development to operational deployment. This recommendation is analogous but focused on NASA. NASA is unique in that it has research and development programs that span from sensor development to operational space deployment, often with other agencies like NOAA or the Department of Defense. Historically NASA demonstrated sensors through a strong airborne program, carried out “integrated field demonstrations” through validation campaigns related to Earth Observing System (EOS) missions and Earth System Science Pathfinder (ESSP) exploratory missions that often included in-situ and airborne measurements, and supported research and application in the use of space observations.

Unfortunately, the airborne remote sensing component within NASA has been weakened for both piloted and unpiloted systems, causing a gap in the sensor development-to-operational deployment road. This weakness is being felt in the sensor development activities because there is no obvious mechanism for easily and inexpensively testing new sensor concepts. This weakness may also extend to operational deployments because the possibility that new sensors, or the operational products from new sensor measurements, may be best delivered through airborne systems rather than through spaceborne systems that cannot be evaluated effectively.

Recommendation 3-1: NASA should strengthen its program in sensor technology research and development, including piloted and unpiloted airborne sensor deployment for testing new sensors and as a platform for collecting and transmitting data useful for applications.

In the current federal budget environment NASA has constrained resources. The space agencies in other countries have similar pressures. Historically there have been joint programs between NASA and international space agencies in developing and launching specific sensors. Two examples are the Tropical Rainfall Measurement Mission (TRMM) with the Japanese Aerospace Explora-

tion Agency (JAXA) and the Calipso satellite with the French space agency, Centre National d’Études Spatiales (CNES).

Regardless of the above collaborations, NASA can do more, particularly in the areas of hydrologic and environmental sciences, to foster both U.S. interagency and international partnerships.

Recommendation 3-2: In addition to partnerships with other federal agencies for the development and testing of experimental sensors that are of a particular interest to agencies, the Nation, and especially NASA, should explore additional strategic partnerships with space agencies in other countries and regions, such as the European Space Agency (ESA), JAXA, CNES, and the Canadian Space Agency (CSA).

Assuming a strengthening of NASA’s sensor development programs as recommended above, there is an opportunity for NASA to take a leadership role in developing integrated ground-to-space sensor networks as envisioned in Chapters 2 and 3 of this report. Some of the necessary coordination could be done through the Office of Science and Technology Policy’s Subcommittee on Water Availability and Quality (SWAQ), as proposed by the NRC (2006c), but this should be only one of many points of contact. This is an area in which an increase in support by NASA could be highly leveraged by other agencies. The decadal survey (NRC, 2007) recommended that NASA increase support of its Research and Analysis (R&A) program, noting that among other purposes the R&A program is “necessary for improving calibrations and evaluating the limits of both remote and in situ data.” Engagement in observatory design would also be consistent with NASA’s Earth Science Implementation Plan for Energy and Water Cycle Research (NASA, 2007), which notes that “[i]n some cases, NASA investments may be required to supplement [planned] activities to ensure that they meet specific needs, for example, in situ measurements of parameters that are essential to validating space based remote sensing, as well as quantities needed but not otherwise measured or derived.”

Recommendation 4-1: NASA and NOAA should work with NSF and other agencies to assure that plans for incorporation of space-based and airborne observations (from both existing and, preferably, planned or proposed missions) are part and parcel of the experimental design of these proposed observatories.

Cyberinfrastructure is a key linkage between observatories, and a way to achieve economies of scale within and between multidisciplinary initiatives. Yet there is insufficient work in exploiting current advances in cyberinfrastructure. The underfunding of cyberinfrastructure has been shown to compromise results and data interpretation at all scales. Advances in computational, communication, and Internet technologies will help develop and deliver “services” to the broad community of users.

The observatories and other national initiatives offer settings and mechanisms for the development and testing of the “next generation” service delivery system. For example, current NOAA plans for the newly funded National Integrated Drought Information System (NIDIS) include the development of a “drought portal” for the delivery of drought information.

The guiding principles important to a next generation system were set forth in Chapter 3. These included employing open architecture solutions to enable the rapid adoption of new hardware and software technologies and nonproprietary and, ideally, open-source software solutions (e.g., middleware, metadata management protocols) and promote the modularity, extensibility, scalability, and security that are needed for observation and forecasting. They also included using community standards (e.g., data transport, quality assurance/quality control, metadata specifications, interface operations) to ensure system interoperability, and providing open access to data and information through customizable portals to ensure timely access to data, information, and forecasts.

Recommendation 5-1: Advanced cyberinfrastructure should not only be incorporated as part of planned observatories and related initiatives to help manage, understand, and use diverse data sets, but should be a central component in their planning and design.

Recommendation 5-2: Utilization of web-based services, such as collaboratories, for the distribution of observations, model predictions, and related products to potential users, should be encouraged.

To achieve efficiencies and economy of scale, it may be useful to create an interobservatory council among at least the NEON and WATERS communities to address interoperability issues and agree on a uniform cyberinfrastructure. An important step forward has been taken by WATERS, in the form of a recently funded “Test-bed Digital Observatory” for the Susquehanna River Basin and Chesapeake Bay, which will attempt to demonstrate the applicability and utility of many aspects of the Hydrologic Information System (HIS) from CUAHSI and the “CyberCollaboratory” from Collaborative Large-Scale Engineering Analysis Network for Environmental Research (CLEANER) for use by both the hydrologic and water-quality communities.

The NRC (2006c) went one step beyond an interobservatory council and recommended that “[s]erious consideration should be given to placing the various NSF environmental observatory programs under a parent organization that could be termed something like the ‘Environmental Observatory Networks’ or EON.’ Such a parent entity would be responsible for cyberinfrastructure development, among other shared activities.” This approach also seems reasonable.

Water and environmental data products from the observatories can be extremely useful for the classrooms of the nation. An exploratory effort to do this—the Student Analysis of Data Driving Learning about the Earth (SADDLE) project—is creating student-friendly interfaces to allow students to interact with real-time and archived databases such as would be associated with a large-scale scientific observatory. There is a need for various studies of this kind to understand how observatory-derived data can be delivered and used in classrooms at various educational levels. It should be noted that the parent organization EON, proposed by the NRC (2006c), would also have a role in coordinating educational outreach by the various proposed NSF environmental observatories.

As discussed in a number of case studies, there is an unmet need to have sensor systems that link plot-scale, in-situ measurements to measurements at satellite scales, and the related data-assimilation systems that merge these data with appropriate predictive models. It is recognized that both NASA and NSF have research programs that support data assimilation. However, these ongoing programs are not sufficient.

Recommendation 6-1: NASA and NSF should develop and strengthen program elements focused on demonstration projects and application of data assimilation in operational settings where researchers work collaboratively with operational agencies.

This recommendation is consistent with the decadal survey (NRC, 2007) recommendation that “NASA, NOAA, and USGS should increase their support for Earth system modeling, including provision of high-performance computing facilities and support for scientists working in the areas of modeling and data assimilation.” It is also consistent with NASA’s implementation plan for water and energy cycle research (NASA, 2007), which includes “Creating a global land and atmosphere data assimilation system for energy and water variables” among its major elements. Existing data sets should be fully exploited in these efforts.

Many of the above recommendations have been especially applicable to observatories, experimental watersheds, and other kinds of fairly focused work. But it is important to look beyond observatories to other, broader, downstream applications of observations.

Chapter 2 makes clear that the Nation has invested significantly in remote sensing, especially in space observations. However, agencies such as NASA and NOAA need to foster activities that will help state agencies and other federal agencies utilize and merge remote sensing measurements with process-based hydrological models for improved management decisions. The availability through the NASA Distributed Active Archive Centers (DAACs) of data from the recent suite of Earth Observations System satellite sensors is significant, and Advanced Very High Resolution Radiometer (AVHRR) and Landsat Thematic Mapper (TM) data are widely used. Yet the case studies in Chapter 4, and in other reports, indicate that remote sensing data and data products are often underutilized in operational and application uses. NASA and NOAA have often encountered difficulties in “Crossing the Valley of Death” (NRC, 2000b) from research to operations. This has often resulted in underutilization of data and data products, resulting in long delays in users and agencies realizing the value of such measurements to society.

Also, the process for identifying the remotely sensed measurement needs of the community and federal agencies, and for the subsequent technological development of such sensors, is often unclear. One example of need is water-quality measurements from space, where only the most basic water-quality variables, turbidity and chlorophyll, have been measured and only as an experimental variable. The scientific basis for water-quality measurements from space needs considerable development, and it does not appear that any NASA program is addressing this problem.

Integrated data products that merge satellite, airborne, and ground observations with model and decision-support products at regional to continental scales, would be of particular value. Demonstration projects at observatories, including remote sensing validation observatories, may be an effective mechanism to demonstrate to users of water cycle and environmental data products how they may better exploit the current investment in space observations in their prediction and management activities.

Recommendation 7-1: NASA should take the lead by expanding support for the application of integrated satellite remote sensing data products. NSF, NOAA, and other federal and state agencies engaged in environmental sensing should likewise expand support for the creation of the integrated digital products that meet educational, modeling, and decision-support needs.

Utilization of remote sensing data in applications and operational settings also depends on assurance by the government to users that these measurements will be maintained into the future so their investment in developing procedures that use the data will be realized. Except for atmospheric remote sensing, which NASA develops for NOAA and Department of Defense to use operationally, and visible wavelength surface imagers (AVHRR and TM), all other NASA-developed sensors are viewed as limited-duration experimental sensors. Transitioning sensors from a research to an operational monitoring mode is critical if users and agencies are to utilize their products as part of their management responsibilities.

However, as yet there is no government strategy for such a transition beyond those sensors used for weather. Current concerns about the elimination on the new NASA Landsat satellite of the infrared (IR) channels, which are used by USDA and state agencies (among others) to monitor irrigation water usage, is a case in point. The elimination is due to budgetary issues and lack of consensus over who will pay for the sensor. Similar issues are occurring on the planned National Polar Orbiter Environmental Satellite System (NPOESS). This is one important impediment for wider usage of remote sensing measurements in applications and by operational agencies.

Recommendation 7-2: Congress, through the budgetary process, should develop a strategy for transitioning NASA experimental satellite sensors to operational systems with assured data continuity so that the Nation’s investment in remote sensing can be utilized over the long term by other federal agencies and users.

A wide variety of water agencies have a need for accurate and timely information from in-situ, airborne, and spaceborne sensors. These agencies may be involved in anything from water supply forecasting to flood forecasting to transportation and recreation. They are often underfunded, with a very modest number of trained scientists and engineers. Because of this, once they develop a system—even an imperfect one—that provides reasonably good information on which to make their decisions, there is strong aversion to new kinds of data, models, and tools that have less of a track record.

Adopting new systems involves start-up costs in the way of model design and calibration, and staff training. Users who are adapted to a given product will likely need to be educated in interpretation and use of the new product in order to assist them through the transitional period. Further, until a new remote sensing system is fully operational with a firm commitment of continued service from the federal agency responsible for it, a local or state water agency may be very conservative about adopting the new technology.

Nonetheless, there are numerous opportunities for both existing and new sensor data and information to be incorporated into planning and forecasting, and such opportunities will only increase in the future.

Recommendation 8-1: Water agencies should be alert for opportunities to incorporate new sensor and modeling technologies that will allow them to better deliver their mission.