Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
4 Implementing Environmental Observatories The need for better understanding and management of the nation's environmental resources over past decades has led to the establishment of several federal, regional, state, and local monitoring and assessment programs. The National Oceanic and Atmospheric Administration's (NOAA) weather and ocean/Great Lakes monitoring, the U.S. Geological Survey's (USGS) national hydrologic monitoring and water quality assessment programs, the U.S. Department of Agriculture (USDA) Forest Service (FS) and Agricultural Research Service (ARS) experimental watershed facilities, and the Environmental Protection Agency's (EPA) Clean Air and Clean Water Act monitoring are examples of federal activities with practical or regulatory drivers. These programs, along with regional, state, and local monitoring and assessment activities, provide data useful to many fields of science and engineering, and contribute to the understanding of complex environmental phenomena. These monitoring and assessment activities provide valuable information to research communities and management agencies. However, there remains a critical need for more integrated and comprehensive approaches to understanding and analyzing environmental processes that ultimately will help us better manage and improve the quality of air, land, and water resources. This need has motivated the planning of environmental observatory and research networks by the National Science Foundation (NSF). These initiatives are at different stages of planning and implementation and in many respects have been developed by separate scientific constituencies and separate directorates within the NSF. No single environmental observatory initiative as currently formulated will be sufficient to provide the integrated data sets, models, and predictive capability necessary to adequately understand and guide effective management of our nation's environmental resources in a setting where large scale, even global factors, must be considered. As a coordinated group, however, they might be. Planning has been underway for the National Ecological Observatory Network (NEON) after several NSF workshops in 2000 addressed the needs for long-term ecological observatories. NEON will focus on large scale and interdisciplinary analysis of ecological systems (NRC, 2004a). The Ocean Observatory Initiative (OOI) is another NSF environmental observatory system at approximately the same stage of development as NEON; it is being designed to help better characterize processes occurring in the ocean. The Geosciences Network (GEON) grew from a series of workshops in 1999 and seeks to advance the field of geoinformatics in support of the geosciences. A national hydrologic observatory network that proposes to address atmosphere, subsurface, and surface dynamics of water movement is in the beginning stages under the leadership of the Consortium of Universities for the Advancement of Hydrologic Science, Incorporated (CUAHSI). NSF's Office of Polar Programs is also working to design a land/atmosphere/ocean-based Arctic Observing 34
Implementing Environmental Observatories 35 Network that would collect, check, organize, and distribute arctic observations. Finally, CLEANER is a proposed network through which problems involving large-scale complex engineered systems and human-stressed environments would be addressed. As noted earlier, during the course of this study, the NSF was considering and planning for the merger of the CUAHSI Hydrologic Observatories and CLEANER. These national environmental observatories are now being planned, each with a somewhat different focus area and each designed to meet somewhat different goals. Yet each observatory has a core focus on transformative developments in the environmental sciences. Because of the differences in the stages of development, these observatory programs will be ready for implementation at various times over the coming years. Even so, it makes sense to think about their coordination to take advantage of existing expertise, experience, and technology, and to avoid duplication of expensive infrastructure, including cyberinfrastructure. COORDINATION OF ENVIRONMENTAL OBSERVATORIES Coordination of Environmental Observatories within NSF Because these individual NSF initiatives have been developed somewhat independently, there is danger of unnecessary and inefficient redundancy in developing data collection standards and infrastructure and in performing various research and associated activities. To effectively observe complex environmental systems and inform their management, a coordinated, administratively integrated effort will be necessary. Indeed, there is a danger that lack of coordination among different observatory designs will make the sharing of data, information technology, and education and outreach activities more difficult if not impossible across programs. As these observatory initiatives develop, communication and close coordination among the programs are essential. It is apparent that the various environmental observatory projects currently in the planning process at the NSF have some similar basic components and needs. These needs include the development of a cyberinfrastructure, coordinated educational activities and outreach, and the co-location of some observatory elements to achieve needed data integration at lower costs. It makes scientific and economic sense to share infrastructure/capital that will meet common needs. The scientific challenges to be addressed by each of the initiatives clearly require an integrated effort across the relevant disciplines (i.e., engineering, ecology, hydrology, earth sciences). For instance, the exchange of materials between terrestrial and aquatic ecosystems should be of interest to both NEON and CLEANER, which further suggests coordination with existing agency-based observatories and environmental assessment programs.
36 CLEANER and NSF's Environmental Observatories The NSF appears to recognize these issues and ongoing efforts to coordinate the observatory initiatives are currently underway. The NSF program officers leading the development of these initiatives have issued joint calls for proposals in the areas of sensor development and cyberinfrastructure that will benefit each of the observatory initiatives. Even more coordination and possibly joint management of these initiatives may be desirable as the programs reach implementation. Serious 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. This parent entity would be responsible for cyberinfrastructure development, educational activities, outreach, and other shared activities across the observatory programs. This entity would facilitate collaboration and coordination among programs and minimize redundancy, while not impeding the progress of individual observatory initiatives. It would also help ensure that the observatory networks are well integrated spatially and across relevant landscape units. An independent expert committee outside of the NSF could provide advice and oversight on the research and support activities of EON. Again, this integration of observatories would contribute to data integration and reduce infrastructure redundancy and associated costs. Coordination of Observatories with Other Agencies and Stakeholders Many federal and state agencies have programs that support water resources research, monitoring, and management through in-house capabilities or by the distribution of grants. Federal organizations having significant water resources research programs include the NSF, USGS, U.S. Fish and Wildlife Service, USDA, EPA, Department of Defense, Bureau of Reclamation, Department of Energy, National Aeronautics and Space Administration, and NOAA (NRC, 2004b). The proposed NSF observatories will need to document their value in the context of existing state, regional, and federal programs that conduct similar types of investigations. Observatories should focus on filling scientific and engineering gaps identified by researchers at universities and managers of programs administered by the government agencies. The new observatory infrastructure needs to be designed to enhance existing infrastructure, if such infrastructure exists, without significant duplication. Expertise in the federal, regional, and state agencies on conducting watershed-based field studies, modeling, analyzing geospatial information, making ecological assessments, working with sensors, addressing water quality issues such as nutrients, toxics, pathogens and invasive species, and managing databases should be of value to investigations at CLEANER observatories. Close cooperation should result in mutual benefits and cost-savings. Cooperation with federal, regional, and state agencies needs to be established at several levels that include: (1) coordination of data collection, (2)
Implementing Environmental Observatories 37 management of data, (3) sharing of techniques such as modeling capabilities and analytical methods, and (4) coordination of activities at sites where there are common interests between an agency program and a CLEANER observatory. The two major national water databases are the National Water Information System (NWIS; managed by the USGS) and STORET (short for storage and retrieval; managed by EPA). Both databases are available on the Internet, although they are not now compatible through one portal. Recent work funded through CUAHSI for its Hydrologic Information System (HIS) has advanced the development of a common portal to access water databases from different agencies. If ultimately successful, it will be a major achievement for the retrieval of water resource information. The NSF has proposed to have the data generated through all NSF-funded observatories available through a cyberinfrastructure. Field-focused activities that are most likely to have common interests with sites selected for investigation by CLEANER are those managed by the USGS, the USDA ARS and FS, EPA, and the NSF Long-Term Ecological Research (LTER) programs. The USGS has several field-based programs, including the National Water Quality Assessment (NAWQA) program that focuses on watersheds and major aquifers to assess the status of, trends in, and causes of changes to the quality of the nation's groundwater and surface waters; the Toxic Substances Hydrology Program, through which long-term research is conducted at sites that represent different types of contamination, and a long-term small watershed program (Water, Energy, and Biogeochemical Budgets (WEBB) Program) at five sites, which focuses on processes in different climate regimes. The USDA supports watershed research through the ARS, FS, and Natural Resource Conservation Service. The ARS watershed network is a nested, multi- scale network designed to assess the hydrologic impacts of watershed conservation and management practices. The USDA FS supports experimental forest and range research sites. Nine of these sites focus on watershed research. The NSF LTER Network is designed to address long-term research to understand ecological phenomena that occur over long temporal and broad spatial scales. The LTER Network involves long-term ecological measurements and experiments at 26 sites in the United States and Antarctica, some of which are in human-stressed environments. The EPA conducts water-related research in its Office of Water and Office of Research and Development. EPA's Regional Environmental Monitoring Assessment Program (REMAP) is a research initiative aimed at helping to monitor and assess the status and trends of national ecological resources, including water quality. Also the EPA supports site-specific investigations such as in the Chesapeake Bay and Great Lakes. CLEANER is different from most of these other programs because it offers a large-scale focus on human-stressed environmental-hydrologic systems from the engineering perspective of identifying and evaluating innovative approaches to the solution of environmental problems. There is also a need to coordinate with non-governmental organizations that are actively involved in addressing environmental issues in many areas of the
38 CLEANER and NSF's Environmental Observatories country. In many major river basins, a variety of non-governmental organizations (i.e., the Nature Conservancy, the Wetlands Initiative, Environmental Defense) are working with various stakeholders to develop solutions to problems in the basins. The observatory networks should coordinate with these efforts to facilitate the transfer of applicable observations and analysis to address relevant problems. Mechanisms to Achieve Coordination and Cooperation Several measures can be implemented to help achieve successful coordination of the new observatories with existing federal, regional, and state programs. Activities that can facilitate coordination include: (1) workshops on specific topics, (2) information requests in proposals about existing infrastructure funded at the federal, regional, and state levels, (3) establishment of agreements with federal, regional, and state agencies, and (4) dialog in committees that have representatives from government agencies. Workshops on specific topics can enhance scientific exchange and encourage coordination and collaboration. Workshops where parties, including university and government scientists and managers, work together to find solutions and identify directions are often more effective in establishing communication than are briefings that tend to become "show and tell" exchanges. The most successful collaboration is often achieved at the scientist- to-scientist level. Workshops should be designed to encourage this type of interactive exchange. By encouraging participation of others, the observatory networks can benefit by sharing costs and avoiding duplication of efforts. Another measure to encourage collaboration with federal, regional, and state agencies is to require in the planning process that the existing infrastructure at a potential site be identified and that the agency supporting the infrastructure be notified and their goals documented. A solicitation of letters of support from the appropriate agency managers or scientists can be encouraged. This process is currently in place in several NSF programs. Collaborative agreements between the observatories and government agencies will provide mechanisms for sharing resources. For example, the USGS and CUAHSI have an agreement, as of 2006, in regard to the USGS Hydrologic Instrumentation Facility (HIF) that covers the rental of equipment, such as that used for gauging streams. The agreement will save the observatories from having to purchase all their instrumentation and develop facilities to test, repair, calibrate, and house it. The instrumentation is located at the HIF and is used by the USGS for its investigations and thus is cost neutral to the USGS. One approach to establishing cooperation with the federal agencies that have water resource research programs may be for the EON coordinator to interact with the Subcommittee on Water Availability and Quality (SWAQ) that was formed by the Office of Science and Technology, Executive Office of the
Implementing Environmental Observatories 39 President. The SWAQ, created in 2003 initially for five years, has a membership of 18 federal agencies to "advise and assist the Committee on Environment and Natural Resources and National Science and Technology Council on policies, procedures, plans, issues, scientific developments, and research needs related to the availability and quality of water resources of the United States" (SWAQ, 2004). Although the future of SWAQ is uncertain, the NSF is a member of the subcommittee, and it is an appropriate group with which to exchange information and establish coordination. CLEANER CYBERINFRASTRUCTURE ISSUES National-scale environmental observing initiatives will need to coordinate and share resources to provide quality data products for accurate scientific analysis and forecasting. The scale of the scientific enterprise and the fact that there are numerous participating communities, agencies, and programs precludes a single home-grown bottom-up approach to observing system development and operations. Operating across distinct jurisdictional boundaries, with disparate policies and technologies, raises a number of challenges--both in infrastructure and operations. By considering these challenges during the planning of cyberinfrastructure initiatives associated with the environmental observatories, there may be some hope of achieving a high-level of system compatibility and integration that will benefit both the quality of the science and the economics of system development and operation. Cyberinfrastructure for environmental observing systems includes hardware (i.e., sensors, actuators, networks, computers, storage platforms), software (i.e., middleware that supports different application components, tools, applications), and the standards and policies necessary for operations. Given the extreme heterogeneity of the technology base, the role of cyberinfrastructure in facilitating system integration will be paramount. A robust cyberinfrastructure can provide common frameworks, components, modules, and interface models that can be used in multiple observatories or applications. There are several cyberinfrastructure activities underway that address aspects of the resource integration challenge. For example, CUASHI has initiated a Hydrologic Information System (HIS) project to integrate access to national water resources data, and there are plans to leverage this cyberinfrastructure for NEON. Extending this activity to include CLEANER, as seems likely given the merger of CLEANER and CUAHSI's Hydrologic Observatories, is one example of the types of cross-organizational collaborations that would facilitate the vision of an integrated EON. In the following sections a few of the key issues are discussed related to creating a common cyberinfrastructure for environmental observatories: (1) policies--rules and guidelines for operations, (2) standards for data, networks, and instruments, and (3) the hardware and software of the cyberinfrastructure. A comprehensive approach to building an efficient and secure, scalable and
40 CLEANER and NSF's Environmental Observatories extensible, environmental observing system will require careful design across all three areas. Policies Many of the challenging issues involved in integrating observing system resources center on governance policies. Who owns the resources (including data and physical resources--sensors, instruments, computing platforms, etc.)? Who can use the resources and under what conditions (including access control, resource contention and scheduling, intellectual property rights, publication, and citations)? Often these policies vary by community or project (e.g., requirements for timely public access to data products). Privacy policies and public disclosure of sensitive data (e.g., wetlands designations) also factor into information sharing activities and applications across observing systems and communities. The current practice in negotiating cross-agency (or cross-project) resource-sharing policies is typically limited in scope and heavily influenced by a small number of participants. A more comprehensive approach would benefit multiple observing system initiatives or the multiple users of a common environmental observatory network. Policies will be needed regarding the access and use of a variety of resources beyond data products, including the field-deployed observing system sensors and instruments, and the data center resources for data management, modeling, and analysis. A services-oriented architecture (see below) provides the engineering framework for building distributed resource-sharing applications; however, it is the policy specifications that will determine the efficiency and usability of such applications. Standards Standards with respect to observatory hardware, software, and research activities, including data collection, documentation, and storage, can facilitate cross-program applications. Standards can also constrain scientists and engineers, especially in an environment of rapid changes in technology. The challenge is to develop standards that can both facilitate multi-disciplinary activities and be adaptable to changing technologies and needs. Especially important are standards for providing metadata descriptors for key resources and infrastructure components. Metadata perform a number of roles in observatory network design and operation, including facilitating application integration and locating, evaluating and using data products, sensors, communication and computer networks, software applications, and the like. For example the ocean observing community's Marine Metadata Initiative1 and the CUAHSI HIS metadata 1http://marinemetadata.org/.
Implementing Environmental Observatories 41 initiative have direct relevance to the CLEANER program. A common set of metadata specifications and tools would facilitate cross-observatory activities. Most of the observing system standards to date have focused on data collection and products. Similar standards need to be specified for other resources including sensors, instruments, and actuators. The ocean observing initiatives (e.g., the Laboratory for the Ocean Observatory Knowledge Integration Grid [LOOKING]2) have been exploring these issues, drawing on evolving standards. Tools and Applications Observing system cyberinfrastructure must provide the tools to implement polices and deliver services. The cyberinfrastructure goals include: 1. Interoperability--the ability to build applications across domains; 2. Extensibility--the ability to add new devices, components, tools, and applications; and 3. Scalability--the ability to add new computational resources to handle increasing demands. These goals can be accomplished by developing: (1) an open architecture based on well-defined standards, (2) a common framework for application development and integration, and (3) an integrated suite of tools for resource discovery, access, and use. The requirement for an open architecture derives from the fact that there are numerous heterogeneous infrastructure components that are needed to satisfy the overall scientific mission. It is not reasonable to presume (or mandate) a particular technology product (e.g., database, geographic information systems (GIS), workflow package, analysis toolbox). In addition, observation system resources come with differences in protocols and performance. For example, there exists a large variety of sensor hardware platforms and software elements (i.e., operating systems, networking protocols, data base systems). Commercial off-the-shelf (COTS) sensors are equipped with transducers for measuring variables ranging from pressure, temperature, humidity, and light, to various complex environmental phenomena (e.g., dissolved oxygen, nutrient concentrations, current profiles). In addition, new generations of biological sensors are being designed and developed. The processing power of available COTS sensors ranges from 4 MHz to 400 MHz. Their radio bandwidth varies from few kilobits/sec to several Megabytes/sec. Typical operating systems available include TinyOS, MOS, Linux, and 2www.lookingtosea.org.
42 CLEANER and NSF's Environmental Observatories Windows CE. There are similar differences in the data center components and computer programming applications (e.g., NET vs. Java web service stacks). Accommodating these heterogeneous components (including legacy resources and researcher/developer preferences [e.g., favorite tools]) dictate an extensible architecture. A commitment to an open architecture is not a commitment to open-source software products. Rather, it is an open set of standards and protocols that specify the functional interfaces to various devices, products, and services. Data Availability and Sharing This committee also recommends the selection of an open source standard for data collection, storage, and sharing. There are two major advantages. First, the data standard would support data sharing and analysis opportunities between CLEANER locations. The more consistent the data protocols, the more effectively the data can be shared among multiple CLEANER stations to examine common trends and responses. Second, definition of an open standard can more effectively expand the influence of the CLEANER program. Not all data will be collected by the CLEANER program lead. Historic data bases might need some conversion and migration to a common platform to provide a basis for research. Even more important, other researchers, federal, regional, state, and local government agencies, and institutes will continue to collect information in or near the CLEANER site areas. Each data collection effort provides an opportunity to expand the scope of the CLEANER data set. The CLEANER program site leads might not have the authority to dictate everyone's data collection protocols within their geographic footprint. However, the availability of an open source standard, similar to those provided for software developers, can encourage and promote consistency in data collection. The open source standard could be provided and disseminated under the auspices of the CLEANER program. The CLEANER network of data sets should be available to all researchers with adequate recognition of those who collected them. Data should be widely available and be shared among the research community. Open standards and data sharing can be critical to the ultimate success and long term success of CLEANER. The current cyberinfrastructure approach to achieving application integration is services-oriented architecture design, building on standards and tools developed from the web/grid services communities (e.g., the World Wide Web Consortium3). This approach is being developed for both NEON and CUAHSI HIS. A focused effort to coordinate these activities could play a significant role in building a common cyberinfrastructure platform for integrating CLEANER (including the CUAHSI HIS), and NEON (and others), i.e., in building EON. 3http://w3c.org.
Implementing Environmental Observatories 43 DISSEMINATION, OUTREACH, AND EDUCATION Large-scale national observatories to assess the status of environmental resources provide opportunities to better understand environmental systems and natural and human-induced perturbations to these systems. This increased understanding should ultimately help water resource managers, for example, forecast the conditions of environmental resources much like the National Weather Service provides weather forecasts. Environmental forecasts might be the status of beach areas due to concerns over pathogens or nuisance algae blooms, adverse air quality events in urban areas (e.g., elevated ozone, particulate matter), turbidity events in water supplies, or the impacts of combined sewer overflows in waters adjacent to urban areas. An improved forecasting capability should motivate increased interaction and communication with resource managers and engagement with the public about environmental resources and problems. This interaction and communication can benefit from improved programs in environmental education and outreach. If CLEANER and the other environmental observatory initiatives fail to transform environmental education and outreach, then they will fail to meet their full potential. The observatory initiative has the potential to make knowledge of the status of environmental resources a basic component of day- to-day life. Given the level of investment and necessary commitment of funds and time needed to design, implement and operate the environmental observatories, it would seem critical that this large-scale science and engineering initiative should be incorporated into science and engineering curricula at universities. Such an educational initiative would seem to be necessary to train the next generation of scientists and engineers on how to effectively use and to continue to improve the environmental observing systems. Furthermore, these networks will be able to convey and animate patterns and processes at spatial and temporal scales not currently possible. This, together with new sensors and other technologies associated with the environmental observatories, just might capture the imagination of younger scientists and help attract them to the fields of environmental engineering and science. SOME NECESSARY CONDITIONS The benefits derived from environmental observatories can only be realized if these observatories are a long-term undertaking. While planning efforts continue, it is important to consider and avoid pitfalls from such an undertaking. The potential benefits have been outlined in the previous chapters. Here we discuss possible "fatal flaws" to avoid should these environmental observatories be implemented. Prior to implementing the environmental observatory network program(s), it is critical to identify through scenario planning any flaw that
44 CLEANER and NSF's Environmental Observatories would make untenable all or part of the program. Examples of fatal flaws for this observatory investment might include: · lack of sufficient funding to create a network on a scale that is needed and resources to maintain and upgrade it over time; · software security failure resulting in data loss or major delays in research; · process failure in selecting research challenges; · inappropriate location and scale of observatories (i.e., critical gaps in geographical coverage, inefficient use of resources on large, complex systems); · loss of public and political support due to inability to connect research results with required actions and desired outcomes; and · inability to train enough engineers and scientists with the interdisciplinary breadth required to operate CLEANER and address the future research challenges of CLEANER. Fatal flaw analyses of possible scenarios such as these can help identify potential "surprises" that could reduce the benefits coming from environmental observatories. Such analyses can also help identify activities which if undertaken might reduce the likelihood of these possible fatal flaws occurring. For clarity, each of the above possible fatal flaw concerns are briefly discussed below. Lack of Sufficient Funding Given the short-term pressures on the availability of funds, lack of sufficient money to maintain these environmental observatories and the databases through the long-term is a realistic concern (Merali and Giles, 2005). The research community needs to be committed to the environmental observatories so that there is a willingness to use programmatic funds, if necessary, for their long-term maintenance and operation. Leveraging or cost- sharing strategies may be necessary to better ensure both short- and long-term funding. Every step must be taken to assure the public's return on the capital investment by securing operating funds sufficient for long periods as part of the capital plan. Expensive observing systems could quickly become useless if operating and maintenance monies become insufficient or unavailable to continue those observations on into the long-term future. Also of potential concern is the source of such operating and maintenance funding. If operating and maintenance monies must come from the research budget of NSF normally available to individual investigators, it becomes a
Implementing Environmental Observatories 45 tradeoff between the benefits derived from CLEANER and those derived by individual scientists and engineers pursuing other research topics. Younger professors in particular are vulnerable to any reductions in research funding that is essential for graduate student support and hence to their professional careers. If maintaining and upgrading the CLEANER environmental observatory network takes away funds from individual research programs, will the gain in understanding from CLEANER be worth any future loss to other environmental science and engineering educational and research programs not directly tied to CLEANER? Software Storage and Security The success of the observatory approach depends on reliable long-term storage and management of time-series and other data. Plans for creating and maintaining databases should consider the need to port databases to new hardware as it is developed in the future and as current technology becomes obsolete. It is critical that collaborators cooperate, share data, and make timely entries into applicable databases. Quality assurance information should be incorporated into the database maintenance program so proper interpretation of observations is ensured. Another challenge is to maintain the accessibility and flexibility of ever-increasing data and databases as well as the interfaces that maximize their use. The security of these data is critical to the identification of long-term trends, yet all databases need to be accessible to multiple users at multiple sites. Process Failure for Selecting Research Challenges Selection of what to observe is an important issue. Failure to measure, collect, and store the values of priority environmental parameters will restrict future research opportunities. Considerable attention needs to be given to the proper identification and selection of the data to be obtained over space and time to address the important current and future environmental resource and ecosystem management issues. In this regard it is critical that all relevant disciplines be active participants in CLEANER (e.g., geochemistry, environmental engineering, hydrology, microbiology, geomorphology, social sciences) to ensure that there will be a comprehensive examination of the problems that may need to be addressed and their data requirements. It is always a challenge to know what data (that may be of little or no significance today) we should be currently collecting for the benefit of researchers and resource managers in the distant future.
46 CLEANER and NSF's Environmental Observatories Mis-location of Observatories, Gaps in Geographical Coverage, and Appropriate Scale for Site Selection We must be certain that observatories are put in the "right" places and that resources are used optimally. In a program such as CLEANER that focuses so fundamentally on water and human interactions with our environment, geographic coverage means that hydrology, and specifically water flows, may determine to a large extent the geographic coverage of observatories. Large river basins should be a focus of special emphasis for CLEANER because of their impact on surface water flows, water quality, and ecosystem functioning. (Furthermore, other observatory programs are not including them.) For example, an important area might be a large agricultural region where groundwater flows are critical to its economy. Major metropolitan areas are also candidates for better understanding and predicting the environmental-human interactions in our built environments. While it is likely that CLEANER will focus on large aquatic systems, resources may not be adequate to study very large, complex systems. There are important tradeoffs in site design and scale relative to available resources that the principal investigators of CLEANER will need to evaluate. Following the development of an important question to be addressed, a system should be selected for study that is large enough to address that question, but not too large to investigate effectively without wasting resources. Another tradeoff is should one region be selected to address all questions of importance, or should several regions be selected based upon their potential to address one or two of the critical questions? All of this implies that the CLEANER program should be driven by where the data will be of most scientific value and where the available resources can be used most cost- effectively, not just for geographical coverage. Covering a range of spatial scales is also important. Smaller systems can serve as more isolated laboratories. For example, the Lake Tahoe basin has a long record of data collection, modeling, and research. This smaller system has the benefit of a very distinct and enclosed watershed and a centralized high alpine lake. For larger systems various "nested" monitoring approaches can be designed to provide both the larger and smaller watershed laboratories. The more detailed systems could be a size (<100 square miles) that is more responsive and reactive. Smaller systems are likely to be more useful when building and testing models and evaluating reactions to changing environmental conditions. By "nesting" within larger systems the analysis of the larger scale and far field impacts can also be evaluated concurrently. Inability to Connect Research with Required Actions All research proposals must be connected to an action plan that produces results and tests hypotheses and eventually makes a positive difference to human conditions. The sequence of converting data to information and then to
Implementing Environmental Observatories 47 knowledge and action must be thoroughly considered and planned before selecting from the many alternatives any particular scientific research program to undertake. Analysis and use of data need to be stressed, rather than simply collecting and archiving data. There is also a critical need to better communicate scientific and engineering findings with resource managers and the public. Public, and thus political, support for continuing these long-term observatories will be essential for their long-term success. Improved linkages with the social sciences will facilitate understanding the need for communication and the communication of findings to important stakeholders. While not the mission of CLEANER, an outcome many stakeholders might wish from a successful and ongoing CLEANER program would be the establishment of environmental monitoring and management centers. These centers would have access to near real-time information and predictive models that together would provide early warnings of potential adverse hydrologic and environmental impacts due to natural or human activities. The establishment of an early warning system based on observation data, especially of resources or species under stress, could help identify emerging problems in time to change land and water management policies to prevent further degradation if not permit some restoration. Such an early warning system must be based not only on what is being observed, i.e., current data, but also on the research stemming from the long-term integrated data sets obtained from the environmental observatories. In 1990, a multi-panel committee of the EPA's Science Advisory Board developed a ranking of risks to natural ecology and human welfare. This ranking turned out to be contrary to the public's perception of risks. From a scientific viewpoint, loss of biological diversity was a higher risk than were ground water pollution and presence of radionuclides (EPA, 1990). Environmental observatories, and indeed these potential environmental monitoring and management centers, must focus on the risks perceived by both the public and the scientific community if emerging problems are to be detected in a timely manner. Inability to Train Enough Engineers and Scientists As discussed above, training future engineers and scientists in the science of remote observing systems and associated cyberinfrastructure will be critical to the long-term success of CLEANER and other observing systems. Training is the job of universities, and graduate programs in environmental subjects cost money. Environmental and ecological engineering departments in universities need to attract bright graduate students to undertake such studies. Training grants and fellowships, perhaps similar to those provided by the public health service that resulted in strong environmental engineering programs of the 1960s, may need to be considered.
48 CLEANER and NSF's Environmental Observatories SUMMARY Planning and development activities are underway by the NSF for several national environmental observatory systems. These systems each have somewhat different goals and missions, but there is also considerable overlap in terms of scope and components of research and outreach activities. In addition, there is a wealth of federal, regional, state, and local environmental monitoring in the United States. These monitoring programs are typically closely linked to the mission of the coordinating agency and are coupled with management of natural resources. It is important for CLEANER to be well integrated into the existing and planned networks of the NSF environmental observatories and various federal, regional, state, and local monitoring activities as well. CLEANER is in a unique position to catalyze linkages between the national science-based observing systems and governmental monitoring efforts. CLEANER is distinct from the other proposed environmental observatories due to its focus on interactions between the natural environment and engineered systems, and particularly human-stressed systems. As a result, CLEANER should be closely aligned with applied issues that are of national significance and associated with government environmental monitoring and assessment programs. If properly coordinated, CLEANER could add considerable value to ongoing government monitoring efforts through data management, and the application of models and other analysis tools. In turn through proper coordination and linkages, CLEANER could add value to the data obtained from other observatories as it focuses on engineering and more applied issues. To facilitate coordination of the national environmental observatories, the NSF should strongly consider establishing a parent organization that could be termed something like the "Environmental Observatory Networks" or EON, which would be responsible for cyberinfrastructure development, educational activities, outreach, and other shared activities across the NSF observatory programs. EON (or some alternative entity, to be determined by the NSF) would facilitate collaboration and coordination among the observatory programs and minimize redundancy. To facilitate and improve linkages between CLEANER and existing governmental assessment programs (federal, regional, state, local), a series of workshops, inter-agency agreements, and other collaborative activities is recommended. There are several areas of common interest across the NSF environmental observatories, including cyberinfrastructure (e.g., data, software, sensors, models), education, and outreach. These activities and initiatives should be coordinated and joint programs should be established among and involving all observatories. A comprehensive approach to building an efficient, secure, scalable, and extensible environmental observing system (or systems) will require careful design across all three areas. A critical activity of national environmental observatories will be outreach to water resource managers and the public, as well as training the next generation of scientist and engineers in the "science of environmental observing systems."
Implementing Environmental Observatories 49 The committee supports the concept of CLEANER and recommends that it proceed with planning and coordination activities. This planning should include an analysis of potential "fatal flaws" that could limit, or possibly negate, the benefits expected from CLEANER. Should the NSF become assured that these potential pitfalls can be avoided or mitigated, CLEANER has the potential of supplementing existing national science-based observing systems and governmental environmental assessment efforts, especially those focused on the interactions between the natural environment and engineered-human systems. A successfully operating environmental observatory network could transform the environmental engineering profession and increase its already considerable contributions to society.
