A great deal of effort has already been expended establishing the scientific and intergovernmental organizational and planning mechanisms for GOOS. During the last five years many recommendations have been passed, both scientific and at the intergovernmental level, in an attempt to see the concept of GOOS made into a real, effective system. Unfortunately, most of these recommendations have borne little or no fruit, principally because there were not sufficient resources to ensure the recommendations were acted upon. There have been other factors, such as the lack of a permanent GOOS Support Office Director, and, it must be admitted, several of the recommendations were unrealistic. This meeting provides yet another opportunity to assess, evaluate and recommend actions for the implementation of GOOS. It would be pointless to revisit issues that are relevant only to GOOS or, as is more often the case, relevant only to the community and agencies concerned with a particular aspect (module) of GOOS.
What this meeting does provide is an opportunity for GOOS to make others in the global observing system community more aware of its objectives and operating principles, and of its products and database, and in turn for GOOS to become acquainted with planning and implementation activities external to its domain. It follows then that the meeting should concentrate on those GOOS issues which demand, or could benefit from, discussion and debate in the broader, multi-disciplinary community concerned with integrated global observing systems.
A detailed background document concerning ocean observations was prepared for the meeting (Smith et al., 1996), and will be published separately.
General Conclusions and Recommendations
First, some general remarks should be made about the relationship between the (integrated) global observing systems and that part which we refer to here as GOOS, the ocean observing system. From the viewpoint of GOOS and its scientific planning, most people involved concede there is a considerable amount of work yet to be done and that an integrated GOOS largely remains in the conceptual phase. Furthermore, at the component (module) level, the various specialist groups are at different stages of development. The Climate and Health Of The Oceans (HOTO) modules are beginning the task of implementation whilst the Living Marine Resources (LMR) and Coastal Zone (CZ) modules are only at a preliminary strategic design stage. Thus there are very few examples of successful integration at the component (or even sub-component level). This is not to be interpreted as implying that the concept of GOOS is flawed, but that GOOS is embarking on a task that is challenging in the extreme, and patience and effort will be required. What experience GOOS does have would suggest that integration and prioritization must be tackled at the sub-component and component levels before attempting it at the GOOS level. What, then, is the prospect of achieving useful outcomes at the next level in this hierarchy, in terms of the overall global observing systems (GCOS, GOOS, GTOS), or "G3OSs" as they are colloquially known? A conclusion we can draw from the experience of GOOS is as follows.
Conclusion 1: GOOS will benefit from an integrated global observing system through the identification of opportunities for constructive cooperation among the three global observing systems.
However the development of overall strategies for the three global observing systems should avoid directly or indirectly duplicating the decision-making and prioritization processes that already exist at the observing system (e.g.,the Joint Scientific and Technical Committee for GOOS (J-GOOS)) or observing system component (e.g., the HOTO module) levels. The interaction between GOOS and the atmospheric global observing system is fundamental to many of the applications presented here and must be strengthened and sustained. There is also an issue of cooperation through common (non-global) regional systems.
From the outset it has been realized that ocean observing systems must be supported by science, both at the planning stage and as part of the evolution of the system. The observing elements also need to be effective and efficient if they are to be maintained long-term and be able to withstand the inevitable cost versus effect pressures. It is for this reason that a definition of "operational" is proposed, that underpins the steps toward prioritization. An "operational" data-gathering system for GOOS should be:
systematic - measurement methods should be supported by a sound scientific rationale;
relevant - measurements should be taken with purpose and due regard of the GOOS objectives;
long term and routine - the methods should not be experimental or ad hoc, but should be based on a history of successful testing and should be capable of implementation over extended periods with regular sampling and minimal reliance on direct scientific oversight; and
cost-effective - observational methods should be efficient and economical.
If a potential measurement system cannot meet these requirements, then it is unlikely to be regarded as a high priority for GOOS, though it may, however, continue to be regarded as important for experimental programmes. The distinction between "operational" and "research" is even more formal within meteorological agencies with regard to numerical weather prediction.
Recommendation O1. A general recommendation is that the global observing systems should seek a common methodology for drawing the distinction between experimental and ad hoc observations on the one hand, and systematic, routine global observing system observations on the other.
Recommendation O2. It follows from conclusion 1 and recommendation O1 that the global observing systems should recognize the fundamental importance of prioritization for the development of operational, long-term observing systems and should encourage the development of common approaches to prioritization.
For the implementation of in situ systems, where limited resources and lack of longevity are often the norm, prioritization is critical. The need for prioritization, and clear demonstration of the importance of observing system elements, which are candidates for implementation, is repeatedly emphasized by those charged with implementation. This does not mean that the recommended prioritizations are fixed. They should be revisited as alternate or improved technologies emerge, or as understanding of the sampling requirements change, though this revisiting should be on a frequency commensurate with the long-term, "stable" principles of the observing system.
Data and information management also tends to have generic issues that cross the boundaries of the different observing systems. GCOS and GOOS have recently agreed to adopt a unified approach to data management through the appointment of a single data and information management panel. While this rationalization is laudable, some very difficult problems are bound to arise as the HOTO, LMR and CZ modules develop. Many of the data collection methods, quality assurance processes, communication techniques and archiving procedures are, according to the "operational" definition above, still in an experimental phase. Progress in these cases may depend on allowing further experimentation and refinement rather than the imposition of strict guidelines. Another thorny issue concerns the "free and open exchange of data and products" principle that has been adopted for GOOS. For the data and information pertaining to the exclusive economic zones of different countries, it may prove very difficult to get agreement on just how this principle should be invoked. Perhaps now, before planning has advanced too far, we should be seeking out the principles and guidelines that are generic and seeking general endorsement of those principles.
Recommendation O3. It is recommended that the global observing systems should seek a set of generic principles and/or guidelines governing the management of data, information and products of the global observing systems. These guidelines should not be restrictive but provide clarification and direction, based on the experiences of existing operational and experimental systems, that will facilitate the development, planning and integrated implementation of global observing system elements.
Remote sensing effectively imposes integrated requirements across the modules of GOOS and, in many cases, across the different global observing systems. For example, surface wind and sea surface temperature information is an important requirement for all modules of GOOS, although each module is likely to specify different sampling characteristics. While the in situ requirements may be met in different ways, the remote sensing requirements will normally depend on a single platform (or, at most two or three). GCOS has put considerable effort into identifying common requirements for GCOS and in many cases for GOOS (GCOS, 1995c). A similar approach should be taken for the global observing systems. However, we must avoid the "wish list" syndrome, where specifications are added without due regard to the actual priority attached to the data by those planning and coordinating the observing system components and we must maintain as far as possible the connections between the specified requirements and the user community and observing system products. Tables of sampling densities and frequencies lose all meaning if they do not have an attached context and clear purpose.
Conclusion 2: GOOS seeks a unified and integrated approach to remote sensing ("All for 1 and 1 for all") whereby its requirements are considered as part of the justification for operational sensors, rather than each observing system seeking to justify sensors in their own right. Furthermore, GOOS strongly commends an approach to remote sensor specifications that takes full account of the basis for prioritization and rationalization between GOOS elements, and which maintains a clear linkage between the specifications and the prime user communities and products that ultimately must provide justification for the sensor.
Specific Recommendations
The discussion of the ocean observing system for climate strongly supports the following:
Recommendation O4. The global observing systems should seek a multi-disciplinary response to the demands for observations and analysis from the IPCC assessment process, and to the state-of-the-environment reporting processes.
GCOS is in part meeting this challenge, but in a more general sense. This recommendation suggests that the IPCC process should be used as a motivation to focus and integrate relevant climate activities. Such a response requires close collaboration, and joint implementation, with research groups. The global observing systems might also assist in coordinating approaches used in assessing the state of the environment.
Recommendation O5. GOOS and the Climate module in particular recommends a cross sectoral approach to calibration and validation of remote sensors with in situ data and notes that the detail of sensor sampling requirements are dependent on the application. For several sensors the applications cross GOOS component and other global observing system's component boundaries. Merged satellite-in situ products are proving more useful than either of the individual (remote or in situ) products.
Conclusion 3: Routine forecasts of the El Niño-Southern Oscillation phenomenon are now feasible. Their existence creates many opportunities for inter-disciplinary collaboration. In particular, products now being developed outside the domain of GOOS, such as agricultural outlooks, depend directly on terrestrial data and indirectly on ocean (GOOS) and atmospheric (WWW) data. This observing system inter-dependency demands cooperation and collaboration in evaluating system design and prioritization.
The discussion from both the climate/coastal physics presentation and that of HOTO identified an opportunity to pursue planning in concert with that being undertaken within GTOS, and to cooperate with and build on the research being undertaken within the IGBP Land Ocean Interactions in the Coastal Zone (LOICZ) programme. This would be facilitated by agreements at the sponsor level that their requirements could be met by a cooperative approach. Therefore it was agreed that:
Recommendation O6. The philosophy, design and implementation of coastal zone measurements within GOOS should be coordinated with similar activities within the other global observing systems, particularly GTOS.
The presentations on health of the oceans living marine resources and biology covered many issues, some of which are common to planning issues faced in GTOS. The following conclusions and recommendations were reached.
Recommendation O7. Sensors and systems (e.g., acoustics and optics) which can provide key LMR data should be included on other platforms (e.g., moorings and VOS) to provide concurrent, complementary and economical information about LMR.
Such an approach not only achieves cost savings but allows immediate access to contextual information such as temperature and circulation fields, which assist interpretation.
Conclusion 4: The development of the GOOS LMR module must involve consultation and cooperation with many groups outside the LMR community, and groups external to GOOS (e.g., Joint Global Ocean Flux Study (JGOFS), Global Ocean Ecoystem Dynamics (GLOBEC), LOICZ).
Recommendation O8. The effective implementation of the LMR module depends on the development of a more timely and effective system for gathering and collating fish stock data. It is recommended that opportunities within the global observing system process be sought in order that such a system can be put into place.
Conclusion 5: The GOOS LMR module, perhaps more so than any of the other modules of GOOS, depends on progress and cooperation with the research community.
