Sustained Time-Series Observations for Ocean Research

Since the earliest expeditions of H.M.S. Challenger in the 19th century, progress in ocean research has been driven strongly by the ability to make new observations - either located in new places (i.e. classical exploration-going to places on Earth that have not been observed before) or new types of measurements that permit natural phenomena or processes to be understood in different ways. The state of knowledge concerning our planet's oceans is built primarily upon this foundation of `spatial' exploration.

As our knowledge of the oceans has improved, observational oceanography has evolved to include `process studies' - where both location and observational strategies are chosen carefully to enhance the likelihood of gaining insight into the physics, chemistry, or biology of one particular active process.

Simultaneously, the realization has grown that few characteristics of the ocean are in steady state _ the ocean and the seafloor beneath are highly dynamic environments. If these processes are to be understood, if new insights are to be gained, if quantitative models are to be validated satisfactorily, then observations are needed over the time scales appropriate to the dynamics of these processes. We know enough today to realize that these time scales span milliseconds to decades, centuries and beyond.

Today's data collection and experimentation in oceanography are carried out most commonly in an expeditionary mode _ a one or two month-long research 'cruise' focused upon a specific problem, or set of problems, led by one or more investigators who claim substantial rights to the data that results from their efforts.

We expect that this expeditionary mode of operation will evolve substantially over the next decade, driven primarily by the growing and well-justified need for sustained time-series observations. This need is clear, not only in our most reliable source of information concerning research trends (the proposals that we receive), but also in essentially all of the community-based planning documents that have been produced in recent years. It is important to the health of the science that the Division of Ocean Sciences (OCE) respond to this evolution by studying whether the facilities and support mechanisms that we provide are optimal to this new way of doing oceanography, and if not, to make the changes that are required.

In doing this we face many challenges. The role of NSF in the support of long-term measurements is often questioned. Is this not just `monitoring'? And is not `monitoring' the purview of the mission agencies that can more appropriately tailor the measurement systems to their defined operational goals?

The most effective response to these valid and important questions is as follows.

The Division of Ocean Sciences should support the data collection that its investigators need to tackle the most exciting and topical problems in the discipline. Today this means that, in many cases, sustained time-series of multiple years in duration are needed. Recent experience shows that new and unexpected discoveries are made when careful measurements are carried out over sustained periods of time. Such discoveries further confirm the complex dynamics of many of the Earth's processes. They challenge, and sometimes require the rejection of, long-standing models for the explanation of what were previously thought to be steady-state phenomena. Examples of this are too many to mention, but span the full range of OCE endeavors - from the unexplained changes in flow rate of water through deep sea boreholes over several years to the determination that pelagic sediment flux onto the seafloor is dominated by discrete `bloom' events.

When investigators work to understand the ocean by making sustained time-series observations they are, in effect, 'exploring-in-time'. The earliest oceanographers made the first great discoveries by conventional spatial exploration - they traveled to new places in the oceans and discovered unexpected phenomena that catapulted their understanding of a particular process to a new level. Today, innumerable examples exist in the published literature of important and sometimes unexpected discoveries resulting from the collection of long time-series data sets. But this is not 'monitoring' - it is the classical combination of hypothesis testing and exploration, but in the time domain, not the space domain. Researchers do not continuously measure some parameter for no good reason! They are continuously developing, changing, and improving measurement strategies and techniques to maximize understanding and insight.

In parallel with this intellectual evolution that is yielding new insights into the dynamics of Earth and ocean systems, the technology that is needed to make the required measurements is evolving at a comparably rapid rate. A number of highly successful pilot experiments using seafloor cables have revolutionized measurement strategies by bringing substantial power and data bandwidth to the seafloor. Satellite-telemetering profiling floats have been used to provide a synoptic view in real time of basin-scale circulation of the upper ocean. Autonomous underwater vehicles with decision-making capability are making measurements on temporal and spatial scales never before possible. And one of the jewels in the crown of the Joint Global Ocean Flux Study (JGOFS) Program is the breadth and significance of the discoveries yielded by the Hawaii Ocean Time-series (HOT) and Bermuda Atlantic Time-series Study (BATS) sites.

And so, if OCE is going to fulfill properly its responsibility to the community to facilitate the highest quality research, we should begin to investigate ways to better enable time-series based research activities. We provide research vessels to investigators at no cost to their research grants, to enable research to be carried out anywhere in the world's oceans. But if an investigator wants to go to one place (or several places) and make measurements for several years, then no standard community-wide infrastructure is available. The concept of the UNOLS fleet some thirty years ago revolutionized ocean sciences in the U.S. by providing open access to the oceans to NSF-supported researchers. Perhaps the time has come to enable a similar revolution today, but in the acquisition of time-series data sets.

With these thoughts in mind, we requested the Ocean Studies Board of the National Research Council to convene a workshop, think about all the issues I have described above, and evaluate their importance. This workshop was held in Key Largo, Florida in January 2000. It stimulated substantial high quality discussion, and we look forward to receiving the report in the near future.

It would not be wise for NSF to develop activities in ocean observations in isolation from the interests and requirements of our sister ocean agencies in the Federal government. Although the exclusive focus of this article has been fundamental research, there are rich and diverse opportunities for major progress towards the attainment of the mission agencies' operational goals through ocean observations. As NSF moves forward to support new sustained time-series research activities, we will do so in close coordination with the other agencies of the Federal government. NSF is an active player within the National Ocean Partnership Program's planning activities for the development of a U.S. Integrated Ocean Observing System (IOOS). It is clear that the operational goals of IOOS and the research goals of NSF-supported investigators are highly synergistic and there is substantial benefit to considering both research and operational objectives when designing components of the observing system.

This is an exciting time in the ocean sciences - the discipline is rich in new ideas, and the development of new technologies is changing the way we observe the ocean. OCE is responsive to these changes and is willing to implement new ways to enable the community's research.