4.1 Review of Status of Science Plan

The U.S. JGOFS Science Plan (U.S. JGOFS Steering Committee, 1990) provides a detailed rationale for the time-series component of the program. The plan suggests that time-series observations extending over many years are the only strategy by which the longer term cycles and trends of the ocean can be observed and understood. Logistical requirements of time-series measurements are a dominant concern and therefore island stations assume a particular importance.

The objectives of the time-series studies within JGOFS are: 1) to observe and interpret the annual and interannual variability in the biology and chemistry of the mixed layer and euphotic zone as forced by physical processes, 2) to observe and interpret the annual and interannual variability in the rates of particle flux and the apparent rates of particle remineralization over the entire water column, 3) to understand the interrelationships between the biological, chemical and physical processes involved in (1) and (2) above, and 4) to provide data on global trends of selected oceanic properties over seasonal and interannual time scales.

In addition, time-series stations are optimal sites for:

1. Detailed process study research and testing of ecological hypotheses and model predictions.
2. Intercomparison of methods.
3. Method and instrument development and testing.
4. Test beds for moored arrays of bioptical, acoustical, geochemical and physical sensors.
5. Surface truthing of satellite data.
6. Data supply for JGOFS modeling activities.
7. Ocean-atmosphere interaction experiments.

4.2 Components of the Time-series Program

The implementation of the time-series part of the U.S. JGOFS Science Plan will be coordinated with other international programs within JGOFS. The main components of a time-series program include:

1. An agreed upon suite of well-defined intercalibrated, core measurements will be made through the water column at regular spacing at approximately monthly to bi-weekly time periods. Particle flux measurements will be made using moored and floating arrays. Other sites at which bi-weekly or monthly sampling is not feasible, e.g. high latitude sites, may also be considered as time-series sites.

2. Shipboard measurements will be supplemented by remote measurement technologies such as surface moorings, tethered vehicles, and autonomous underwater vehicles (AUVs). Sensor development will allow for an expanded parameter set to be included in remote instrumentation.

3. As various process cruises within the U.S. JGOFS context may require an elucidation of the temporal framework for the study, time-series measurements will be expected to be incorporated in the implementation plans of the process cruises in different basins.

4. The U.S. JGOFS time-series program will therefore provide expertise to the international JGOFS program for establishment of new time-series stations as well as providing input for time-series programs implemented to resolve temporal questions within the U.S. JGOFS process studies.

4.3 Review of Current Activities

At present there are two manned time-series stations as part of the U.S. JGOFS program. These are located near the islands of Bermuda and Hawaii. Both have conducted monthly cruises since October, 1988. The island locations and histories of the institutions responsible for the measurements make logistics easier than in other areas. Both of the existing time-series stations are located in oligotrophic gyres of the Atlantic and Pacific oceans, with the Bermuda station showing strong seasonal patterns of mixing and production.

4.3.1 The Bermuda Atlantic Time-series Station (BATS)

BATS was established in early 1988. The station is located at the Ocean Flux Program (OFP) site (3150N, 6410W) 80 km southeast of Bermuda with a water depth of 4,800m. Monthly hydrographic cruises commenced in October, 1988. The site is near the location of Werner Deuser's (WHOI) mooring site for deep sediment traps, a continuing record of deep sinking fluxes that began in 1978. Hydrostation S lies 22 km southeast of Bermuda and the 40 year time-series record of biweekly hydrographic sampling is continuing at this station. Thus, BATS continues and builds upon a long tradition of time-series research in this area. To date over 75 cruises have been carried out. In addition to the major monthly cruises, sampling frequency has increased during spring with a set of short cruises to document the yearly over-turn and bloom period.

4.3.2 The Hawaii Ocean Time-series (HOT)

The HOT program is a multi-investigator, interdisciplinary research program that has successfully established and maintained a deep, open ocean hydrostation for the purpose of observing and interpreting oceanic variability. The HOT program has two major components: JGOFS and WOCE. This cooperative arrangement has resulted in a close collaboration between the JGOFS and WOCE program scientists at the University of Hawaii and will undoubtedly yield a higher quality data base than either of the individual groups could have assembled alone.

In October 1988, a permanent site was established at a location referred to as Station ALOHA (A Long-term Oligotrophic Habitat Assessment). The station is approximately 100 km north of Oahu at 2245N, 15800W, and is near a TOPEX/Poseidon satellite crossover point. It is 50 kilometers away from the steep topography associated with the Hawaiian Ridge. The water depth is 4,750 m, the seafloor is without relief, and the sediments are predominantly composed of pelagic red clay. Station KAHE, located off Kahe Point at the 1,000 m isobath is also routinely occupied during monthly cruise operations. Between October 1988 and the end of 1994, fifty research cruises were conducted to Station ALOHA.

4.3.3 Measurement Parameters

The following parameters are routinely measured at the BATS and HOT sites:

1. Temperature and salinity versus depth with CTD

2. Dissolved oxygen by Winkler titration (used to calibrate CTD sensor

3. Salinity by conductivity using an Autosal 8400A (used to calibrate CTD sensor)

4. CO2 by coulometry, alkalinity by Gran titration

5. Dissolved nutrients (nitrate+nitrite, phosphate, silicate) by standard colorometric techniques or a Technicon Autoanalyzer II

6. Particulate organic carbon and nitrogen by high temperature combustion CHN analyzer

7. Phytoplankton pigments; chlorophyll a and phaeopigments by fluorometry; 19 photosynthetic pigments by HPLC

8. Bacterial abundance by the acridine orange or DAPI direct count method (BATS) or flow cytometry (HOT)

9. Primary production at 8 depths using trace metal clean sampling and incubation methods and 12 hour in situ incubations

10. Sinking particle fluxes by 72 hour sediment trap deployment, traps located at 3 to 4 depths between 150 and 500 m depths

The results from both stations are sent to the U.S. JGOFS Planning Office at the Woods Hole Oceanographic Institution and then forwarded to the U.S. NODC in Washington DC., where they are available to the community as a whole.

4.3.4 Other Activities

Other activities within the U.S. JGOFS effort include method development for time-series measurements, interaction with international JGOFS programs involving standards, reference materials and development of measurement protocols. Part of this component could include training and advice for other international time-series programs.

4.4 Detailed Implementation Strategy

4.4.1 Near Term

The location of future time-series stations as part of the U.S. JGOFS effort must be determined by the overall scientific goals of the U.S. JGOFS program. The design of each time-series sampling program must be determined by the nature of the environment, the scientific questions involved and the major logistical considerations. Therefore the manned-ship-cruise effort presently at Bermuda and Hawaii needs to be continued and needs to be supplemented by new automated time-series technologies (e.g., moorings, automated underwater vehicles) as they emerge to provide a linkage between the different approaches. Future time-series stations would use the appropriate combination of ships, moorings or AUV's determined by the scientific goals. These two existing sites are critical as test beds for existing and emerging technologies for time-series measurements world-wide. They are also essential for training and other educational activities within the JGOFS program as a whole. Many of the areas of interest to JGOFS are in logistically difficult areas, therefore a commitment to technological development is essential.

4.4.1.1 Future Time-series Sites

Selection of future time-series sites will be based upon the following criteria:

• Will it provide dynamic interpolation and temporal history at a JGOFS process study site? Time-series measurements will play integral roles in process study cruises. For many areas it is envisaged that these would be automated stations (moorings, AUV's; see Dickey et al., 1993). For future process studies, the times-series efforts should begin prior to the main process study.

• Will it provide critical information in areas where interannual and seasonal variability is likely to be important for the parameters of interest to JGOFS? These sites are likely to be some combination of automated and ship-board operations. Stations that have very demanding logistics requirements (e.g., Southern Ocean) should be entirely automated.

• Will it provide critical temporal resolution of seasonal atmospheric and oceanic interactions for specific experiments such as the iron fertilization hypothesis?

Most resources are presently available for the two existing time-series sites to carry out the work detailed above. Additional resources are required to fund the core measurement work above until 1998, with the addition of bio-optical, acoustical, geochemical and physical mooring-remote sensor development at both sites. These research activities require approximately $1M per annum per site, excluding 60-75 days of shiptime. Additional resources would be required to expand the scope of the research or to automate the existing measurement programs. For example, the addition of a physical/bio-optical mooring would require a supplement of at least $400K per site with continuing costs of approximately $150K per year.

4.4.2 Medium Term (0--2 years)

Resources will be required for various time-series aspects of the process studies planned or in the planning stages. These resources will be required within the design of each process study and will be listed as part of the individual process study. As the long term plan for reoccupation of process study sites becomes more firm, it will be imperative that the time-series stations be established to provide a link between these periods of major occupations.

4.4.3 Long Term (0--5 years) Automated In Situ Instrumentation for JGOFS Time-series Observations

Automated in situ instrumentation is needed for time-series observations of bio- geochemical, bio-optical, acoustical and physical variables at present and future JGOFS time-series sites for several important reasons (see review by Dickey, 1991; Dickey et al., 1993).

1. The processes controlling the flux of carbon in the upper ocean have dynamic ranges in space and time of at least nine orders of magnitude (Figure 1). These processes depend on biogeochemical, bio-optical, acoustical, physical oceanographic, and meteorological variables.

2. Ship-based sampling, while critical for detailed and more comprehensive observations, can span only a limited portion of these ranges because of logistical and financial constraints. Thus, it is important that high-resolution time-series data (including currents) be collected at the present time-series sites as well as future sites and that horizontal/vertical spatial data be collected in the vicinity of these sites. New sensor designs enable collection of many of the principle ocean variables with time scales as short as 1 minute.

3. Although remote sensing of the very near surface can provide important sea surface temperature, ocean color, and sea surface elevation data on scales as small as a few kilometers to 10's of kilometers with repeat orbits of order of a day. It is important to note that a) clouds frequently obscure SST and color images, b) only the upper centimeters are sensed for temperature and optical depths of m's to 10's of meters are sensed by color images, and c) some problems with SeaWiFS color images collected near any land mass are anticipated. The sampling of the subsurface ocean on horizontal scales of 100's of meters is presently limited to ship-based operations (e.g., tow-yo's). This is generally prohibitive because of shiptime costs and the limited sampling (bi-weekly to monthly at best). For this reason, development and utilization of autonomous underwater vehicles (AUV's) is highly desirable.

4.4.3.1 Sensor Development

In order to detect physical climate change, relatively simple measurements of temperature are crucial. In terms of biogeochemical changes, variables such as pCO2, total CO2, alkalinity, dissolved oxygen, water clarity, pigment biomass, and phytoplankton productivity are the master variables. The unambiguous interpretation of measurements used to determine these quantities remains a challenge, however, considerable progress in making high-resolution, long-term bio-optical, biogeochemical and acoustical measurements has been made and rapid advancements in sensor and system development seem likely. Therefore, automated sampling at time-series sites is not only important, but technologically feasible. This will require a commitment to testing and use of established sensors as well as improvement and development of new sensors.

4.4.3.2 Site Selection

Global scale changes cannot be determined from a limited number of time-series sites. However, changes at these sites can be used to monitor long-term trends and indicators of significant changes in oceanic biogeochemical and bio-optical variables, just as the measurements at Mauna Loa have provided compelling evidence of atmospheric CO2 concentration increases. Therefore it is intended to take advantage of the core data measurements at the Bermuda and Hawaii sites to test some of the present and future technologies and add to the temporal data base.

In addition, remote sensing of ocean color and the derivation of upper ocean pigment biomass and primary productivity on regional and global scales will be done with SeaWiFS. The requisite algorithms rely on in situ observations of bio-optical variables. Therefore, inconsistencies between satellite and in situ sensors in temporal and spatial sampling scales must be interpreted and corrections derived at these stations. Additionally, satellite-derived ocean color and temperature measurements are often obviated by cloud or water vapor conditions. Further, the understanding and interpretation of fundamental measures such as fluorescence remain problematic. Therefore, intensive and extensive shipboard sampling at mooring sites will be important.

4.4.3.3 Data Acquisition

One of the advantages of automated sampling is that data collected from such systems, in principle, can be transmitted in near real-time to shore-based laboratories around the world. The ARGOS satellite communication system is presently used by programs such as the Tropical Ocean Global Atmosphere (TOGA) program to distribute meteorological and physical data in near real-time. The near real-time capability is useful for continuous monitoring, planning sampling strategy, and for insuring data retrieval in the event of instrumentation loss or major malfunction in data recording. An advanced version (greater data transmission capabilities, especially bandwidth and volume) of ARGOS will be needed for the large number of variables collected at high data rates. Presently, near surface instruments are easily linked to such communication systems and technological advances are being made in transmitting data from depth acoustically (e.g. Dickey et al., 1993).

4.4.3.4 Description of Platforms

The three primary modes of collecting in situ data from automated systems will be from moorings, tethered vehicles and AUV's. Several successful experiments have now been done using moored bio-optical, physical, and geochemical sensors in the open and coastal ocean (e.g., review by Dickey, 1991). A detailed discussion is presented below as these platforms will be essential in the long term for JGOFS as well as for systems such as the Global Ocean Observing System (GOOS).

These devices are important for the JGOFS time-series program as they may be used:

• to establish relationships between biogeochemical processes (particularly carbon transport from the euphotic layer) and physical forcing;

• to intercompare several methods for the determination of primary production from moored instrumentation and sediment traps

• to make accurate vertical, horizontal, and temporal estimates of pigment biomass and phytoplankton productivity leading to fluxes of carbon

• to enhance understanding of relationships between primary production and export of carbon from the upper ocean and to model the flux of carbon through the upper ocean

• to maximize the accuracy of future regional and global satellite pigment biomass and phytoplankton productivity estimates.

4.4.3.5 Moorings

Development of interdisciplinary moored instrumentation systems has been driven in part by the need to increase the time domain of bio-optical, biogeochemical, and acoustical sampling. The selection of appropriate sensors and inherent constraints such as power consumption, data storage, and biofouling are of special concern for this method of deployment. The potential of obtaining moored bio-optical data has only begun to be realized. Yet data have been collected in several regions of the world ocean (e.g., Sargasso Sea, the coasts of California and New York, Equatorial Pacific, and Iceland) from moorings (see review by Dickey, 1991; Dickey et al. 1993).

Some of the moored sensors include: thermisters, conductivity sensors, vector measuring current meters, strobe fluorometers, natural fluorometers (683 nm upwelling radiance), photosynthetic available radiation (PAR) sensors, beam transmissometer (660 nm), dissolved oxygen sensors, spectral downwelling irradiance sensors (410, 441, 488, 520 and 560 nm), and spectral upwelling radiance sensors for the same wavelengths. Data are typically recorded every 1-4 min. Because of the high temporal resolution, it is possible to do spectral analysis of bio-optical as well as physical data in the open ocean. The processes contributing to variability shown in resulting spectra include diel cycles of phytoplankton, tides, inertial currents generated by passing weather systems and wind events, and internal gravity waves. Variability in stimulated fluorescence and beam attenuation coefficient are related to phytoplankton biomass. Recently, estimation of primary production using models and data obtained from time-series bio-optical measurements has become possible. Some of the methods which can be used to estimate primary production using mooring data have been presented by Kiefer and Mitchell, (1983), Emerson, (1987), Siegel et al. (1989), Kiefer et al. (1989), Dickey (1991), Marra et al. (1991), and Stramska and Dickey (1993).

4.4.3.6 Tethered Vehicles

Remotely Operated Vehicles (ROVs) are an emerging technology that show great promise for addressing specific research questions related to JGOFS goals. These vehicles are tethered to a surface platform, usually a research vessel and are controlled from the surface through the tether. The surface connection frees these vehicles from many power and data transmission constraints. They can transmit live video images and other wide-bandwidth data to the surface in real time. They can have sophisticated sampling and manipulation equipment onboard and controlled from the surface. Conversely, they are constrained by the length and hydrodynamics of the tether system to measurements near the surface platform, although some of these constraints will lessen as new tether management systems become available. The most apparent uses for these vehicles in the JGOFS context is in conducting specific, process-oriented experiments on ocean biogeochemistry. They allow direct observations of the many organisms that mediate midwater and benthic organic transformations. They allow the quantitative enumeration, collection and incubation of marine snow particles, the dominant type of sinking particles in many parts of the ocean. They are a direct extension of the human eye and laboratory in to the deep sea and are more cost-effective than submersibles that actually carry humans to depth.

4.4.3.7 Autonomous Underwater Vehicles (AUV's)

Exploration of the subsurface ocean has for the most part been done using ships, submarines, submersibles, or remotely operated vehicles and thus required human resources. These methods are costly and require considerable manpower. Thus, there is renewed interest in the development of autonomous underwater vehicles (AUV's) which could be used for regional and global sampling. AUV's can be thought of as ``robotic submarines. A generic AUV may be defined to be a ``free-swimming,'' untethered vehicle with its own power supply, propulsion unit, computer intelligence systems for decision making, navigation, etc., communication links and telemetry, and system and scientific sensors and discrete water samplers. Various institutions have been involved in AUV development over the course of the past 30 years.

The current enthusiasm for AUV's is spurred in part by technological advances in miniaturized computers and artificial intelligence. In addition, a broader suite of oceanographic sensors and water samplers is now available. Some of the necessary electronics, computer control, and scientific sensor capabilities are being utilized for moored instrumentation used for bio-optical and geochemical as well as physical observations (sampling domain of minutes to months).

4.4.3.8 Summary of Automated Instrumentation Needs for JGOFS

Many of the present uncertainties in the estimation of carbon budgets are caused by 1) undersampling (aliasing), particularly in time and space (Dickey, 1992; Wiggert et al. 1993), 2) uncertainties and limitations in existing measurement techniques involving primary production and sediment trap methodologies (Jahnke, 1990), and 3) the lack of concurrent physical, bio-optical, and biogeochemical data (Dickey, 1991; Dickey et al. 1993).

Observations from moorings, ships, and satellites all have sampling advantages and disadvantages. It is anticipated that the moored, tethered and AUV observations at time-series sites will be most useful in reducing many of the present ambiguities and provide new insights into the complicated carbon cycle within the framework of JGOFS. Therefore resources for such technology will be part of the overall JGOFS budget.

All of these new technologies will feed into a new integrated data assimilation system Global Ocean Observing System (GOOS) modeled after the global meteorological stations.

4.4.3.9 Resources Required for Automated Instrumentation Needs

It is unreasonable to assume that JGOFS will bear the burden of full funding for moorings, tethered submersibles or AUV prototype design and field testing. Other federally sponsored research programs, for example the ongoing $10M Office of Naval Research AUV project, will provide invaluable engineering and feasibility data. A fully instrumented JGOFS-AUV (e.g., Ocean Voyager II) is expected to cost $250K and have annual operating costs of approximately $150-200K.

Deep-sea physical/bio-optical moorings have been successfully deployed in numerous locations worldwide. Construction costs vary considerably depending upon configuration and, especially, instrumentation, ranging from $200K to $2M.

Because of their inherently high rates of data acquisition, continuously recording instruments require an additional investment in Data Management at a level of approximately $100K per site.