Outline

I. INTRODUCTION
II. 1996 OVERSIGHT COMMITTEE & ITS CHARGE
III. THE REVIEW PROCESS
IV. THE REPORT
A. Core Measurements
1. General Protocols
2. Quality Assurance and Measurement Compatibility
a. Parameters For Which Standards or Certified Reference Materials are Available.
b. Parameters For Which There are Community Efforts For Methods Comparison
c. Parameters For Which There Should Be Exchange of Personnel and Direct Comparison
d. Parameters For Which Direct Intercomparison is Warranted
e. Non Core Measurements
B. Data Products and Their Distribution
C. Resolving Spatial Variability
D. Resolving Temporal Variability with Moored Arrays
V. RESOURCES
A. Laboratories and Ships
B. Schedule for Core Funding
VI REFERENCES
VI. APPENDICES
Appendix I Bermuda Agenda
Appendix II Hawaii Agenda
Appendix III Pigment Intercalibrations

U.S. JGOFS Time-Series Oversight Committee

1996

I. INTRODUCTION

In the earliest stages of JGOFS, plans for the program's intensive process studies emphasized regions with strong physical forcing. It was, however, realized that the rates and fates of primary production in oligotrophic gyres of the major ocean basins might not be as spatially and temporally invariant as had often been assumed.

One of the few oceanic time-series to document temporal variability in physical, biological, and biogeochemical processes, which could influence seasonal and annual rates of primary production, was the three-year study of Menzel, Ryther and co-workers (cf. Menzel and Ryther, 1961) near Bermuda (Station S). In the late 1950's and early 1960's they observed strong interannual variability in the magnitude, timing and persistence of the annual cycle of primary production, and subsequent analyses of upper ocean density structure and oxygen content at Station S by Jenkins and Goldman (1985) revealed multi-year trends and reversals in these features.

Given the primary productivity data from Station S thirty-five years ago, and the fact that routine hydrographic data had been collected in the intervening period, a North Atlantic oceanic JGOFS time-series site near Bermuda was a logical choice. While a comparable historical base of observations was lacking for an oligotrophic region of the North Pacific, data from several years of cruises conducted by McGowan and co-workers (cf. Hayward, 1987) had revealed interannual variability in this oligotrophic region as well.

When JGOFS began to plan its time-series stations at oceanic sites, it was anticipated that these investigations would require at least a decade of continuity to document and understand the regulation of processes that influence the atmosphere-ocean exchange of carbon in the central ocean gyres.

Accordingly, the primary objectives of the oceanic time-series stations, established by U.S. JGOFS near Bermuda (Bermuda Atlantic Time-Series, AKA BATS) and near Hawaii (Hawaii Ocean Time-Series, AKA HOT) in 1988, were to

1) study the annual and interannual variations in the biogeochemistry of the upper ocean,

2) observe the annual and interannual variations in the rates of particle flux and particle remineralization in the water column,

3) establish relationships between the biological and chemical processes as well as biogeochemical responses to physical forcing, and

4) provide data relevant to global trends of selected properties over decadal time scales.

Successes of the U.S. JGOFS time-series stations in addressing these objectives are now widely evident in journal publications and particularly noteworthy is the recent collection of papers in Deep-Sea Research II special volume (Karl and Michaels, 1996). It is now clear that previously unrecognized similarities and differences exist in the organization of upper ocean ecosystems in these two regions and in the manner in which they process carbon. The extent of interannual variability and the nature of longer term trends are increasingly well established for these two oceanic sites, and with access to this information becoming increasingly widespread it is being used by many research groups in synthesis and modeling efforts to improve precision on estimates of oceanic components of the carbon cycle.

II. 1996 OVERSIGHT COMMITTEE & ITS CHARGE

The U.S. JGOFS Steering Committee has at intervals of a few years charged an ad hoc oversight committee to visit and review activities at the Bermuda and Hawaii time-series stations. Membership of the 1996 team included: David Archer, Jim Bishop, Francisco Chavez, Scott Doney, Jon Sharp, and Jim McCarthy (chair).

The charge to the 1996 review team included assessing whether these two stations are

1) meeting the basic core measurement requirements of JGOFS,

2) maintaining consistent analytical protocols and performance between the two stations, and

3) sustaining timely generation of data and its dissemination to the community of users.

At the onset of U.S. JGOFS, a set of properties and processes were identified as sufficiently central to the mission of JGOFS as to be given 'core' measurement distinction. It was envisioned that these measurements would constitute the basic underpinning of the time-series stations and process study campaigns. Methodologies would be standardized and intercompared between various stations and studies to the extent possible. The pursuit of specific research hypotheses relevant to the location or time of the study were to build upon this common measurement capability. Moreover, the time-series stations were expected to play a special role in these core measurement endeavors due to the continuity of their support and opportunity they provide as a test bed for new and improved methods. It was also envisioned that their on-line analytical capabilities might from time to time be engaged to serve the needs of the process studies. The core measurement data were also expected to be a readily and widely distributed product of U.S. JGOFS science.

Given that the Bermuda and Hawaii time-series stations are in their eighth year of U.S. JGOFS funding, the review team was also charged to consider

1) the current suite of observations and rate measurements and determine if they need to be enhanced or de-emphasized, and

2) whether the new understanding that has emerged from results to date calls for new measurements.

In the course of their review the oversight team was also mindful of the need to

1) anticipate the likely state of the time-series stations at the end of their first decade of results,

2) assess the case for continuing some of the current observations beyond the projected lifetime of JGOFS,

3) provide guidance on the implementation of new observing technology at the time-series stations, and

4) consider interfaces with other national and international ecological and marine observing systems.

III. THE REVIEW PROCESS

Prior U.S. JGOFS Time-Series Oversight Committees submitted reports in 1990 and again in 1993. In the course of the first review a number of issues were raised regarding measurement methods. The second focused more on infrastructure and research products, and provided forward looking recommendations regarding method development.

The 1996 oversight team reviewed the recommendations of previous teams, and gave special attention to

1) lingering (or new) problems relating to intercomparabilty of data from these two stations,

2) community awareness of, access to, and interest in these data,

3) limitations to synthesis activities that rely upon these data,

4) timely modernization of these stations with implementation of new moored and automated instrumentation, and

5) plans for continuing these time-series observations after the period of JGOFS (and WOCE) support.

Meetings of the Committee were convened at Bermuda Biological Station for Research (23-25 March), and the University of Hawaii School of Ocean and Earth Sciences and Technology (29-31 May). Agenda for these meeting are appended.

IV. THE REPORT

The following report clusters around topics arising from the list above. Each has its own set of findings and recommendations.

A. Core Measurements

1. General Protocols

Overview:

Over the history of the JGOFS time-series stations there has been convergence in the measurement procedures for core parameters. Initially scientists at BBSR and SOEST largely used methods that had worked well for them in the past, and this has been true for JGOFS process studies as well. With time consensus has been attained among groups of JGOFS analysts for many of these methods, and this is essential in order to merit community confidence in the comparability of data among different activities of JGOFS. At the onset of the time-series stations it was envisioned by many in JGOFS that BATS and HOT would set standards for the rest of the community with consistent high quality analytical capabilities. Moreover, it was expected that these programs would vet new methods and on occasion serve as analytical centers for other JGOFS activities. BATS has been highly visible in promoting consistent protocols, with Tony Knap assuming responsibility for orchestrating the international JGOFS IOC/SCOR protocol manual (UNESCO 1994). In addition, BATS and HOT have generated independent series of protocol manuals.

Findings:

There have been convergences in protocols relating to several core measurements, and protocols for both programs are increasingly well documented. However, discrepancies continue to appear in comparisons between BATS and HOT data (see below), and in some instances the cause is difficult to determine because the two programs use different methods of sample preparation and/or analysis. The Committee was given drafts of BATS Method Manual, Version 4 and HOT Field and Laboratory Protocols manual 2nd edition

Recommendations:

Efforts should be redoubled to resolve lingering differences in program protocols for core measurements. We applaud the updating of these two programs' respective protocol manuals. However, we believe that the JGOFS community would be best served by now generating a common U.S. JGOFS Time-Series Station Manual that clearly specifies, and justifies any differences between site specific program protocols.

2. Quality Assurance and Measurement Compatibility

Overview:

At both time-series stations it is essential that samples are collected and core parameters analyzed at the highest level of accuracy and precision. This is required to produce data that can address temporal changes at each station and differences between the two stations (presumably differences between the Atlantic and Pacific oligotrophic northern gyres). To assure analytical quality and compatibility of BATS and HOT efforts, a combination of approaches is needed that includes use of standard or certified reference materials, participation in community analytical comparisons, and direct two-station intercomparisons. Below, findings and recommendations are listed in groups according to approaches that have been used or that should be used for this quality assurance and program compatibility. At the end of this section is a discussion of parameters that are not currently considered core at both stations.

a. Parameters for which standards or certified reference materials are available.

Findings:

Both stations are using the same CTD system (Seabird SBE-911), and 12 L Niskin bottles, with adequate calibrations of sensors for reliable depth and temperature measurements. Discrete bottle samples are taken for measurement of salinity at both stations and conductimetric analysis at both is calibrated using IAPSO standard seawater. Although different salinometers are used, direct comparisons of salinity performed in 1992 and 1995 on paired samples shows differences less than 2 milliunits on the practical salinity scale. Dissolved oxygen is measured at both stations using the modified micro-Winkler method with Metrohm 556 Dosimat auto-titrators. Titrators of different vintages are used and hence automation of the titrations is different. The BATS station uses Baker Dilut-It iodate standard while the HOT station uses a CSK certified iodate standard. Both stations measure total dissolved inorganic carbon (DIC) with the SOMMA automatic gas extraction coulometric method. Both use Dickson's Certified Reference Materials (CRM) for calibration checks and both demonstrate excellent performance on these CRM checks. Both stations are measuring alkalinity using modified Gran plot titrations. The BATS station uses a "closed-cell" titration while HOT uses an "open-cell" titration. Experimentation is underway at the HOT station using Dickson's DIC CRM as a calibration check for alkalinity. The HOT program is measuring pH with great precision using the Byrne colorimetric method.

Recommendations:

Little is needed to further verify compatible analyses for salinity because the measurement is absolute in that conductivities of samples are ratioed to the same calibration standard. Similarly, the use of the CRM for the DIC analysis gives good indirect evidence of excellent compatibility and accuracy. However, agreement should be reached regarding alkalinity titration and both stations should use the same procedure. It should be possible soon, with routine alkalinity analyses on the ampoules, to use Dickson's CRM for alkalinity. Periodic exchange of paired samples between the two programs for salinity, DIC, and alkalinity will assure continued compatibility.

The dissolved oxygen analysis could be improved by both programs using the CSK iodate standards. Also, the newest model of the Dosimat is highly automated and gives excellent performance and this could eliminate any possible individual quirks of the computer systems used for the current oxygen analyses. This would also greatly reduce the time for this analysis.

Consideration should be made to add colorimetric pH measurements as a core parameter and to use the Dickson CRM as a preliminary cross-calibration material.

b. Parameters for which there are community efforts for methods comparison

Findings:

Both programs are using high temperature combustion infrared analysis for measurement of dissolved organic carbon (DOC). Since samples are not filtered both stations are actually measuring total organic carbon. The BATS program is using a homemade instrument while the HOT program uses a modified Ionics instrument. Both programs have participated in the Sharp community DOC methods comparison with good preliminary results: however, a direct comparison between the two programs showed slightly higher values by the HOT method. Both programs are measuring total dissolved nitrogen (hence deriving dissolved organic nitrogen, DON) by UV oxidation followed by colorimetric analysis. A direct comparison between the two stations on paired samples did not show as good of a correlation as is desired.

Both programs are measuring pigments with HPLC separations and fluorometry using protocols set up by Bidigare. Both have participated in a preliminary methods comparison set up by Bidigare with differences up to two-fold for all pigments.

Recommendations:

Reference materials will be made available in the next year for continued calibration of DOC analyses and both programs should use these reference materials. While current thinking does not invalidate the older methods for DON, there is serious question about the accuracy of DON as is usually measured. A broad community methods comparison for DON will be started by Sharp in the next year. Both programs should participate in this comparison which will probably work out of either Bermuda or Hawaii for some of the activity.

It is very difficult to do a methods comparison for pigment analyses; a more systematic and carefully controlled comparison is needed to give better confidence that the two programs are obtaining compatible data. Further evaluation should be made of slight differences in the protocols used for HPLC pigment analyses and both programs should continue to work with Bidigare to resolve apparent differences, see Appendix 3. Ideally this comparison can be made as part of the SeaWifs calibration. With all three of these parameters, additional direct comparisons should be made with paired samples from both stations.

c. Parameters for which there should be exchange of personnel and direct comparison

Findings:

In general, measurements of microbial populations and activity are not well standardized in the marine research community. The two time-series stations are ideal for better establishment of routine methods and protocols for these measurements. Enumeration of primary producers is covered under the HPLC pigment analysis and was already discussed. At present, the BATS program does bacterial enumeration with DAPI stain and epifluorescent microscopy while the HOT program uses flow cytometry. At the HOT station it is clear that the more detailed information from the flow cytometry is necessary because of the predominance of Prochlorococcus and Synechococcus, bacteria that function as photoautotrophs not as heterotrophs. The epifluorescence method enumerates all procaryotes without regard to trophic status.

Both programs estimate primary productivity with similar approaches, using clean sampling techniques and dawn to dusk in situ incubations. Comparison of annual production composites suggest different levels of production for the two stations, higher at Bermuda. Microheterotrophic production is estimated regularly at BATS using tritiated thymidine uptake; less frequent estimates of microheterotrophic production is done at HOT with tritiated leucine uptake. Both programs are estimating fluxes from the photic zone using Vertex style floating sediment traps placed near the bottom of the photic zone. While the procedures for trap use is similar at the two programs, depth of deployment and calculation of the flux are slightly different at the two stations. In addition, processing of the trap material is significantly different, especially the method for removal of "swimmers". While there has been some limited opportunity to calibrate floating sediment trap fluxes with 234Th at BATS, this has not been possible at HOT.

Recommendations:

Efforts should be made to make the microbial enumeration and production measurements more compatible between the two programs. There is no easy way to exchange samples for these comparisons. Therefore personnel should be exchanged on cruises and both methods run side-by-side at both stations.

Both prior oversight committees have expressed concern regarding the different methods used at BATS and HOT for removing sediment trap "swimmers". The intercomparison attempted recently by deploying HOT protocols at BATS was inadequate. It is essential that personnel from BATS join a HOT cruise and compare directly the methods used by the two programs for "swimmer" removal. While this work will be time-consuming, incur travel costs, and may necessarily displace lower priority (non-core) tasks from the cue on this cruise, there can be no substitution for this comparison.

Serious consideration should be given to augmentation of the on-going upper ocean floating sediment trap work at both BATS and HOT with a modest 234Th calibration effort.

d. Parameters for which direct intercomparison is warranted

Findings:

Although nutrient analyses are considered routine, there is concern that the international oceanic community is not making these measurements very accurately. Both the BATS and HOT program do routine analyses of nitrite, nitrate plus nitrite, silicate, and phosphate using well-established colorimetric methods and Technicon Autoanalyzers. There is a slight protocol difference between the two programs since the HOT program does not filter samples and the BATS program does using an 0.8 µm Nuclepore filter; this should not cause a significant difference between the two efforts. On at least two occasions, paired samples were collected and analyzed by both programs. The results were only fair with some large discrepancies found for all analytes. However, it must be recognized that the intercomparisons were not performed under ideal conditions or with careful enough controls for this intercomparison to be reliable. Both programs participated in an international community methods comparison for nutrients run by Kirkland, but it is evident that the analyst involved for the BATS program was not representative of the current effort there. Both programs analyze particulate C/N using similar procedures; filtration on 25 mm GF/F filters, storage of filters in foil in a deep freezer, and later analysis by high temperature combustion gas chromatography. The BATS program acidifies filters at the time of analysis to remove inorganic carbon; the HOT program does not. The BATS program uses a CEC 240-XA Elemental analyzer and the HOT program has previously used a Perkin Elmer 2400 CHN analyzer and is currently converting to use of a Europa Roboprep system with gas chromatography. Although there is no reason to suspect significant differences in the particulate C/N analyses of the two programs, there is need for demonstration that the two are measuring the same thing since there has never been an extensive methods comparison for particulate analyses. JGOFS process cruise data show that POC/PON measurement results are highly dependent on equipment used, sample handling, filtration methodology and even the investigator. Given that different methodologies are used at the two sites, it is unknown if differences seen as dependent upon differences in methods

Recommendations:

Particulate C/N, and nutrients should be subject to direct intercomparison by using paired samples from both stations run at both facilities. A single individual should collect or supervise collection of the paired samples on both BATS and HOT cruises. Personnel at both programs should be prepared to perform the analyses in a fashion representative of what is normal for that program and exchange data in a timely fashion. Comparisons need to be thorough with respect to sample collection and handling, including filtration steps, with must be documented in detail. Direct comparisons, especially for the parameters without good reference materials or accepted standards, are very important and should be viewed as a high priority.

e. Non core measurements

Findings:

There are other parameters not considered core and that are measured regularly on only one of the time-series stations. These include low-level nitrate and phosphate measurements, done at HOT but not at BATS. Dissolved organic phosphorus and particulate phosphorus are measured at HOT but not at BATS. As already mentioned above, HOT measures pH while BATS does not. Considerable optics measurements are made at BATS, but not at HOT.

Both programs have deep moored sediment trap components that are technically considered outside the core program. In addition, both programs have auxiliary research efforts measuring macrozooplankton populations, while neither is assessing microzooplankton populations.

There are yet other measurements that while not routine in either program, should be considered in at least a pilot fashion. For example, recent presentations relating to primary production at both sites have indicated that the role of nitrogen fixation may be substantially greater than previously expected in these gyres, and yet this process is not being routinely measured.

Recommendations:

Thought should be given to how important it is to know fine details of the nutrient cycles in the photic zone. If found to be important, perhaps low-level nutrient measurements (nitrate, phosphate at BATS, and possible ammonium at BATS and HOT) should be added as core parameters. If not, routine low-level nitrate and phosphate measurements at HOT might be discontinued if they displace higher priority parameters.

Since the major questions in long term trends at these two ocean stations pertain to carbon fluxes, every effort should be made to monitor the carbon cycle with maximum sensitivity. The relatively simple colorimetric pH measurement is a major analytical breakthrough. The precision of this method is on the order of +/- 0.001 pH units and the absolute accuracy on the order of 0.005 (and this could improve very soon with better calibrated buffers and with use of Dickson's CRM). The current measurement precision of pH is comparable to that of DIC and perhaps better than that of alkalinity. Considering the uncertainties that still remain in the thermodynamics of the carbonate system, the addition of pH measurements to the suite of carbon system parameters will improve our overall constraints on ocean carbon. Since pH can be measured with a precision similar to that of DIC and probably with greater precision than alkalinity, it will constrain the carbonate system to have this measurement. Thus, pH should be added as a core parameter , continuing the effort initiated in the HOT program and adding this measurement to the BATS program.

Since both stations are measuring dissolved organic and particulate carbon and nitrogen, but only HOT is measuring dissolved and particulate organic phosphorus, consideration should be given to adding organic phosphorus as core parameters and establish these two measurements at the BATS program.

An evaluation should be made of microbial production measurements considering accurate estimates of photoautotrophic and heterotrophic carbon utilization and whether more significant information could be obtained with routine measurements of nitrogen fixation.

Serious consideration should be given to the value of the macrozooplankton and deep sediment trap studies. If they are essential components of the information base required for these programs to meet JGOFS objectives relating to the processing and downward flux of particulate carbon, then they should be added as core parameters. This would both ensure that high priority is given to of these measurements and enhance comparability between the two programs.

B. Data Products and Their Distribution

Overview:

Both time-series stations have been collecting data for nearly 8 years, most of that time without seeking assistance from the JGOFS Data Management Office with respect to on-line distribution of data, data parameter naming and formatting consistency. As the U.S. JGOFS program begins its synthesis phase, it is important to have a time-series database that is as up to date as possible, easily accessible and well organized.

Findings:

As of January 1996 BATS and HOT data available to the community in digital form were current to September 1993 and mid 1995, respectively. The value of time-series data is their currency and high quality. Excessively long delays in quality control procedures can compromise the value of the time-series data set to program PIs.

Recommendations:

Every effort should be made to quality control and make core measurement parameters available as fast as possible in digital form. BATS and HOT should strive to have data ready for release approximately 6 months after each cruise. In no case should data availability lag beyond the required two year period as formulated by U.S. JGOFS and NSF policy. Hard copies of station data reports should continue to be produced and distributed as in the past, but should not delay the electronic release of the data.

Findings:

The BATS data are inconsistent with regard to naming parameters from year to year. Parameters were added to the data over the years, creating a difficult puzzle to sort out when determining whether "p11" from year 2 is the same as "p11" from year 5. The data team at Bermuda has worked out these problems for their MATLAB tool by utilizing 'knowledgeable' .m files. Once the data are loaded into .mat files (binary matrix) using these .m (procedure) files, the data are regularized, i.e., wherever, and by whichever name salinity is stored in the input file, it is always salinity in the .mat file.

BATS and HOT are inconsistent with one another in naming of JGOFS 'core' parameters.

Recommendations:

Both time-series sites should make every effort to work with the U.S. JGOFS Data Management Office with a goal of serving their time-series data by way of the U.S. JGOFS data system. If possible, these data should be served from the machine at each of the sites where the most-final (or current version) of the data reside. This will best assure the community that the data being accessed is that which has received all known corrections. Programming effort on the part of BATS and HOT will be required.

Software (methods) serving data from the two sites should rectify the problem of variable naming differences between the two data sets wherever possible and name core data variables consistent with JGOFS standards.

Findings:

BATS and HOT parameter units are inconsistent with one another. For example BATS particulate organic carbon data are reported in units of µg/kg where as HOT uses µmol/kg.

Recommendations:

The two time-series stations should adopt consistent units for reporting core JGOFS parameters, and the mole unit standard of JGOFS should be adopted. For non core measurements, agreement should be reached for common names.

Findings:

Each of the sites has developed a MATLAB-based analytical and graphics tool, for use locally to facilitate data analysis and quality assurance procedures. These are very useful to personnel working with these data on site, and may prove useful to interested investigators using MATLAB on their own computers.

Recommendations:

To the extent that is feasible and allowed under licensing with The Mathworks, Inc., the .m files comprising the BATS MATLAB tool and the HOT MATLAB tool (known as HOT-DOGS) should be shared with the scientific community.

Findings:

Both HOT and BATS have ancillary measurements collected as part of their ongoing science and supported with JGOFS funds. A renewed effort should be made to incorporate these data into the time-series data bases. .

Recommendations:

Tony Knap and Dave Karl should send letters to the investigators who have received funding to make ancillary measurements at their respective sites, asking that these data be submitted to the U.S. JGOFS Data Management Office. These letters will be copied to both the U.S. JGOFS Executive Committee and Christine L. Hammond, as head of the U.S. JGOFS Data Management Office.

C. Resolving Spatial Variability

Overview:

Over the last decade, the ocean biogeochemical community has grown increasingly aware that spatial and temporal variability occurs over a wide range of scales. This shift has been driven by the deployment of new instrumentation (e.g. biogeochemical moorings; SEASOR spatial surveys) and by field programs such as the JGOFS North Atlantic Bloom Experiment and the time-series program. As at any other location in the ocean, the observed signals at the time-series stations are influenced both by local change and advection. It is crucial that we characterize these three-dimensional effects to determine the representativeness of the time-series data to the larger subtropical gyre scale and to constrain the underlying biogeochemical dynamics. That is, to fully utilize the valuable resource of the time-series data, we need to better understand the regional context in terms of the impact of both large-scale and mesoscale circulation patterns on the local biogeochemistry.

Findings:

The potential for aliasing of the two time-series records due to the proximity to land, the so-called "island effect", has been recognized since the conception of the program and has been raised in previous oversight committee reports. Both programs have attempted to address this issue through a combination of near-shore stations (Kahe and Hydrostation S) and transects away from the islands. Bermuda has also used their annual validation cruise to sample a grid of stations about the BATS site. These results, while highlighting the existence of large-scale gradients and mesoscale variability about the time-series sites, also suggest that the direct island biases at the time-series stations are minimal.

Recommendations:

The data from the earlier validation and transect cruises should be released to the community, and where possible the resources and ship-time should be made available to continue such efforts at least on an annual basis. The current underway sampling programs carried out in transit to the time-series stations are also relevant in this regard and should be continued and/or expanded where possible. In particular, the BATS program should explore the possibility of acquiring at a minimum a subsurface profiling capability similar to the towed optical plankton counter currently at HOT.

Findings:

One increasingly apparent result from the time-series stations is that the biogeochemical cycles at both Hawaii and Bermuda can not be interpreted in strictly local or one-dimensional contexts. The inter-cruise variability at each station is strongly modulated by the advection of mesoscale events passed each station. Further, the important underlying mechanisms governing the overall carbon cycle may depend on three-dimensional processes. The ecosystem dynamics at both stations appear to be driven in part by vertical nutrient fluxes from mesoscale "eddy-pumping", while at Bermuda the passage of a large-scale, seasonal advective feature has been suggested as a possible cause of the large, and as of yet unexplained, summer draw-down of total inorganic carbon at Bermuda.

Recommendations:

The on-going time-series program should be supplemented with directed, regional process studies to address the specific three-dimensional scientific questions raised by the time-series data (i.e., eddy variability and nutrient pumping at both Hawaii and Bermuda, and carbon budget closure at Bermuda).

Findings:

A three dimensional or regional context for the mesoscale variability in the time series data can be diagnosed from satellite images of SST and cloudiness. Both time series efforts operate in close collaboration with satellite observation groups who maintain exemplary data archives on the world wide web, but satellite imagery is not clearly available corresponding to each of the time series cruises.

Recommendations:

The collection and archiving of satellite images should be incorporated into the dissemination of shipboard time series data. Perhaps the responsibility for archiving images directly relevant to time series data should be shouldered by time series data management personnel.

Findings:

Despite the submission of a number of 3-D process study proposals affiliated with the time-series programs, no direct funding for such work has become available though conventional pathways. The lack of funding has arisen in part because of the perceived large costs, relative to regular single or multi-authored proposals, associated with 3-D studies and because of questions regarding the actual implementation of such field work. Many of the problems and issues surrounding the feasibility of 3-D studies were raised in a JGOFS planning documents, and it appears that many of the technical hurdles may be overcome by combining field data with the new generation of data-assimilation models.

Recommendations:

Moderate-scale regional process studies for the time-series should be considered as a component of the Synthesis and Modeling Program phase of U.S. JGOFS. As a first step, a detailed implementation plan for an integrated 3-D program should be developed for one or both of the time-series stations integrating field work, mooring and satellite remote sensing data, and numerical modeling.

Findings:

Basic surface meteorological data (winds, air temperature, humidity, longwave and shortwave flux) are needed in order to interpret and model the physical variability observed at the two time-series stations. Unfortunately, because of the lack of mooring capabilities during the early years of the time-series, continuous meteorological records are not generally available. This situation has recently improved with the installation of IMET packages on the two ships regularly used by BATS and HOT, and with the inclusion of meteorological packages on the proposed moorings.

Recommendations:

Shipboard meteorological data are generally treated as core measurements in the U.S. JGOFS process studies, and BATS and HOT should endeavor to process and incorporate both the ongoing and historical shipboard IMET data into their data bases. Ancillary data, such meteorological tower data from Bermuda and Hawaii, should also be released if possible. All future moorings should include, at a minimum, a standard surface meteorological package.

D. Resolving Temporal Variability with Moored Arrays

Findings:

The paradigm of a low productivity, stable oligotrophic ocean is called into question by recent observations from the time-series sites at Bermuda and Hawaii. Recent evidence, from deep sediment traps and continuous moored observations, suggests episodic inputs of nutrients and subsequent responses by phytoplankton, that have been mostly missed by the monthly shipboard sampling.

Recommendations:

In order to better resolve episodic events moored observations of a few critical properties should be made an integral part of the time-series programs.

Findings:

There have been numerous efforts to gain funding for moored observations of bio-optical chemical and physical properties at Bermuda and Hawaii. To date, reviewers and funding agencies have considered these proposals difficult to support. The Oversight Committee believes that this stems partly from the general scientific community either not trusting the measurements and/or underestimation the utility of moored observations. Through different means, the Bermuda and Hawaii sites have managed to acquire moorings for deployment at the time-series sites. Several deployments have been made at Bermuda while the initial deployment at Hawaii is scheduled for Fall 1996. These efforts are currently under-funded.

Recommendations:

While the current mooring efforts are greatly under-funded, it is crucial that these efforts continue. The moored data should be made available to the community as soon as possible in an effort to convince skeptics of their potential use.

Findings:

Biological and chemical instrumentation for moored application has only recently become available and many of the measurements are mostly proxies for those made on ships. There are new instruments expected to come on-line over the next several years that will be able to be used in situ and may partially replace the shipboard measurements

Recommendations:

Co-principal investigators and private corporations, with expertise in moored instrumentation, should be approached and invited to use the Bermuda and Hawaii sites as test beds for instrument development.

V. RESOURCES

A. Laboratories and Ships

BBSR and SOEST have made extensive use of existing facilities, and augmented these where necessary, to support the JGOFS time-series station activities. The Oversight Committee was impressed with the efficient use of space and technical assistance at both host locations. Over the course of this work significant improvements have been made with regard to ship capabilities at BBSR. The BATS program has been well-served by the R/V Weatherbird II. Her recent modifications to the labs and staterooms have improved her at-sea capabilities, and she should be able to provide adequate support for the BATS core measurements, ancillary investigators and mooring activities into the next century.

Regrettably, the situation is less stable at SOEST. To date, the HOT program has used 11 different UNOLS, non-UNOLS government, and private charter vessels to support their field sampling programs. However, the University of Hawaii's flagship, the UNOLS research vessel Moana Wave has been the most commonly used platform supporting 43 of the first 75 program cruises and all mooring activities. She is an excellent platform fully capable of supporting the HOT program science requirements. The recent additions of a differential GPS system that will enhance the quality of the existing ADCP data sets, a new suite of shipboard meteorological data sensors, a well-designed uncontaminated seawater intake system and a portable conducting cable winch for the deployment of optical profiling equipment will make this good ship even better. Because most of the crew members have sailed with HOT program scientists many times, the shipboard needs of this science are well understood. These cruises run efficiently with a high yield per unit effort, and experience little down-time during their four days of operation. When the Moana Wave is at sea supporting other science programs, the HOT scientists need to use an alternate, typically less satisfactory, platform to continue the scheduled time-series observations.

The fate of RV Moana Wave beyond 1997 is uncertain. There is a possibility that she will be retired and SOEST will acquire instead a Class 1 vessel suitable for global ocean projects. The Oversight Committee registers its concern that such a vessel, although perhaps a fine asset for SOEST, is far from an answer to the needs of the HOT program. A glance at the ship schedules for major vessels in the UNOLS fleet gives little clue as to the locality of their home ports. A ship suited and scheduled for local waters is essential for the HOT program to remain credible. The JGOFS Steering Committee should make its views known on this matter, both to SOEST and UNOLS.

B. Schedule for Core Funding

BATS core support: Current funding through April 30, 1998. Renewal proposal has been submitted to continue this work from May 1, 1998 to April 30, 2001.

HOT core support: Current funding through July 31, 1998. Renewal proposal has been submitted to continue this work from Aug. 1, 1988 to July 31, 2001.

WOCE core support: Current funding through July 31, 1998. Roger Lukas will submit a renewal proposal in 1997 for a continuation of this effort.

VI. REFERENCES

Hayward, T.L. (1987) The nutrient distribution and primary production in the central North Pacific. 34, 1593-1627.

Jenkins, W.J. and J.C. Goldman (1985) Seasonal oxygen cycling and primary production in the Sargasso Sea. Journal of Marine Research, 43, 465-491.

Menzel, D.W. and J.H. Ryther (1961) Annual variations in primary production of the Sargasso Sea off Bermuda. 7, 282-288.

UNESCO (1994) Protocols for the Joint Global Ocean Flux Study (JGOFS) core measurements. Intergovernmental Oceanographic Commission, Scientific Committee on Oceanic Research, Manual and Guides, 29, 119-124.

VII. APPENDICES