INTERGOVERNMENTAL OCEANOGRAPHIC WORLD METEOROLOGICAL COMMISSION
( of UNESCO) ORGANIZATION

SUMMARY REPORT
SECOND AD-HOC CONSULTATIVE MEETING
ON THE GLOBAL TEMPERATURE-SALINITY PILOT PROJECT
(A PROPOSED COOPERATIVE IGOSS-IODE PROJECT)

Marine Environmental Data Service
Ottawa, Canada
25-28 July 1989

(July 1989)

TABLE OF CONTENTS FOR THE OTTAWA MEETING

1. Opening remarks and election of Chairperson

2. Presentations

2.1 Overview of the GTSPP

2.2 Overview of TOGA Data Management

2.3 Overview of JEDA Center

2.4 VNIIGMI-WDC USSR Summary

2.5 AODC Summary

3. Implementation of BUFR

4. Review of the Status of Implementation of Communications

5. Specification of Quality Control Procedures

6. Review of GTS Monitoring

7. Review of the Scientific Requirements Document

8. Role of GTSPP in the WOCE Data Management System

9. Review of the Project Implementation Plan

9.1 Table of Contents for the Project Plan

9.2 Relationship between the French WOCE DAC and the GTSPP

9.3 VNIIGMI-WDC-B Participation in GTSPP

9.4 Schedule and Review Process for the Project Plan

9.5 Project Management

10. Review of Historical Data Aspects of the Project

11. Review of the Action Plan

12. Other Business

13. Approval of Summary Report and Recommendations of the Meeting

14. Closure of the Meeting

Annexes for the Ottawa Meeting

Annex A - Agenda

Annex B - List of Participants

Annex C - GTSPP Project Goals

Annex D - List of Documents

Annex E - Provisional Agenda

Annex F - Table of Contents for the Project Proposal

GENERAL SUMMARY OF THE WORK OF THE OTTAWA MEETING

1. Opening Remarks and Election of Chairperson

The meeting was opened on Tuesday, July 25, 1989, by Dr. W.G. Doubleday, A/Assistant Deputy Minister for the Canadian Department of Fisheries and Oceans, who welcomed the participants (Annex B) from the international data centres and oceanographic research community. Dr. Doubleday reviewed a number of unique characteristics of the GTSPP including the concept of the continuously updated database scheme within which more accurate data replaces earlier versions of the telecommunicated data, thus providing increasingly more useful data for science programs. Dr. J.R. Wilson, Director of the Marine Environmental Data Service (MEDS), also greeted the participants, noting the increased interest in the Project expressed by representatives from France, the Soviet Union, the Tropical Ocean Global Atmosphere (TOGA) and World Ocean Circulation Experiment (WOCE) offices, and other agencies since the first meeting of the ad-hoc working group held in January 1989 in Washington, D.C., USA. Mr. Gregory Withee, Director of the U.S. National Oceanographic Data Center (NODC) was unanimously elected Chairman of the meeting by the participants.

The Chairman outlined a brief history of the international oceanographic community's interest in creating a timely global ocean temperature and salinity dataset of known quality in support of the World Climate Research Program (WCRP). This history dated back to the 1981 IODE meeting in Hamburg, Federal Republic of Germany, the 1984 IGOSS/IODE meeting in Tokyo, Japan, leading to preliminary discussions by the Australian Oceanographic Data Centre (AODC), the MEDS, and the NODC during an IGOSS-IODE January 1988 meeting held in Ottawa, Canada on IGOSS/IODE data flow. Discussions continued at the June 1988 workshop on ocean data files convened in Washington, D.C., USA, and the pilot project was formally recognized and endorsed at the November 1988 meeting of IGOSS. The first ad-hoc consultative meeting was subsequently held in January 1989 at the NODC in Washington, D.C., USA.

The purposes of the present meeting were: to gather information required in view of preparing a project overview document which could be presented at the January 1990 IODE meeting in New York, USA in order to seek formal endorsement for the GTSPP by the IODE; to improve plans for the Project; to discuss quality control guidelines for all participants; and to discuss communication/exchange formats for temperature and salinity data to be interchanged on a global basis.

2. Presentations on the GTSPP and TOGA Sub-Surface Centre

Under this agenda item five informational presentations were made to the meeting.

2.1 Overview of the GTSPP

IGOSS has been exchanging temperature and salinity data in near real-time using the GTS for many years. These data are compiled by special centres and sent to the IODE system for long term archiving. Today new developments can be applied to improve this data management. The use of compact formats (such as BUFR) for the GTS improves the flexibility of transmission. New telecommunication circuits and software can protect the integrity of the data. Work stations and interactive graphic software improves the data quality assessment process. Inexpensive mass storage devices, such as CD-ROM provide for distribution of the data in a form that can be read on inexpensive readers on small computers. It is the intent of the GTSPP to take advantage of all of these improvements to develop a prototype data management scheme for temperature and salinity data which will serve as a forerunner to International data management projects of the next decade.

The Project goals (Annex C) were re-stated and suggested real-time and delayed mode data flows for the GTSPP were reviewed. It was stressed that several operational requirements were prerequisites in order to address global science problems. In particular, the need to increase the participation of the oceanographic community in real-time data submissions, the need to acquire delayed-mode data faster within data centres, and the need to accurately monitor all the data collected.

2.2 Overview of TOGA Data Management

An overview of the TOGA project was presented by Jean-Paul Rebert, Director, TOGA Sub-Surface Data Centre, ORSTOM in Brest, France who outlined its successes since inception in 1985. An accumulated dataset of over 100,000 oceanographic profiles for the tropical area from 30 S to 30 N since 1985 was noted. Also noted were the delay times in months to years in data reception of delayed mode versus real-time data at the TOGA data centre.

The meeting noted that in part, the GTSPP was based on the successes of the TOGA Subsurface System, and that the data flow patterns, i.e. between scientific institutes and data centres, were similar. The meeting recommended that the GTSPP take care not to interfere in any way with the existing TOGA data flow.

Interest by other GTSPP participants centred on the TOGA quality control procedures, particularly with respect to the potential integration and comparison of model results with in-situ observations.

2.3 Overview of the JEDA Center

The JEDA Center is a joint centre co-located at the Scripps Institution of Oceanography (SIO) in San Diego, and at the National Oceanographic Data Center (NODC) in Washington, D.C. NODC's role in the JEDA Center is to be responsible for the collection and dissemination of subsurface thermal data; as an operational centre, NODC is designed to be more responsive to data requests than the JEDA Center at SIO would be. The role of SIO in the JEDA Center is to provide for scientific quality control of the subsurface thermal data and to prepare monthly analysis products.

Statistics were presented which showed that roughly 30% of delayed-mode data, after 2 years from the observation times, were never received. The decline in total worldwide data collection since 1980 were also noted. The declining funding for XBT probes was presented as the major reason for this latter trend.

2.4 VNIIGMI-WDC USSR Summary

Soviet initiatives and their decision to participate in this project (see Section 9.3) were welcomed by the participants. Of note were the one million worldwide oceanographic profiles presently archived, and the interest to participate in the GTSPP by striving to exchange data within 6 months of observation. Recommendations on a 4-level data structure related to data precision were noted, as well as Soviet advances in modern statistical methods for quality control of oceanographic observations (see Section 9.3).

2.5 AODC Summary

Australian representatives explained the possibility of the Australian Bureau of Meteorology expanding the area of coverage for their subsurface thermal analysis maps. In relation to the operation of the Australian SOC, a quality control manual for XBT data is being developed. Participation in GTSPP would include data exchanges of quality controlled data within six months of observation. It was also noted that interest in GTSPP has been expressed by the Australian Institute of Marine Science.

3. Implementation of BUFR

Technical information related to the Binary Universal Format for data Representation (BUFR) was made available to the participants. Limited practical experience was reported on the use of the BUFR code amongst GTSPP participants, however independent test results by the NODC and the Soviet participants indicated significant reductions in transmission volumes could be expected from its use.

The participants identified the need for approval of extensions to the BUFR coding system and code tables for oceanographic requirements including quality control and that these would have to be formulated and submitted to the WMO Commission on Basic Systems for approval. It was noted that such extensions would benefit both meteorologists and oceanographers since both communities needed to use telecommunicated data from the other.

The participants also recognized the requirement for software to decode BUFR formatted data into symbolic formats in ASCII, as well as to develop a capability to convert BUFR to the GF-3 format and vice versa.

The recommendation from the previous meeting to adopt the BUFR format as the GTSPP standard for exchange of telecommunicated data was endorsed by all participants. The reasons for this choice were:

a) BUFR has international acceptance by the IOC/WMO member countries, particularly the meteorological community, and will be used extensively by the major weather forecasting centres;

b) existing international communications facilities (GTS) presently accept BUFR message traffic;

c) flexibilities of the self-documented BUFR code can provide for the inclusion of quality flags within the messages which is a requirement for the GTSPP;

d) the relative ease in adopting oceanographic parameters within the present BUFR coding system; and

e) the availability of compression techniques which suit oceanographic requirements.

Finally, there was a recommendation that all GTSPP data centres review their individual requirements for respective internal processing formats in light of new requirements to handle quality flags, data exchanges in BUFR code, and to meet overall GTSPP monitoring requirements. It was noted that adoption of similar processing formats would simplify the exchange of software between participants. This item is to be discussed again at the next meeting.

4. Review of the Status of Implementation of Communications

The first Ad-hoc Meeting on the GTSPP identified the requirement for Canada to implement a connection to the NASA Space Physics Analysis Network (SPAN) in order to implement the necessary telecommunication of data to and from the US NODC, the SIO, and the AOML. The meeting noted the importance of communications to the Project and encouraged all GTSPP centres to seek interconnections to common telecommunications networks.

Canadian participants reported on a joint effort of the Canadian National Research Council, the Department of Energy Mines and Resources, and the Department of Fisheries and Oceans, to establish a Canadian SPAN (CSPAN) communications network in Canada with access to the US SPAN. MEDS expects to begin testing this link in October 1989 and will report on its progress at the next meeting.

5. Specification of Quality Control Procedures

The first ad-hoc meeting of the GTSPP recommended the production of a manual which would detail all of the quality control procedures that were to be applied by project participants. This was deemed necessary because quality control was going to be applied at different locations and a consistent approach was desirable. MEDS was assigned the task of producing a draft manual for consideration.

Discussions covered a range of issues related to the latest draft of a GTSPP Quality Control Manual and some wider issues. The following issues and actions were decided:

a) It was agreed that a participating data centre would preserve the original data values as received so that they can be available to users. This will allow for a complete reanalysis of the data at some future time. The method by which these data would be made available was considered to be an issue of implementation which individual centres can decide for themselves.

b) It was recognized that if the original data are preserved and available, it is permissible, in certain circumstances, to change data as long as the change is identified. The present draft Manual recommends changes only when the identification, position or time information are clearly wrong. However, in the context of real-time data, identifiers of unknown platforms must be preserved as received and be flagged in some fashion to indicate that they are considered to be erroneous. While no decision was reached, the next draft of the Manual will relax the restrictions imposed on changes that are allowed and solicit comment.

c) The concept of a minimum level of quality control through which data must pass before they are available to users was deemed unnecessary in light of the decision reached about preserving data as originally received. It was considered more important that clear documentation of the quality control steps applied to the data be available to users at all stages of processing and that a record of the stages passed be preserved with the data.

d) The choice of IGOSS flags for the Manual was welcomed. The general concept of a version parameter indicating the stages of quality control through which data have passed was considered appropriate except that an extension was made to permit centres to apply procedures not documented by the Manual. The discussion on this point attempted to balance the desire of data centres to use all facilities at their disposal in assessing data quality. Resolution on this matter was not reached.

The problem is that a data centre often does not have the flexibility within its processing system to allow the skipping of certain quality control procedures even though another centre has effected them. So, for the moment, there will be some duplication of effort between participants. On the other hand, the existence of a Manual will permit centres to focus their software development activities for the project to those areas where they will need to upgrade their capabilities. In so doing, it may be possible that software developed at one centre for one stage of quality control processing will be adaptable for use at another centre and thereby developments can be shared. This latter is useful to optimize resource utilization as well as to deal with data received which do not fall within the context of the GTSPP.

As the Project advances, centres will rely more and more on other participants to effect quality control procedures described by the Manual. New procedures will be suggested and adopted as the Manual is upgraded. The inclusion of new procedures into the Quality Control Manual will require a formalism for adoption. It is recommended that at the same time that the first version of the Manual is agreed to, a review mechanism be established to ensure the Manual is kept current.

e) It was agreed that quality control flags should be kept with data at both the individual data value and at a profile level. It was agreed that users be permitted the flexibility to access data based on the flags and that the flags be included in any data exchanges between centres and users.

f) The group endorsed the style of the draft Manual. A few specific issues were raised concerning the tests in the draft Manual. These issues are included below for completeness. Suggestions were also made about further tests that could be included. These will be forwarded to Mr. Keeley of MEDS. While not expressed explicitly by the group, it was clear that the present version does not represent the final draft. Accordingly, Mr. Keeley will need to consider the comments made and incorporate these into another draft. No timetable was discussed.

g) Discussions reflected a need to preserve not only data but also information about the data, referred to as meta data. This information is useful to both data centres and users in assessing data quality. Any data handling system must be capable of managing both the data and the meta data. However, for data qualification and retrieval purposes, a simple data flagging scheme is needed, and the one proposed in the draft Manual was recommended.

The following specific comments were made.

- The use of the word "level" was considered to be confusing with other uses of this word in data processing. The word "phase" was suggested.

- It is desirable that the model established by GF3-PROC for freely available software be pursued here.

- There is no problem in being able to both upgrade and downgrade flags as the result of further tests conducted on the data.

- More attention must be paid to ensure test procedures apply to both real-time and delayed-mode data. It must be clear how each test may be applied to data from the different sources.

- Attention must be paid to the resolution of the problem of dealing with near and exact duplicates.

- Test descriptions should discuss how data must be ordered prior to each test.

- The Manual should have some discussion of the hierarchy of tests, including scientific tests, that would occur after the testing described by the Manual.

- A graphic representation of the entire decision tree of the tests would be desirable.

- The current group helping in the preparation of the draft Manual needs to be confirmed. The present membership is Warren White, Syd Levitus, Ben Searle, Susan Kerig, Jean-Paul Rebert, Nikolai Mikhailov, Bob Fish, Chris Noe, John Withrow, and Bob Keeley.

- The Manual should be very clear in stating that when a version flag is attached to data it means that all tests described in the Manual at that level and at all lower numbered levels have been performed.

- Interpolation schemes must be considered explicitly by the Manual. Participants at this meeting will forward information to Mr. Keeley.

- The text of the Manual needs a clearer description of the use of version and QC flags including indications of scientific quality control applied to the data.

- The Manual needs to address the identification of duplicates to the same level of detail as the QC tests in the present draft.

Because of the interest in the quality control measures being used at the TOGA Centre in Brest, the meeting requested Jean-Paul Rebert to bring a document describing these procedures to the next GTSPP meeting. NODC and AODC also agreed to provide a document of their quality control procedures for the next meeting.

6. Review of GTS Monitoring

Results of an IGOSS GTS monitoring exercise carried out during a week in July 1989 were discussed. MEDS also presented results of its own monitoring exercise which compared IGOSS BATHY-TESAC traffic received at the NODC and at MEDS independently for January and February, 1989. While MEDS results showed an encouraging pattern, the exercise identified discrepancies of the order of 20% in data reception between the two centres. The Soviet participants reported similar incidences of data losses over the GTS.

It was decided that MEDS conduct a monitoring intercomparison study amongst GTSPP participants (Australia, Canada, France, USA, and USSR) for three months of GTS BATHY-TESAC data (Sept., Oct., Nov., 1989) and distribute the results to the participants.

To effect this intercomparison, it was decided that countries presently involved in GTSPP assist by recording each month's BATHY and TESAC data received and transmitted on GTS at their GTS Centre. This data should be forwarded, on a monthly basis, to MEDS. MEDS will undertake a comparison of each data set, removing duplicates and retaining all unique observations. With a minimum of reformatting, computer system utilities could be used to sort and compare data from each source. Information could also be obtained on the number of duplicates and on the number of `unique' observations available from the GTS. Doug McLain volunteered to request the same from the European Centre for Medium Range Weather Forecasting (ECMWF).

In addition, it was recommended:

a) that the IOC Secretariat should be encouraged to distribute the "IGOSS Statistical Evaluation Sheet" on a monthly basis to all participants, even if these reports are only compiled 2 to 3 months after the month in question. This would allow GTS users to verify the results;

b) that monitoring of delayed-mode data amongst GTSPP participants should also be carried out at an inventory level of information.

Further to this point, the existence of different streams of real-time, delayed-mode and historical data, which can be managed differently by the Data Centres participating in the GTSPP, increases the need of a clear knowledge of the status of the data sets collected, processed and exchanged within this framework.

It was therefore recommended that all the Centres participating in the GTSPP create, maintain and update regularly, an inventory of their data. These inventories should be built in such a way that they can be easily compared to those of other Centres.

The structure and content of statistical tables easily manageable on micro computers and containing relevant information common to all these inventories will be defined, prepared, and submitted to the next meeting for discussion; and

c) that a scheme for the systematic monitoring of quality of upper ocean observations be elaborated. Statistics on the quality flags attributed to observation reports should be consolidated in time for each platform, in order to detect the occurrence of any flaw in the data coding, the measurement or the equipment, which may be the cause of systematic errors. These monitoring results could be fed back to the Principal Investigators or Manager of the observing system for corrective action.

7. Review of the Scientific Requirements Document

The participants discussed the review document by S.L. Patterson on this subject and were in general agreement with the majority of its content. Participants agreed that GTSPP would not meet WOCE WHP quality and precision standards but that on the other hand, real-time data would be helpful to WOCE in regards to its data management requirements. The participants noted the possible future expansion of GTSPP based on acquisition of new oceanographic parameters and welcomed the guidance in this area. The GTSPP, however, did not agree that these datasets were "targetted more at the retrospective data users", although this was indeed a component of its potential user community.

Steve Pazan noted that low-density VOS lines supply data with aereal coverage sufficient to estimate large scale fluctuations in upper-level heat content, which is necessary for flux studies. The high density lines are incapable of such heat content estimates; they are designed to measure horizontal fluxes and mesoscale fluctuations.

The GTSPP participants felt that this document would be useful as an overall guide for the development of the Project, but would require greater elaboration/clarification on the "meta" data and "timeliness" components to suite its particular needs. The US NODC agreed to continue to try to improve the document and make a revised draft available to the workshop.

8. Role of the GTSPP in the WOCE Data Management System

Allyn Clarke, from WOCE, and Bert Thompson, from the WOCE IPO, gave overviews of the experiment's goals and specific objectives, the proposed operational program of core projects, and the observational component associated with these core projects. The meeting noted that quite a lot of information about the experiment is available through OMNET (address WOCE.IPO) and that information on data were available through the Data Information Unit at the University of Delaware, U.S.A. in the form of an on-line catalogue.

Allyn Clarke noted that GTSPP data would be of considerable use to the WOCE in its core project for global descriptions on the evolution of the heat and salt distributions in the upper oceans. He also noted that present XBT sampling in the Pacific and the tropical oceans (TOGA) were close to meeting present WOCE upper ocean thermal requirements and that the GTSPP could assist further through acquisition of ICES fisheries salinity data and other additional XBT and salinity data. Another area in which the GTSPP could assist would be in identification and acquisition of time series transects, particularly in the South Atlantic, to supplement present project objectives (e.g. PANULARIS, LINE P, etc.). Steve Pazan noted that increased sampling of salinity in the western North Pacific, the western North Atlantic, and in the tropical Oceans (TOGA) would be necessary to estimate interannual geostrophic dynamics.

The meeting noted the high accuracy requirements (0.002 C, + 0.002 PSU) for temperature and salinity data in deep ocean (greater than 2000m) by the WOCE WHP, but that 200-2000m precision requirements could be relaxed to (0.02 C, 0.05 PSU). For some WOCE applications, (0.3 C, 0.3 PSU) for depth less than 200m, assuming there are no systematic errors at these levels, could be useful data due to the large variability at these depths.

With respect to WOCE data management, the meeting noted the general wish of principal investigators not to provide high precision datasets to the general community and that such data would not be available for at least the first year after the project. There was, however, some possibility for GTSPP access to WOCE data with respect to the Upper Ocean Thermal objectives where both principal investigators and the VOS programs are expected to participate in both real-time and delayed-mode aspects of the program. The same comments were also noted in the US-WOCE presentation by Scripps. There were suggestions that TESAC formatted versions of the data could be released as "quick look, less precise" data useful for many purposes without jeopardizing the scientist's first right to publication.

The relationship of the French WOCE DAC and the GTSPP is discussed in Section 9.2 of this report.

9. Review of the Project Implementation Plan

9.1 Table of Contents for the Project Plan

The meeting examined the need for a number of documents in order to achieve the previously defined aims. However, it was agreed that only one document would be produced, having three main sections and one annex, the annex giving credit to those agencies and individuals that constitute the Project Planning Contributors.

It was agreed that a document with the following title was the most appropriate at the moment:

"Global Temperature-Salinity Pilot Project
A Cooperative IGOSS/IODE Project"

The purpose of this document is to promote GTSPP, encourage participation, and seek support and endorsement from the parent bodies (IGOSS/IODE), the scientific community and existing projects and experiments such as TOGA and WOCE. The project plan would also serve as the principal document for the upcoming GTSPP workshop. (A proposed agenda for the workshop is included as Annex E.)

The table of contents for a revised document were discussed and are included as Annex F.

9.2 Relationship between the French WOCE DAC and the GTSPP

The meeting then discussed the relationship of the French WOCE DAC to the GTSPP.

France is involved in a number of activities including:

a) - operation of a TOGA/WOCE Upper Ocean Data Centre;

b) - operation of a TOGA Data Centre;

c) - participation in the GTSPP.

These functions were described at a conceptual level as being independent. However, it was recognized that there will be a lot of interaction between these activities. The Working Group stressed that it was important to realize that these are co-ordinated projects that are assisting and complementing each other. Data flow within each area will not be in conflict.

9.3 VNIIGMI-WDC-B Participation in GTSPP

The meeting noted with appreciation the contributions of our Soviet colleagues with respect to the GTSPP implementation plan and welcomed their decision to join the project.

Besides increasing ocean temperature and salinity data flows, coming within the International Oceanographic Data Exchange; developing and implementation of coordinated approaches to data quality control and their formatting on the carrier; developing means of computerizing data management and processing; the following can also be done at VNIIGMI-WDC:

- Regular merging of GTS data with delayed data for the Atlantic and the Pacific and on this basis preparation of annual (real-time plus delayed) and historical data sets within Phases 1 and 2 of the Project. Then it is highly desirable that the data available at the oceanographic data centres of the European countries, Japan, specialized centres and international oceanographic organizations be utilized (e.g., ICES). The cut-off time for sending real-time data to MEDS is one week, delayed data - five to six months after the observation year. The time of completing Phase 2 "Historical Data" depends on the financial resources of the USSR NODC for data transfer from tables to technical media and the time of data submission to the USSR NODC.

- Preparation and publication of "Temperature and Salinity Monitoring Bulletin. North Atlantic and North Pacific" and annual reports on the availability and space-time distribution of temperature and salinity data available at VNIIGMI-WDC.

Preparation of the data base for the Arctic regions requires much effort in connection with estimating the data streams, finalizing the development of the data acquisition process and some organizational activities. The possibility of undertaking the above activities requires a more detailed consideration. What is already known is that most of the historical tables for the Arctic have the form of differently formatted tables and their transfer onto computer-compatible media will be rather money and labour consuming.

9.4 Schedule and Review Process for the Project Plan

The document is required by IOC by the beginning of October to be ready for IODE in January. Consequently, the following schedule and review process was established:

a) August 20 - Ben Searle to re-draft document based on results of Working Group meeting suggestions;

b) August 20 - Transmit draft to NODC, MEDS, TSDC in Brest, and ITPO via OMNET;

c) mid-September - NODC, MEDS, TSDC in Brest, and ITPO to add relevant diagrams and any additional text;

d) mid-October - Draft document will be circulated to all ad-hoc meeting participants for review and, in addition, a draft will be sent to Bert Thompson at WOCE Office for review by relevant WOCE Scientists;

e) End of October - Forward document to IOC for translation of executive summary and distribution of published document in time for GTSPP Workshop and IODE TC meeting in January 1990.

9.5 Project Management

The meeting discussed the on-going management of the Project, recognizing that as the number of participants increased, planning meetings could become unmanageable. It was agreed that a Steering Committee should be established to coordinate and oversee the Project. The meeting suggested that it was also necessary to undertake a scientific review of the project. As a result of these discussions, it was recommended that a Steering Committee be formed which would consist of 2 to 3 scientists from outside the project and 3 to 4 GTSPP participants.

10. Review of Historical Data Aspects of the Project

There are two aspects of this problem. The first has to do with the quality and availability of data which already are held within RNODC and WDCs in the IODE system. The second has to do with acquiring the data which have not yet been submitted to the IODE system.

For the data already held by the IODE, there is a need to regularize the quality control procedures. By this is meant that each centre doing quality assessment must be prepared to deliver a complete, written description of exactly what tests are done. Under the GTSPP this is being addressed by the writing of the Quality Control Manual. It could be considered to broaden the scope so that the Manual represents tests recommended by the IODE. As part of the quality control process, scientific input will be required along the lines of the JEDA model. If this can be incorporated into the quality control procedures that normally are done at data centres, so much the better. In any case, scientific input and comment to these procedures is necessary.

Once these procedures are documented and standardized, there will be the need to process all of the historical data through the new procedures. This will be a substantial task. It will be necessary to look at how the processing load can be shared amongst various data centres and to develop a plan within IODE to schedule this reprocessing task.

For the data not in the IODE centres, it is necessary to develop an acquisition strategy. First, there will be a substantial effort in locating where the data are held and what they are. This could be tackled through the science organizations within SCOR, through the CCCO and through commitments by member states. The collating of all of this information likely is the task of the WDCs unless some other centres volunteer. The reports of data will have to be sifted through to identify the data sets of greatest importance to the GTSPP according to the acquisition strategy. Then the data will have to be acquired and prepared for input to the archives. A scientific advisory group may be required here.

An initial review at NODC has found of the order of 300,000 oceanographic stations that are in manuscript form. Some tests have been done to develop automated means of scanning the data into ASCII files to computerize these data. The tests to date have been promising and are continuing.

It was also stated that real-time data can and should be used to assist in the identification of data sources for acquisition of more delayed-mode data for the project. Some concern was expressed on the quality of data due to changes in measurement technology over the last century.

All of this activity to acquire data not within the IODE system is a time consuming exercise. Before embarking on this there needs to be a careful plan that will have the support of the major agencies that would be involved. This plan should discuss how the work can be partitioned to the various interested parties so that all may be completed in a reasonable time frame, but spreads the burden as much as possible. The meeting noted the importance of this topic and recommended that it be discussed at the upcoming Workshop.

11. Review of the Action Plan

The action plan was reviewed in the light of the discussions and the decisions of the meeting, new participation by France and the USSR, workshop preparations, and was updated as appropriate.

12. Other Business

Steve Pazan gave a presentation which showed how an analysis of the data presently in the archives could be used to examine the information content. The results of such an analysis would be of value in assessing higher priority to data which have been collected but not yet reached the archives.

Australia will investigate the possibility of hosting the next GTSPP meeting in October, 1990.

The meeting noted the need for a plan for data and information management within the Project. It was suggested that this should be discussed at the next Meeting.

13. Approval of the Summary Report and Recommendations of the Meeting

The meeting report was read paragraph by paragraph and adopted as modified.

14. Closure of the Meeting

The Chairman expressed his thanks to all participants for the substantive progress made at the Meeting with special thanks to MEDS for hosting the meeting. Special thanks was also expressed to Mrs. Mira Clark for coordinating all the arrangements for the meeting. The meeting closed at 12:15 p.m., Friday, July 28.

ANNEX A

Global Temperature-Salinity Pilot Project
25-28 July, 1989
Room 1444

AGENDA

1. Approval of Agenda and Workplan

2. Presentations on the GTSPP and TOGA Sub-Surface Centre

3. Implementation of BUFR

4. Review of the Status of Implementation of Communications

5. Specification of QC Procedures

6. Review of GTS Monitoring

7. Review of the Scientific Requirements Document

8. The Role of the GTSPP in the WOCE Data Management System

9. Review of the Project Implementation Plan

10. Review of the Historical Data Aspects of the Project

11. Review of the Action Plan

12. Other Business

13. Approval of the Summary Report and Recommendations of the Meeting

14. Closure of the Meeting

ANNEX B

LIST OF PARTICIPANTS

Irving Perlroth
DOC/NOAA/NESDIS
National Oceanographic Data Center
1825 Connecticut Avenue, Room 422
Washington, D.C.
20235 USA TELEMAIL: NODC.WDCA
I.PERLROTH

Susan Kerig
DOC/NOAA/NESDIS
National Oceanographic Data Center
1825 Connecticut Avenue, Room 422
Washington, D.C.
20235 USA TELEMAIL: NODC.WDCA

Bob Fish
DOC/NOAA/NESDIS
National Oceanographic Data Center
1825 Connecticut Avenue, Room 422
Washington, D.C.
20235 USA TELEMAIL: NODC.WDCA

Greg Withee
DOC/NOAA/NESDIS
National Oceanographic Data Center
1825 Connecticut Avenue, Room 422
Washington, D.C.
20235 USA TELEMAIL: NODC.WDCA
G.WITHEE

Dr. Vyacheslav I. Melnikov
Senior Scientist
All-Union Research Institute of
Hydrometeorological Information
- World Data Centre
USSR State Committee for
Hydrometeorology TELEPHONE: 2-59-09 Obninsk
6, Korolyov Str. TELEPHONE: 546-39-50 Moscow
Obninsk, Kaluga Region FAX NO.: 2556684
249020, USSR TELEX: 412633 INFOR SU

Mr. Nikolai N. Mikhailov
Chief of Laboratory, Oceanographic Data Centre
All-Union Research Institute of
Hydrometeorological Information
- World Data Centre
USSR State Committee for Hydrometeorology
6, Korolyov Str. TELEPHONE: 2-56-78
Obninsk, Kaluga Region
249020, USSR

Mr. Ben Searle
Australian Oceanographic Data Centre
P.O. Box 1332
North Sydney
NSW 2059
Australia TELEMAIL: [BSEARLE/NESDIS]
TELEMAIL/USA

Mr. Garry Hopwood
Australian Oceanographic Data Centre
P.O. Box 1332
North Sydney
NSW 2059
Australia

Mr. Jean-Paul Rebert
Centre ORSTOM de Brest
IFREMER BP 70
29263 Plouzane
France TELEMAIL: ORSTOM.BREST

Mr. Joel Martellet
International TOGA Project Office
World Meteorological Organization
41, Avenue Giuseppe-Motta
P.O. Box 2300
CH-1211 Geneva 2
Switzerland TELEMAIL: INTL.TOGA

Mr. Doug McLain
Oceanographer
National Ocean Service
Ocean Applications Group
NPS, FNOC, Building #4
Monterey, California
93943-5005 U.S.A. TELEMAIL: OAG.MONTEREY

Mr. Steve Pazan
Scripps Institution of Oceanography
University of California at San Diego
La Jolla, California
92093 U.S.A. TELEMAIL: S.PAZAN

Mr. Roger Bauer
Compass Systems Inc.
4640 Jewel Street
Suite 204 TELEPHONE: 619-270-5230
San Diego, California
92109 U.S.A.

Mr. Allyn Clarke
Bedford Institute of Oceanography
P.O. Box 1006
Dartmouth, Nova Scotia
B2Y 4A2 TELEMAIL: BEDFORD.INST

Dr. J.R. Wilson
Director
Marine Environmental Data Service
Physical & Chemical Science
12082-200 Kent Street TELEPHONE: 613-990-0264
Ottawa, Ontario FAX NO.: 613-993-4658
K1A 0E6 Canada INTERNET: wilson@meds-sdmm.dfo-mpo.gc.ca

Mr. J. Gagnon
A/Chief
Data Management & User Services Division
Marine Environmental Data Service
Physical & Chemical Science
1202-200 Kent Street TELEPHONE: 613-990-0260
Ottawa, Ontario FAX NO.: 613-993-4658
K1A 0E6 Canada INTERNET:gagnon@meds-sdmm.dfo-mpo.gc.ca

Mr. J.R. Keeley
Chief
Ocean Information & Systems Division
Marine Environmental Data Service
Physical & Chemical Science
1202-200 Kent Street TELEPHONE: 613-990-0246
Ottawa, Ontario FAX NO.: 613-993-4658
K1A 0E6 Canada INTERNET: keeley@meds-sdmm.dfo-mpo.gc.ca

Mr. Bert Thompson
WOCE International Project Office
Institute of Oceanographic Sciences
Deacon Laboratory
Wormley, Godalming
Surrey GU8 5UB
United Kingdom TELEMAIL: WOCE.IPO
B.THOMPSON

ANNEX C

GTSPP Project Goals

1. To create a timely and complete data base of ocean temperature and salinity data of known quality in support of the World Climate Research Program and of national requirements.

2. To improve the performance of the IOC/IODE and WMO/IOC IGOSS data exchange systems by actively seeking data, exercising the data inventory, data management, and data exchange mechanisms as they are intended to work and recommending changes where necessary to meet national and international requirements.

3. To demonstrate through a widely distributed monitoring report produced on a regular basis, information on the performance of the IODE and IGOSS systems.

4. To improve the state of historical data bases of ocean temperature and salinity data by developing and applying improved quality control systems to these data bases.

5. To improve the completeness of these historical data bases by digitization of historical data presently in analog or manuscript form and by including digital data not presently at a WDC.

6. To distribute copies of portions of the data base and selected analyses to interested users and researchers.

ANNEX D

LIST OF DOCUMENTS

1. Report of an ad-hoc consultative meeting on the Global Temperature-Salinity Pilot Project, 23-25 January 1989, U.S. National Oceanographic Data Center, Washington, D.C. (February 1989).

2. Subsurface data management and tracking for TOGA, 31 May - 1 June 1989; J.P. Rebert.

3. ITPO Comment on the report of an ad-hoc consultative meeting on the Global Temperature- Salinity Pilot Project (IOC/IODE - Cons. VL/Inf. 9).

4. On creating a global deep-sea data set at the All-Union Research Institute of Hydrometeorological Information - World Data Centre (VNIIGMI WDC) and participation of VNIIGMI-WDC in the Global Temperature-Salinity Pilot Project (GTSPP); L. Brenerman, V. Melnikov, N. Mikhailov.

5. BUFR - A Binary Representation; W. Thorpe.

6. Composite BUFR Quality Control Fields; Roger A. Bauer.

7. CSPAN Network Considerations.

8. IGOSS Monitoring Results, July 1989.

9. Monitoring of IGOSS Data Flows.

10. Review of Ocean Science Data Requirements for Temperature and Salinity (Draft), 18 July 1989; Steven L. Pattersen.

11. GTSPP Action Plan.

12. Comments on the draft Q.C. manual, Chris Noe.

13. Quality Control Manual (19 June DRAFT), J.R. Keeley.

14. Appendix 1 of Q.C. Manual (18 July), J.R. Keeley.

15. Draft Implementation Plan (July 1989), B. Searle.

16. WOCE Report of 1st meeting of the Data Management Committee, May 1989.

17. Marine Data Platforms - An Interactive Inventory , Seneira, Woodward and Noe).

18. Proposed new data flows for the GTSPP (FROM ITPO).

19. IOC Circular Letter No. 1227.

20. JEDA distribution of SD2 and BT figures by year.

21. MEDS Data Management System; Compengserv, Ltd. MEDMAN: JULY 1989.

22. WOCE Goals and major observational components.

23. Intelligent Scanning System, Kurzwell 5000.

ANNEX E

Global Temperature-Salinity Pilot Project Workshop
January 15-16, 1990
United Nations, New York

Provisional Agenda

1. Introduction and Welcome

2. GTSPP Overview - Objective and Goals

3. GTSPP Interfaces with Existing Programs
- IGOSS/IODE
- TOGA
- WOCE

4. Scientific Participation

5. Project Elements
- real-time
- delayed-mode
- historical

6. Data and Information Management Strategy

7. Proposed Project Management

8. Proposed Implementation
- National Data Centres, Designated National Agencies, Institutes

9. Discussion by Member States

ANNEX F

TABLE OF CONTENTS FOR THE PROJECT PROPOSAL

1. Brief Executive Summary

2. Overview

2.1 Introduction - providing the rationale of the project and taking relevant sections from the draft Implementation Plan. It would address the scientific needs and requirements and include details of previous meetings and a list of participating countries.

2.2 Objectives and Goals - will include the existing objectives as determined at the first ad-hoc meeting and would describe the benefits of the project to participants and Member States.

This section would attempt to encourage participation but would qualify the areas of the project where involvement will be beneficial.

2.3 GTSPP Interface with Existing Programs - this section will indicate strongly that GTSPP is attempting to build on existing programs such as the TOGA data model and is intended to complement and not replace existing mechanisms.

A diagram of IODE/IGOSS data flow would be included together with an examination of GTSPP's relationship with both TOGA and WOCE and their data centre functions.

2.4 Science Support - the science component of GTSPP would be described in general terms.

2.5 Project Elements - this would examine the general levels of involvement and participation within the project such as global, regional and national aspects. It would examine the near-real time and non real-time data flow. The individual elements, and a brief description of each, would be taken from the draft Implementation Plan.

2.6 Project Priorities - this section would use the relevant section from the draft Implementation Plan and would include a simplified graphical version of the time schedule given as Table 2 in the report of the first Ad-Hoc Meeting.

3. Implementation Scheme

3.1 This would expand on the Action Plan drafted by MEDS but would not include individuals names or the specific time schedule.

3.2 A generalized data flow based on the Figures 1 and 2 from the paper titled "Proposed New Data Flows" would be included.

4. Project Management

4.1 The parent bodies would be described briefly.

4.2 The need for a project Steering Committee will be examined. A committee incorporating 2-3 people from science and 3-4 people from participating agencies would monitor and assist in the development of GTSPP. The Steering Committee would also undertake scientific reviews of the project.

Annex 1 - List of Project Planning Contributors

This will be a list of all people who have attended the initial GTSPP ad-hoc meetings or have contributed to the planning in some way.