GTSPP Project Plan

SUMMARY

1. INTRODUCTION

1.1 Purpose of the Document
1.2 Rationale for the Project
1.3 Scientific and Other User Requirements
1.4 Background of the GTSPP Project

2. OBJECTIVES OF THE GTSPP

3. BENEFITS OF GTSPP FOR MEMBER STATES

4. MEMBER STATE PARTICIPATION IN THE GTSPP

5. GTSPP INTERFACE WITH EXISTING PROGRAMMES

5.1 Interface with the TOGA/WOCE Subsurface Data Centre

6. SCIENCE SUPPORT

7. PROJECT ELEMENTS

7.1 Near Real-time Data Acquisition
7.2 Delayed Mode Data Acquisition
7.3 Communications Infrastructure
7.4 Quality Control Procedures
7.5 Continuously Managed Database
7.6 GTSPP Data and Information Products
7.7 Data Flow Monitoring
7.8 Historical Data and Long Term Time Series

8. PROJECT PRIORITIES

9. PROJECT MANAGEMENT

9.1 Parent Bodies
9.2 Steering Group

APPENDIX

Terms of Reference for the Steering Group on the GTSPP
Composition of the Steering Group for the First Intersessional Period

SUMMARY

The science of oceanography is at the beginning of a new era. Researchers are addressing vitally important problems that demand the support of cooperative global data collection and exchange programmes. As a result, it is now necessary to develop the existing oceanographic data exchange mechanisms to provide the necessary timely and complete exchange taking full advantage of technological advances, such as satellite data transmission systems, CD-ROM's and artificial intelligence based information management systems.

The Global Temperature-Salinity Pilot Project (GTSPP) is intended as a trial system, that will provide the model for a global marine science data exchange system to serve the needs of present and future international science programmes and of the programmes such as the WOCE Goal 2 which will provide for the application of WOCE science on a global basis. Its immediate task will be to create a complete data and information base of ocean temperature and salinity data captured both in real-time and submitted in fully processed form weeks to months later. GTSPP data is intended to support the World Climate Research Programme (WCRP), its associated programmes and all types of national requirements from fisheries operations to fundamental research.

This Project will be drawing on the considerable experience of existing data exchange programmes, particularly the Integrated Global Ocean Services System (IGOSS) and the International Oceanographic Data and Information Exchange (IODE). It will also introduce the important science element into the data flow cycle. GTSPP will be implemented in the manner of the Tropical Ocean/Global Atmosphere (TOGA) data flow model which has been very successful in increasing data flow and improving the management and quality of oceanographic data from tropical oceans. GTSPP will build on the Joint Environmental Data Analysis (JEDA) principle by introducing scientific analysis into the quality control of data and incorporating this technique into the data management function.

It is hoped that the creation of GTSPP will encourage all Member States to contribute to the development of the global data exchange programmes. The Project has been structured to allow the participation of Member States at all levels, from the technologically complex computer based management functions through to the task of acquiring historical data sets on a national scale. All Member States will benefit as the more complete and higher quality global data set evolves. In addition, the new technology and mechanisms used to operate GTSPP will gradually flow to the Member States.

GTSPP is a project within IODE and IGOSS and will build on these existing systems. The Project will be upwardly compatible, evolving over time while retaining a structure that will enable participation at all levels.

1. INTRODUCTION

1.1 Purpose of the Document

The purpose of this document is to provide an overview of the Global Temperature/Salinity Pilot Project (GTSPP) and the actions required to implement and operate it. The benefits for Member States and how they can participate in the project will be described. The interaction with existing data management programmes will be shown. The project's elements, priorities and critical implementation problems will be discussed. This document is aimed at the scientific research community, in general, and the maritime research community, in particular, as well as the IGOSS and IODE data managers.

GTSPP is a major international undertaking and will require extensive cooperation between all Member States in order to be successful. This document will identify the elements necessary to implement the Project allowing the Member States to evaluate their participation and plan for the necessary computing facilities, communications mechanisms, and data acquisition and quality control procedures. The logistics of the Project are large and it is intended that this Document will provide the information necessary to avoid areas of duplication with existing systems. It will also provide a framework on which development activities and operations can be based.

1.2 Rationale for the Project

For a number of years it has been clear that there is an urgent need to improve national and international oceanographic data management procedures. Both the scientific community and the data management agencies have raised valid concerns with the existing situation. Since the most sophisticated user of oceanographic data is the research community, data managers have realized that it is necessary to involve researchers in the quality control process to assure the result.

The United States National Oceanographic Data Centre (NODC), in an effort to promote and improve data management activities in the USA, established a cooperative venture with Scripps Institute of Oceanography. This project is called the Joint Environmental Data Analysis (JEDA) Centre and has already provided the utility of a scientific node in data quality control. The benefits of JEDA are two fold:
(i) JEDA is steadily improving the quality and the usefulness of the data held nationally by NODC and distributed to other NODCs and RNODCs, and
(ii) the activities of JEDA are resulting in the production of a number of very useful products.

GTSPP will implement this method of quality control. It is anticipated that the benefits of JEDA will develop on a global scale and result in a higher quality global data set.

The GTSPP objectives and goals are described in more detail in Section 1.2. In addition, it is intended that the GTSPP will develop the prototype for data management systems for global management of all important ocean parameters. It will also introduce international standards for quality control and will assist the marine research and data management communities to further their cooperation.

The TOGA Subsurface Data Centre has demonstrated the feasibility of the proposed GTSPP proactive approach to the building of the global temperature-salinity database. The transmission of data to the TOGA Centre has increased from less than 10,000 per year in 1985 to over 17,000 per year in 1987. Also, the time taken for the higher resolution delayed mode data to replace the near real-time data is steadily being reduced.

1.3 Scientific and Other User Requirements

The scientific community, in general, and the World Climate Research Programme (WCRP), in particular, require high quality, timely, and complete global ocean temperature and salinity data sets. It is the intention of GTSPP to provide these data sets.

Oceanographic data provides a direct measure of the changing environment. It is required to calibrate and evaluate descriptive and predictive ocean models. Comprehensive and accurate observations of the worlds oceans and their circulation are fundamental to the study of climate change. The distribution of ocean temperature and salinity are of particular importance in providing the necessary four dimensional description of circulation and heat flux and its variability. These are primary objectives of TOGA and WOCE.

Temperature and salinity data are also used in a wide range of research and non-research areas, including recreational and commercial activities such as fisheries. In conjunction with the scientific requirement mentioned above, the need for a timely, high quality ocean temperature and salinity database is clear.

To meet users needs, the data in the GTSPP databases must be documented to agreed scientifically justifiable standards. The scientific community must therefore be fully involved in the project, particularly in the quality assurance aspects. The GTSPP systems thus will be capable of identifying and providing data selected by level of accuracy and reliability to meet scientific, engineering, and operational requirements.

When data are provided to a scientist or other user from the GTSPP database, it will have been passed through a number of quality control checks. These checks are documented in the GTSPP Quality Control Manual. The user will know how reliable his GTSPP data are. This will reduce the need for users to undertake their own quality checks. GTSPP will retain data of all levels of reliability. For example, if a better version of the observation arrives, such as a fully processed version of a CTD observation by the scientist who collected it, this would replace the lower quality IGOSS TESAC message.

1.4 Background of the GTSPP Project

The original concept for the Global Temperature/Salinity Pilot Project (GTSPP) was raised at the Second Joint IOC-WMO Meeting of Experts on IGOSS-IODE Data Flow. This meeting was held at the Marine Environmental Data Service (MEDS) of the Department of Fisheries and Oceans in Ottawa, Canada (18-22 January, 1988) and was attended by a number of experts from France, Canada, USA, Federal Republic of Germany, Japan and Australia.

Under the Meeting's Agenda Item 5, "Identification of Ways to Overcome Problems of Non-Real-time Submission of BATHY and TESAC Messages through IODE Channels", discussions were held regarding the need for "unique, timely, global databases for a variety of oceanographic parameters". The US NODC described their experiences with the Joint Environmental Data Analysis (JEDA) Centre. The meeting agreed that a similar principle could and should be used on a global scale in order to advance international data exchange programmes.

The meeting developed a draft proposal for a project to develop a very high quality, scientifically acceptable global ocean thermal data set.

At a meeting of The Group of Experts on RNODC's and Climate Data Services (Wormley, UK, February, 1988), participants from IODE, WOCE and TOGA discussed the WCRP ocean data needs and the role of the IODE system in meeting those needs, and gave substantial credence to the GTSPP concept. A recommendation "...that a pilot project (global thermal database) be started to prove the value of the concept..." was supported by the participants of the meeting.

At the Fifth Session of the Joint IOC-WMO Working Committee for IGOSS (Paris, 14-23 November, 1988), a plan proposing a pilot project on global temperature and salinity data management was presented to the Committee by Canada and the USA. This plan differed slightly from that developed in the January, 1988 Meeting of Experts in Ottawa with the inclusion of salinity data as a requirement for the global data set. The addition of salinity as a parameter resulted from a number of meetings and discussions held in the USA and Canada between the scientific user community and the data management agencies. The meetings included the NODC/ERL Workshop on Ocean Data Files (13-14 June, 1988) and various other informal discussions.

The IGOSS Working Committee expressed great interest and enthusiasm for the proposal and accordingly adopted Recommendation 4 (JWC-IGOSS V).

The First Ad-Hoc Consultative Meeting on the GTSPP was held at NODC in Washington, DC (23-25 January, 1989). This meeting was the first opportunity to discuss the project in detail and to examine some of the technical areas of concern. A number of potential problem areas were identified, the most significant relating to the quality control of data. Considerable discussions resulted from this topic and many aspects were left unresolved. However, most of the identified difficulties have now been addressed and the solutions are included in the GTSPP Quality Control Manual.

The Meeting developed a set of goals and defined the Project's major elements. The Summary Report of this meeting, titled "Report of an Ad-Hoc Consultative Meeting on the Global Temperature/Salinity Pilot Project (An IGOSS-IODE Programme)", has been provided to the IOC and WMO Secretariats.

The Second Ad-Hoc Consultative Meeting on the GTSPP was held at the Marine Environmental Data Service in Ottawa, Canada (25-28 July, 1989), and further advanced the Project. This meeting also enabled participants of existing programmes, including TOGA and WOCE, to describe their experiences and thoughts on oceanographic data management. The relationship between these programmes and GTSPP was discussed at great length and it was stressed that GTSPP would complement the existing data flow mechanisms.

A Summary Report, titled "Second ad hoc Consultative Meeting on the Global Temperature/Salinity Pilot Project (A Proposed IGOSS-IODE Project)", has been provided to the IOC and WMO Secretariats.

A Workshop on the GTSPP was held prior to IODE XIII in New York (15-16 January, 1990). Many aspects of the project were discussed and clarified. The formation, terms of reference, and composition of the Steering Group on the GTSPP were also defined. The recommendations of the Workshop were discussed at IODE XIII and the results of the discussions can be found in the report of that meeting.

2. OBJECTIVES OF THE GTSPP
(i) To create a timely and complete data and information base of ocean temperature and salinity data of known quality in support of the World Climate Research Programme (WCRP) and of national requirements.
(ii) To improve the performance of the Intergovernmental Oceanographic Commission (IOC)/ IODE and World Meteorological Organization (WMO)/IOC IGOSS data exchange systems by actively pursuing data sources, 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.
(iii) To disseminate, through a widely distributed monitoring report produced on a regular basis, information on the performance of the IODE and IGOSS systems.
(iv) To improve the state of historical databases of oceanographic temperature and salinity data by developing and applying improved quality control systems to these databases.
(v) To improve the completeness of these historical databases by the digitization of historical data presently in analogue or manuscript form and by including digital data not presently at a World Data Centre (WDC).
(vi) To distribute copies of portions of the database and selected analyses to interested users and researchers.

3. BENEFITS OF GTSPP FOR MEMBER STATES

GTSPP has an obvious international objective, but it is also intended that Member States will benefit at the national and regional level. Figure 1 is a graphic depiction of the benefits of GTSPP to member states and of the manner in which the benefits will flow into the member states through their IODE and IGOSS centres. In particular, Member States will benefit from GTSPP because of:
(i) increased quantities of temperature and salinity data becoming available,
(ii) improved timeliness of the data and in distributing them to the User Community,
(iii) the global standardization of the quality control procedures and the production of regional and global data products,
(iv) standardization of data formats and processing methodologies, and
(v) the introduction of new state of the art methods and technology in communications, artificial intelligence, quality control, and data management and its progressive transfer to Member States.


Increasing the rate at which data are available to the user in real-time will benefit national programmes using operational oceanographic data. As greater quantities and more timely data become available for near real-time analysis, the reliability and accuracy of operational products for national users will increase and the efficiency and safety of operations will improve.

The standardization of data formats will simplify data exchange nationally as well as internationally, and will further the development of standard software packages for quality control, data management and data product development. This software will form part of the technology transfer that is an important by-product of GTSPP.

4. MEMBER STATE PARTICIPATION IN THE GTSPP

GTSPP has come about as the result of a meeting of experts on data flow and has now expanded to involve a number of countries. It is hoped that other Member States will participate actively, providing assistance in running GTSPP and using their experience and expertise to advance the capabilities of the Project. There are a wide range of activities involved in establishing and operating this Project. The best progress will be realized if contributions come from many Member States so that a heavy workload is not imposed on a small number of active participants. If the workload is spread:
(i) the work tasks can be completed in a timely manner,
(ii) a large number of agencies gain a sense of involvement and are actively contributing to GTSPP, and
(iii) the workload of individual agencies is reduced to a level where it can be undertaken in a complete and successful manner without compromising national activities.

There are many areas where Member States can make valuable contributions. Within their own marine science communities, Member States can:
(i) encourage an increase in the number of temperature observations transmitted in near real-time,
(ii) encourage an increase in the number of salinity observations transmitted in near real-time,
(iii) undertake quality control checks according to the GTSPP Quality Control Manual on data collected by national programmes,
(iv) improve mechanisms allowing the more timely submission of both near real-time and delayed mode data,
(v) encourage national research agencies to develop data and information products as part of the scientific quality assurance process,
(vi) actively acquire historical temperature and salinity data that has not previously been exchanged, and
(vii) provide advice and assistance to the project in the areas of data management, quality control, communications, and product development.

One particularly important area that needs improvement is the submission of near real-time salinity observations. Globally there are very few salinity data transmitted over the GTS. Member States should make every effort to encourage their research vessels to submit TESAC messages.

All Member States will automatically play a part in GTSPP through their normal IGOSS/IODE participation. To participate more actively in the GTSPP, Member States can write to the IOC Secretariat, the WMO, or to the Chairman of the GTSPP Steering Committee outlining the areas in which they wish to contribute.

5. GTSPP INTERFACE WITH EXISTING PROGRAMMES

The GTSPP goal is to work in cooperation with any existing programmes collecting and managing temperature and salinity data. GTSPP will complement rather than duplicate existing data flow systems. In particular, GTSPP is working with the TOGA/WOCE Subsurface Data Centre in Brest as described below. There is no intention to duplicate the existing mechanisms and great care will be taken to ensure that the Principal Investigator is not confused about where to submit his data. GTSPP will liaise closely with all existing programmes to ensure that these potential problems do not occur. It is important to stress that GTSPP is not intending to replace the existing systems, but rather work within systems, such as IGOSS and IODE, and attempt to complement their capabilities. GTSPP is a pilot project and it is intended that through its achievements oceanographic data exchange and management activities can be strengthened.

5.1 Interface with the TOGA/WOCE Subsurface Data Centre

TOGA/WOCE Principal Investigators will continue to be encouraged to submit their delayed mode data sets according to existing arrangements to the TOGA/WOCE Subsurface Data Centre within six months of acquisition. If the data are not already in the GTSPP, the TOGA/WOCE Centre will provide a copy to the GTSPP. The Subsurface Data Centre will provide services to its TOGA/WOCE clients from the database maintained in Brest.

WOCE has indicated that GTSPP data would be of considerable use in the core project for global descriptions on the evolution of the heat and salt distributions in the upper oceans. GTSPP would also be of value to WOCE by identifying and then acquiring temperature and salinity data sets that are not in the existing system. These data sets, such as time series transects, would supplement the present WOCE data objectives.

At the same time it was considered very important that GTSPP compliment and not disturb any existing working data flows. Consequently, the GTSPP data flows have been designed specifically not to interfere with the existing TOGA/WOCE data flows to the Centre in France.

6. SCIENCE SUPPORT

Scientific acceptance and involvement is crucial to the success of GTSPP. Without the support of the scientific community and the major international research programmes, GTSPP will not reach its goals. A major effort will be to actively seek support from the scientific community. This will occur through a number of mechanisms including:
i) general publicity,
ii) inviting representatives of major field and scientific programmes to participate in GTSPP Meetings,
iii) sending GTSPP representatives to meetings of major programmes to assist in the data management >planning aspects and actively promote GTSPP, and
iv) requesting members of IOC Missions to promote GTSPP during visits to Member States.

Scientific support is also important to the project since a major factor in developing a high quality database involves the scientific analysis and evaluation of the GTSPP data sets. It is essential that GTSPP is not seen as simply a data management exercise undertaken by data centres. The Project must be developed as a combination of data management and scientific endeavor which will result in a timely and complete database of temperature and salinity data of known and agreed standards of quality.

7. PROJECT ELEMENTS

Figure 2 is an attempt to present the rather complex data flows of national and international programmes within which GTSPP must find its place. The boxes in the Figure represent generic centres. A given international IGOSS or IODE centre may fit within several boxes in carrying out its national and international responsibilities. The following sections discuss this Figure in terms of eight essential elements of the GTSPP.

These eight essential elements are given below. All of these must be functioning before the Project can be considered fully operational.

The elements of GTSPP are:

(i) Near Real-time Data Acquisition
(ii) Delayed Mode Data Acquisition
(iii) Communications Infrastructure
(iv) Quality Control Procedures
(v) Continuously Managed Database
(vi) GTSPP Data and Information Products
(vii) Data Flow Monitoring
(viii) Historical Data Acquisition and Processing.

7.1 Near Real-time Data Acquisition

Near real-time data acquisition within GTSPP will depend primarily on the Global Telecommunications System (GTS) of the World Weather Watch of WMO and the telecommunications arrangements for BATHY and TESAC data established by IGOSS. In addition, there will be a number of enhancements. These enhancements will consist of acquiring copies of other real-time data sets from any other available sources via a high speed network or other means to ensure that the most complete real-time data set is captured.

In Figure 3 the box labelled "Real-time Temperature and Salinity Data" represents the data flow for the acquisition of real-time data. The data flows through national programmes and centres, the GTS insertion centres, and IGOSS SOCs to the GTS. The "Real-time Assembly and QC Centre" in MEDS will acquire data from its GTS node on a daily basis and from those RNODCs-IGOSS that agree to do so on a weekly or monthly basis. The participation of the RNODCs-IGOSS in this manner is the key to effective real-time data flow monitoring and the creation of the most complete global database. The "Real-time Assembly and QC Centre" will compile the global data set from the various sources, perform additional QC if necessary to bring the quality to the level of the GTSPP QC Manual, and forward the data to the US NODC, where it is added to the continuously managed database. Section 7.6 describes the characteristics of the continuously managed database. Note that Figure 3 does not imply that all data flows from top to bottom through all boxes to MEDS.

In addition to the IGOSS programme there are a number of existing or planned near real-time data acquisition programmes, including the TOGA Ship of Opportunity Network and the WOCE Hydrographic Programme, that will deliver data to GTSPP. A major effort will be to encourage the submission of salinity data in the form of TESAC messages in near real-time.

The weekly data set will be transmitted to the US NODC via SPAN where it will be added to the GTSPP database.

NODC will accumulate a monthly set of near real-time data, which will be merged with any delayed mode data received for the relevant period. Each monthly data set will be forwarded to the agency responsible for the scientific evaluation and quality assurance of the data. The scientific agency will undertake an analysis of the data covering their region of interest. These data would be retained by the agency for a period of time, possibly up to one year. The complete, scientifically quality controlled data set would be returned to NODC and included in the GTSPP Database.

The aspects of the real-time data flow through NODC to the Scientific QC centres are included in the box marked "Continuously Managed Database" in Figure 4. The flow on the right hand side of the box shows the data proceeding through the "duplicates identification/replacement management" processors into the continuously managed database. At the end of the month the data are retrieved through the "position/time" filters and sent to the appropriate QC centre. When the data (or flags) are sent back to NODC, they feed in at the top again through the "duplicates identification/replacement management" processors and the database is "updated/corrected" as required.

Note that the regular route for real-time data to the box marked "Operational Clients" in Figures 2 and 3 is not affected by GTSPP. This route provides for uninterrupted flow of data for weather and operational forecasting through the national weather services of member states.

7.2 Delayed Mode Data Acquisition

GTSPP will utilize, to the extent possible, the existing IODE data network and processing system to acquire and process delayed mode data. The box entitled "Delayed Mode & Historical Data" in Figure 5 shows delayed mode data flow in graphic form. The data flow into the continuously managed database in the US NODC through a "Delayed Mode QC" process. This process is analogous to the QC carried out on the real-time data and is to conform to the specifications of the GTSPP QC Manual. In some cases, where appropriate arrangements can be made, this box and the QC it implies will actually exist and be performed in another national oceanographic data centre on behalf of NODC.

Having proceeded through the delayed mode QC process, the data then follow the same route as the real-time data through the rest of the CMD process, however, on a different time schedule because of the more irregular times of arrival. During the merging of the data into the CMD, any duplicates occurring between near real-time and delayed mode data sources will be identified with the highest quality duplicate being retained as the active CMD copy.

Many of the data collected in recent years reach IODE with the delays in submission resulting from the time taken to process, quality control, and work up the data. GTSPP will attempt to encourage Principal Investigators to submit these data more rapidly within the constraint of having first right to publication of their data (where applicable), and will encourage agencies who are not already doing so to submit their data.

Delayed mode data will usually be exchanged in GF3 on magnetic tape. However, in some instances where high speed networks are available, data will be telecommunicated. Member States without the ability to exchange data in GF3 are encouraged to submit data in any form they can. GTSPP will attempt to provide assistance to overcome any data exchange difficulties. The WMO format BUFR will continue to be evaluated for use within GTSPP.

Due to the temporal and spatial variability of historical data, it will be difficult to produce compatible and consistent products. However, the results of any scientific analysis will still be of interest for many applications and will be distributed by GTSPP where appropriate.

7.3 Communications Infrastructure

The ability to acquire data, and to ensure that it is provided to the GTSPP Continuously Managed Database system in a timely manner, will be dependent on the communications infrastructure. The initial operations of GTSPP will involve a range of communication mechanisms which will include:
(i) satellite ship to shore (international),
(ii) HF radio ship to shore,
(iii) high speed network, data centre to data centre,
(iv) magnetic tape, data centre to data centre or National Meteorological Centre (NMC) to data centre,
(v) GTS NMC to NMC.
Over a period of time, the objective is to reduce the range of communications methods used and to focus on the more recent and more effective developments, such as satellite systems and digital data networks such as the NASA Space Physics Analysis Network (SPAN).

Communication mechanisms will also be important in the dissemination of data, information products and information concerning the project. Some of the products have near real-time uses, the applicability of which can be greatly enhanced by their timely distribution. The use of electronic mail systems will be one method explored for the dissemination of such GTSPP information and products.

Figure 2 does not contain any specific information on communications infrastructure. However, the characteristics of the communications infrastructure are implied to a certain extent by the arrows on the diagram.

7.4 Quality Control Procedures

The quality control element is crucial to the successful implementation of GTSPP. It is essential that all data maintained in the GTSPP Continuously Managed Database (CMD) is of a known quality and that the scientific community, including the major global research programmes, accept the validity of the quality control flags attached to all GTSPP data.

The philosophy of GTSPP is to create a timely and complete database, and this will be achieved by retaining all data that meet the minimum level of quality. Once an observation has successfully passed the first level of checks (which will consist of relatively simple data structure, format and range related tests), it will not be removed or deleted from the CMD except in the case of exact duplicates.

The GTSPP Quality Control Manual covers the checks that will be undertaken by participating data centres. It describes the approach of using a small number of quality control flags by using a series of data quality levels to identify data validity.

The quality control levels will range from a minimum series of checks up to the more sophisticated numerical assessments. All the data that have been checked and have passed the lowest quality level criteria will be included in the CMD with the appropriate flag and it will then be available to the user.

A significant feature of the GTSPP quality control procedures is that they will form the basis for a global quality control standard. If the criteria used in each level of the data evaluation can be accepted by the scientific community, the data management role of this project will be simplified. It is very important that the scientific community has confidence in the methods used to check the data. If this confidence can be gained it will be a major achievement for GTSPP. To this end, the scientific community has been involved, and will continue to be involved, in reviewing and agreeing to GTSPP quality control procedures. Scientific quality control will be undertaken by national research institutes that have considerable expertise with the water masses in their area of interest. Scientific quality control is accomplished by such means as mapping the data and various calculated variables or using the data in some other analysis that is sensitive to neighboring data. This is the most rigorous test of an observation and it will identify more subtle problems than will the more general techniques described in the QC Manual. Such scientific assessment will significantly increase the acceptability and validity of the GTSPP data set.

7.5 Continuously Managed Database

All the activities of GTSPP, including data acquisition, communications, data quality control, scientific analysis and analysis product development result in the creation of the global temperature and salinity data set. To manage this data set, a Continuously Managed Database (CMD) is being implemented.

As data are acquired in both near real-time and delayed mode, they will be added to the database. Delayed mode data calibrated and quality controlled by the originator replaces the data obtained in near real-time. That is, data that have been quality controlled to a higher level will replace the lower level data. The Continuously Managed Database will therefore hold the most current and highest quality data set. It will be continuously refined as additional quality checks are undertaken. The term replace here means replace as the active copy of the observation in the database. As stated earlier, observations that have passed quality control and entered the database are not removed. They are flagged to indicate that a higher quality version of the observation exists in the database.

The system supporting the CMD System will provide a number of functions including:
(i) Overall management and archiving of GTSPP data.
(ii) Monitoring of data acquisition and data flow.
(iii) Development and production of data management statistics and information products.
(iv) Production of subsets of data in response to requests.
(v) Management of updates of databases on a regular and frequent basis.
(vi) Detection of duplicate data, including exact and inexact duplicates and duplicates between operational and delayed mode observations.
(vii) Management of meta data which describe the actual data set.
Subsets of GTSPP data covering areas of national interest or global data sets will be provided to Member States on request.

7.6 GTSPP Data and Information Products

In developing the GTSPP data set, a range of by-products will be produced. These products result primarily from the quality assurance procedures undertaken by both data management agencies and research organizations.

At the present time, various research and operational agencies produce data analysis products in response to national demands. With careful planning and coordination, these analyses can be produced using GTSPP data, providing a better quality product from more and higher quality data. By dividing ocean areas into regions of responsibility, and by developing mutually agreed guidelines on the format and content of the products, GTSPP can obtain a global coverage. By having the scientific experts work in ocean areas with which they are familiar, the necessary local knowledge find its way into the products.

It is intended that GTSPP products will be issued on both a monthly and annual basis, and may include the following types:
(i) spatial distribution of data used in each analysis field,
(ii) sea surface temperature,
(iii) sea surface salinity,
(iv) mixed layer depth,
(v) vertically integrated upper ocean heat content,
(vi) dynamic height,
(vii) 300 meter temperature, and
(viii) anomaly values of the above data types.

The development and issue of data and information products fulfills three main objectives. These are:
(i) To provide a detailed inventory of data acquired by GTSPP giving valuable information to the user community and assisting them in defining and specifying their needs for data.
(ii) To provide a useful product for researchers who will use the available analyses to assist and further their own studies.
(iii) To promote and advertise the GTSPP by demonstrating both the uses that the data can be put to and showing the effectiveness of GTSPP as a data management model, and as an important contributor to global research programmes.

Data and information products will be disseminated as widely as possible and via a number of means including mail, electronic mail and bulletin boards.

7.7 Data Flow Monitoring

The implementation and operation of GTSPP will involve the use of considerable resources, including manpower, computer systems, and communications networks. In order to maximize the effectiveness and the data returns for the investment, it is essential that the data flow of the system is carefully monitored.

The monitoring programme will assist all participants, including data contributors and user groups, to have knowledge of the effectiveness of the project and to modify aspects that are not performing efficiently. It will also provide an indication of improvements in the system as the project develops.

In particular, GTSPP will exercise the monitoring procedures developed in IGOSS and IODE and assist with the further development and improvement of these procedures.

GTSPP data flow monitoring will aim to meet the following objectives:
(i) provide information on the data flow in GTSPP for both real-time and delayed mode data,
(ii) identify problem areas and bottlenecks in data flow and information distribution,
(iii) provide statistics that can be used to promote GTSPP, and
(iv) provide feedback to the data contributors, data centres and data analysis centres.

There are a number of aspects that can be measured that will give an insight into the effectiveness of the project. The data flow mechanisms can be monitored regarding timeliness of data exchange from collection to inclusion in the database for both real-time and delayed mode data. Details relating to how much data reaches the database can be produced. The time delays in acquiring, quality controlling and adding delayed mode data to the database will also give an indication of the effectiveness of the project.

7.8 Historical Data and Long Term Time Series

An important component of non real-time data acquisition relates to the potentially large set of both original and processed data that has not been exchanged within the oceanographic community. GTSPP will work with Member States to actively pursue these historical data sets and submit them via IODE channels. The acquisition of non-digital data of this type will form a second phase of GTSPP activity once the digital, delayed mode data have been processed and archived.

The objective of this project element is to make available a global, historical data set for climate research purposes. This global data set will be built with the help of the World Data Centres, IGOSS and IODE Centres, and other institutions. The resulting database will then be available to the scientific community.

Candidate data sets for this database are located in existing IODE archives, in RNODC's, regional data centres, in World Data Centres and in various institutional and international programme data centres located world-wide. Data will be identified as to their physical location, type, recording media, and priority for entry into the historical database. This work will be coordinated by the Steering Group through SCOR, the CCCO, and commitments from individual countries.

Once identified, data will be forwarded to one of several assembly centres which will perform one or more data assembly functions. Those functions include activities such as key entry, optical scanning, data formatting and quality control. Each assembly centre may specialize in one of the functions, or it might perform all functions for a specific ocean region. The activity of acquiring non-digital data from outside the IODE system will be time consuming and must be carefully planned.

The World Data Centres (WDCs), Oceanography, will play a major role in building the historical component of the GTSPP database. WDCs will locate, and seek to acquire, historical data sets based upon general priorities established by the GTSPP Steering Group in consultation with the international scientific community. The following priorities will guide this work:
(i) Areas with sparse data coverage, either geographically or temporally;
(ii) Data needed to support special scientific programmes or to better define unusual scientific events; and
(iii) Data required for the continuation of on-going time series data sets.

The WDCs will utilize a two-pronged approach to augment the historical database of the GTSPP as follows:
(i) Digitization of non-digital data currently held in WDCs. The WDCs plan to coordinate the digitization of these data sets and envisage the possible use of optical scanning techniques to accomplish this.
(ii) The WDCs will identify, locate, and seek to acquire, historical data sets currently held by various organizations and individuals in the international community. Some of the methods envisioned to accomplish this task include:
    a) Utilization of marine data inventories, such as ROSCOP (and its successor, the Cruise Summary Report), to determine the availability, location, and point of contact for candidate historical data sets.
    b) Utilization of lists of National Oceanographic Programmes (NOPs), published for many years by IOC, and of Declared National Programmes (formerly published by IOC).
    c) Use of relevant electronic inventories or "Bulletin Boards" where appropriate historical data sets are described.
    d) For selected historical data requirements, as determined by the Steering Group, WDCs may communicate directly with relevant data contacts.

While undertaking these historical data acquisition tasks, GTSPP will recognize the extreme importance of time series data sets to climate research. A major emphasis will therefore be placed on continuing to build the existing time series data sets. As an integral part of GTSPP, particular attention will be paid to the enhancement of such data sets as Ocean Weather Stations currently under development by WDCs.

8. PROJECT PRIORITIES

There are a number of aspects of the project that have a higher priority than others and, therefore, may require more immediate planning. Also, some components must be developed to a certain level before the project can be considered operational. The higher priority areas include:
(i) establishing real-time data flow monitoring procedures,
(ii) implementing procedures to improve the capture of data,
(iii) completing the development of a `scientifically' accepted quality control manual,
(iv) design and implementation of the Continuously Managed Database system,
(v) implementing necessary high speed communications links between centres, and
(vi) 'recruiting' and coordinating the appropriate agencies to undertake the scientific quality assessment (including data analysis and product development) to provide global coverage temperature and salinity data.

One of the more significant aspects of the Project is to monitor real-time data flow on a global scale. Monitoring exercises will be conducted in order to evaluate the quantity of data that exists throughout the near real-time (IGOSS) system. Mechanisms can then be established to ensure that all data that is transmitted in real-time is captured by GTSPP and added to the Continuously Managed Database. This will ensure that minimal data is lost within the communication systems.

Before data can be entered into the database, it must be quality controlled to a minimum level. Therefore, it is necessary that the quality control procedures be developed and agreed upon by the scientific community. This will be one of the first objectives of the Steering Group.

Another high priority of the project is the Continuously Managed Database. This system must be developed, tested and implemented before GTSPP data can be managed. A priority of the Steering Group is to ensure that the CMD algorithms quickly reach a level of functionality adequate for operational use.

A primary objective of GTSPP is to improve the timeliness and completeness of data sets for the scientific community. To achieve this, it must be a priority to develop and implement the communications facilities that are needed by GTSPP to complement the existing systems. Otherwise data can not be transferred within the necessary time frames.

9. PROJECT MANAGEMENT

9.1 Parent Bodies

GTSPP is a cooperative IODE-IGOSS Project and, therefore, has a relationship with the Technical Committees of both programmes. Consequently, the parent bodies for GTSPP are the Intergovernmental Oceanographic Commission and the World Meteorological Organization.

9.2 Steering Group

At the Second ad hoc Meeting on GTSPP in July, 1989, it was determined that a Steering Group should be formed. The Workshop on the GTSPP, (New York, USA, 15-16 January 1990) developed Terms of Reference and a suggested composition for the Steering Group. Both the terms of reference and the composition for the Steering Group were accepted as Recommendation IODE-XIII.4 at the meeting of the Committee on International Oceanographic Data Exchange (IODE-XIII, New York USA, 17-24 January 1990).

The Steering Group will be responsible for coordinating the implementation of GTSPP, continuing development and overseeing the operation of the Project. The Steering Group will continuously review the Project, and actively promote and encourage participation in GTSPP. Scientific guidance to participants will also be provided via the Steering Group which includes members from the scientific community. It will be the responsibility of the Steering Group to prepare, maintain and distribute relevant documentation and to report to IODE and IGOSS contacts, representatives of scientific programmes and appropriate Committees (IGOSS and IODE).

The Steering Group will initially be comprised of representatives from Australia, Canada, France, the USA, and the USSR; representatives from WOCE and TOGA; and an invited scientific advisor. The IGOSS-IODE and IODE-IGOSS Rapporteurs will contribute and representatives from the RNODC's - IGOSS will be invited to participate in Steering Group activities.

The complete terms of reference and composition for the Steering Group are contained in the Appendix.

APPENDIX

Terms of Reference for the Steering Group on the GTSPP
(i) Complete development of the GTSPP Project Plan and Implementation Schedule following the principles described in the current draft.
(ii) Complete development of the GTSPP Quality Control Manual.
(iii) Meet semi-annually at the expense of the participating countries to review the status of the implementation and further develop the GTSPP.
(iv) Actively promote the GTSPP and provide information to the users of GTSPP services, such as the planners of international Science Programmes.
(v) Provide scientific and technical guidance to GTSPP participants in the implementation and further development of the scientific and data management aspects of the six GTSPP elements including:
    a) near real-time data acquisition
    b) non real-time data acquisition
    c) communications infrastructures
    d) quality control procedures
    e) continuously managed database
    f) GTSPP data and information products
(vi) Prepare, maintain and distribute documentation relevant to operation of the project.
(vii) Report after each meeting and as otherwise necessary to keep all IODE and IGOSS contacts, as well as the representatives of the science programmes, informed on the status of implementation of the GTSPP.
(viii) Submit status reports on the GTSPP to the sessions of the Committee on IODE and the JWC-IGOSS.
(ix) Prepare and submit to the sessions of the Committee on IODE and the JWC-IGOSS a report on the status of IGOSS-IODE data flow.
(x) The group will select a Chairman at its first session and will review the Chairmanship bi-annually. The composition of the group is as proposed below. The composition of the group will be regularly reviewed in consultation between the Chairmen of IODE and IGOSS.

Composition of the Steering Group for the First Intersessional Period
(i) One representative from each of the participating countries (initially Australia, Canada, France, USA, USSR) as chosen and funded by the countries, to provide expertise to the project. These representatives may be accompanied by one or more experts;
(ii) A representative from WOCE and TOGA, each nominated by the appropriate Scientific Steering Group, to provide scientific guidance to the project;
(iii) An additional invited scientific advisor selected by the Chairman of the Steering Group on the GTSPP;
(iv) The IGOSS-IODE and IODE-IGOSS Rapporteurs;
(v) Representatives from the RNODC's-IGOSS will be invited to participate in this Group.