Attendees were Bob Keeley (MEDS) Bob Molinari (AOML), Rick Bailey (CSIRO), Doug Hamilton (NODC), Penny Holliday (WOCE IPO), Yutaka Michida (JODC), Steve Diggs (WHPO), Marie-Claire Fabri (IFREMER).
1. QC of Pacific Data
Scripps is no longer receiving funding to carry out scientific quality control of Pacific Ocean data. The NODC liaison officer in the southwest worked with Scripps to complete the QC of data through 1994 but can not continue. The problem is being attacked on three fronts.
In the short term, Bob Keeley will contact Warren White at Scripps to see if it would be possible to employ someone there to undertake QC of the 1995 data immediately. Bob Molinari has some funds that might be used to this purpose.
Bob Molinari will develop and test some automated procedures to identify temperature profiles that need closer scrutiny. Rick Bailey has offered to collaborate in the development. It is estimated that the development and testing will be completed within a year. A report will be prepared to document the capabilities of the test procedures. Assuming reliable tests can be developed, AOML would then be prepared to take over scientific quality control of Pacific Ocean data.
Yutaka informed the meeting of a new group formed in Japan called the Marine Information and Research Centre. This group is directed by Dr. Nagata, and will assist JODC in the development of oceanographic products. It is known that they are intending to prepare a climatology of the western Pacific. Bob Keeley will prepare a letter to be sent to Dr. Nagata. The letter will request MIRC to take on the scientific quality control of Pacific Ocean data and to extend the area of their climatology to the entire Pacific. Positive results from AOML's work will be available to MIRC.
There is interest in developing more capable quality control test procedures than can reduce the visual scrutiny required with the main goal of being able to identify profiles that require additional attention. Rick Bailey told us of some AI work in Australia. There is also the possibilities of using EOF analyses. This will be discussed at the next formal GTSPP meeting to see if there are resources available to pursue these ideas. Results from Bob Molinari's work will be available about this time.
2. Transition of Expertise between Science Centres and Data Centres
This is an important goal of GTSPP. In Australia, the people employed by AODC and trained by CSIRO have left AODC and so the impetus for this transfer has faltered. In the US, the NODC liaison officer in San Diego had started this task, but has been withdrawn from such work now. At the moment, this goal is farther away than it appeared a year ago.
CSIRO is trying to move the operational running of the XBT programme to BMRC. They will attempt to transfer the expertise to the appropriate people at BMRC.
The chairman of GTSPP will send a letter to the Director of NODC to encourage him to support the goal of transfer of expertise to data centres. Such support could be to help in the development of software and techniques that will automate the identification of data that require closer attention. (See item 4 below as well).
Bob Keeley discussed the software that MEDS uses to identify duplicate cruises. Bob Molinari expressed interest in the software. He will contact MEDS as appropriate.
There was no representative of AODC at the meeting and the above report was met with a difference of opinion after the meeting.
3. Brest Participation
Marie-Claire informed the meeting that a new director has changed the focus of their data centre. They will no longer take part in on-going projects that have no definite end point. Rather, their activities will be more project oriented and confined to the northeast Atlantic and Mediterranean. The consequence is that their relationship to GTSPP (and UOT) will change to a more distant participant. They will still provide data collected by France to GTSPP and still wish to receive data collected in waters of interest to them. They also offered to preserve GTSPP data in that format (if circumstances should arise to warrant this), although they would not be loaded into a database.
4. P3 format
NODC has developed a new archive structure for profile data (known as P3). This structure is similar but not the same as GTSPP. There is the intention to migrate all of the diverse profile data archives into this one structure. There is some question whether or not P3 can contain all of the information now held in the GTSPP structure. Doug agreed to explore this in detail and to prepare a report to be circulated to GTSPP members. This report will be distributed by the end of March, 1998.
NODC is undertaking to be the archive for all WOCE data. It is expected that UOT and WHP data (at least) would be managed using the P3 structures. At the DPC meeting, NODC offered to hold a small meeting with representatives from WOCE DACs to discuss the functionality needed by P3 to meet WOCE requirements. Since UOT uses GTSPP format, GTSPP representatives should attend this meeting with Doug's report in hand to ensure that UOT/GTSPP concerns are voiced.
There was some question about whether the new P3 archive will conform to the concepts of a continuously managed database in support of GTSPP. In addition, GTSPP invests substantial effort to compile metadata as discussed above. The chairman of GTSPP was asked to prepare a letter to go to NODC expressing the requirements for continued support of a CMD and preservation of metadata. This letter will be passed to the co-chairs of DPC for their comments and the letter sent with joint authorship of DPC and GTSPP.
5. Data submission
All data from 1990 to 1994 have passed through scientific quality control and been returned to NODC. Data for 1995 have been QCed by AOML and returned to NODC. CSIRO will attempt to complete this work and submit the data in time for issuance of the WOCE CD. As noted above, scientific QC of 1995 data from the Pacific are unlikely to be done in the near future. NODC has the necessary software to merge the returned data back into the GTSPP archive preserving the flags and metadata added by the science centres. This has been done for Atlantic data to the end of 1995, and will be completed for Pacific data to the end of 1994. There are still some problems in merging the data from CSIRO but this is being worked on now.
Discussions turned to what can be done to be more active in pursuing data. First, Bob Keeley has shown that high resolution XBT data collected through the SEAS programme do reach NODC more quickly than other data. There are still things to be done here. NODC can prepare a report showing how many of the low resolution profiles have been replaced by high resolution profiles. Rick will send out a call for all 1997 data to be sent to NODC. Doug will prepare a report in April of 1998 for data received to the end of 1997 identifying which low resolution profiles are still not replaced by high resolution profiles. He will send the NODC report to SOOPIP chairman (Rick Bailey) for action. Rick will look into getting the ship operator's file to NODC so that they can connect the ships to the operators.
Yutaka reported that approximately half of the XBT data collected by Japan uses digital recorders rather than strip charts. Presently JODC sends XBT data sampled at standard depths to NODC. He will look into sending the digital XBTs at 1 m resolution now rather than at standard depths.
JODC has sent low resolution XBT data to MEDS at the end of each month. These are profiles sampled at 4 depths only. This data transfer has been intermittent lately and Yutaka agreed to look into this to improve the data transfer to MEDS. He also agreed to send the latest table that provides ship call signs to match the ship codes used in the data.
Data from P-ALACE floats deployed by Woods Hole should shortly appear on the GTS. Data from floats deployed by AOML should also begin soon. Both of these sources use floats that report temperatures only. Also, no delayed mode data can be expected from these floats. Bob Keeley will look into the floats deployed by Canada. Marie-Claire will look into the PIRATA floats of France and let Keeley know if the data from these could be expected to appear on the GTS.
CSIRO, France, Japan and AOML deploy XCTDs. AOML is working with Noumea on evaluating them. Rick and Bob will cooperate to try to get these data (as appropriate) on the GTS. Yutaka will look into the same problem for XCTDs deployed by Japan. Keeley reported that modifications to the TESAC code to permit information about probe and recorder type (as now sent in BATHY code) have been sent to the appropriate WMO committee. He will check on the progress of this and report to the others.
Depending on if other profile types are associated with T and S profiles, some of the CTDs received at NODC are placed in their profile archive, not in the GTSPP archive. Trying to match TESACs to high resolution profiles in GTSPP can fail because of this. Doug agreed to check on these practices at NODC to ensure that matches are properly done and include profiles in both GTSPP and profile archives at NODC.
Some questions were raised about whether there were enough data received from earlier years of GTSPP to justify a re-analysis. Doug agreed to generate a report for each year of GTSPP that shows the date that data entered the archives for data collected in each individual year.
6. Interaction with other programmes
At a J-DIMP meeting last July in Tokyo, Keeley reached agreement with Hasagawa of JMA to explore cooperative arrangements between GTSPP and NEAR-GOOS. Discussions have progressed such that MEDS will soon make the necessary connections to begin providing real-time data to NEAR-GOOS likely on a 3 files/week schedule. Keeley will then pursue how to receive data in real-time from NEAR-GOOS. Since JODC is managing the delayed mode side of NEAR-GOOS, Yutaka and Doug agreed to begin discussions on making the necessary arrangements to accomplish exchanges of delayed mode data.
Keeley attended a SOOPIP meeting last year representing GTSPP. At that meeting GTSPP took on one of the tasks (formerly with the IGOSS Coordinator) of tracking the performance of ships in making reliable real-time measurements month by month. Keeley and Bailey are cooperating to develop a pictorial representation of how well XBT sampling is being performed. This product will be posted on MEDS web site and will be updated monthly. The presentation of the first map will appear in late February and show results starting in January, 1998.
AOML has received funding to run a GOOS Centre. This centre will coordinate all of the NOAA observing networks including drifters, observations from VOS and NOAA ships, and data from the TAO array. The NOAA ships and VOS ships report surface meteorological observations, XBTs, and thermosalinograph (TSG) data. AOML will optimize and monitor the data collection to ensure effective sampling and ensure delivery of the data to modeling centres. NODC will be the archive for these data. At the same time, NODC is developing ways to manage data collected by ships underway. Doug and Bob Molinari will discuss what steps need to be taken to ensure a smooth transfer of data to NODC.
MEDS offered to transfer the contents of its archive of TRACKOB (real-time TSG data) to NODC or AOML as appropriate when they were ready to receive it. MEDS also offered to continue to receive and decode TRACKOB data and forward these to AOML to assist in them getting their operation running quickly.
Bob Molinari noted that many of the objectives of GTSPP coincide with the goals of DPC. So, the interaction between science and data centres is supported by DPC and is central to the operation of GTSPP. Bailey raised the concern that GTSPP needed to be careful to ensure that the data collection, end-to-end process that is desirable to GOOS is not lost. It was suggested that GTSPP join the DPC or operate under its auspices. It could then influence others in developing the data handling procedures of other types of data of interest to DPC. If/when data management is incorporated in CLIVAR, GTSPP practices then would be included. This arrangement would ensure the scientific contacts needed by GTSPP. Keeley agreed to write a letter that discusses the terms under which GTSPP would feel comfortable as part of DPC and to circulate this to members. He will also discuss this idea with the chairman of GTSPP.
7. WOCE profiles and other data types
Scientific quality control has been confined to temperature profiles even though some salinity data are collected by GTSPP. Both CSIRO and AOML have the intention to expand their quality assessments to include salinities, but no schedule for these developments were given. It should be noted that real-time salinity data do pass through data centre quality control.
8. Review of Action Items from Previous Meetings
a) Participants are requested to seek out data from profiling and undulating instruments that collect subsurface temperature and salinity data and put them into the system in delayed mode. A standard is required which describes any additional metadata to be carried with the data. MEDS and NODC are to produce a draft standard for comment by all participants.
Action: NODC and MEDS
Keeley will prepare a document describing its experience with these data. He will also talk to the ICES MDM to circulate the document that they have prepared on this.
b) All centres are requested to be proactive in getting the delayed mode data for which real-time versions have been received and forwarding it to the CMD in the US NODC.
Action: All Participants
See discussions of this meeting.
c) MEDS is to archive TRACKOB data received from the GTS and make it available to IODE in delayed mode.
Action: MEDS
MEDS is doing so,
d) The WOCE Science Centres will continue to be involved in QC but a transfer of expertise to data centres must be done as the Science Centres will probably not be able to continue to do all the QC as they do now. Formal links (national, and where possible, international) are recommended (e.g., personnel visits, exchanges, commitments to training by both sides, etc.)
Action: Science Centres, NODC, MEDS
See discussions of this meeting.
e) A study is to be undertaken to examine the feasibility and how to implement use of a common climatology for GTSPP QC by all participants.
Action: MEDS
This is not considered appropriate. As basin climatologies are developed (such as the Atlantic and Pacific) these will be adopted.
f) The present procedures for notification of ships that routinely report bad data are to be modified. MEDS will send reports directly to the ship operators to save time. The SOOP Co-ordinator will continue to contact operators that continue to submit bad data in spite of such a notification.
Action: MEDS and SOOP Co-ordinator
MEDS is doing this.
g) Procedures are to be implemented by NODC and MEDS to remove low resolution profiles from the CMD when the high resolution profile has been found to be bad by the originator and therefore has not replaced the low resolution version.
Action: NODC and MEDS
As long as the high resolution profile (flagged wrong) is received by NODC the low resolution profile is flagged as a duplicate. What also needs to happen is the flags of the real-time profile need also to be changed. Doug will look into the procedures needed to accomplish this at NODC and with discussions with Keeley sort out how MEDS real-time archives will also be notified. (is this needed?)
h) MEDS and NODC are to prepare and circulate an agreed definition on the contents of all fields in the GTSPP format.
Action: MEDS and NODC
MEDS and NODC will be updating their web sites based on the UOT CD. This concern is met by the CD. Each centre will assume responsibility for the contents that it is decided they will maintain. Keeley will work with Hamilton to make this happen.
i) Data centres are encouraged to retain unstructured metadata and supply on request. Not to be exchanged routinely within GTSPP.
Action: MEDS and NODC
This is being done.
j) Develop a procedure for assigning parameter and other codes to new variables so that common codes are always used for the same variable. Procedure and Codes are to be posted on a WWW page.
Action: Brest, AODC, NODC, MEDS
This will result from installing CD contents on web sites.
k) Data centres are to attach a time to reports that do not have a time field and set the date-time flag to 5.
Action: MEDS, NODC
Done at MEDS.
l) Standards are to be set, and metadata for collection, processing systems in regard to fall rate, etc. are to be obtained and appended for all delayed mode data as a routine activity.
Action: Data Centres and Suppliers of Data
This is done by data centres.
m) MEDS, NODC and Scripps are to develop a policy in regard to judicious use of the history record.
Action: MEDS, NODC, Scripps
Doug will prepare a document that describes what has been done as a consequence of merging data returned from science centres. This will be put on the CD
n) GTSPP and GODAR to establish agreed-upon standards for metadata to be retained with the data, QC procedures, the meaning of the QC flags, and duplicates identification. A document is to be prepared for the next IODE meeting.
Action: Chairman of GTSPP Steering Committee, Project Leader of GODAR
Not done
o) A simpler format for circulating GTSPP data is to be developed by MEDS, IFREMER, Scripps and AODC.
Action: MEDS, IFREMER, Scripps, AODC
It was decided to make available more information about how to read the GTSPP format, but not to sacrifice information to a simpler format.
p) MEDS is to co-operate with the Canadian GTS Centre in Dorval in pilot monitoring projects for oceanographic data on the GTS.
Action: MEDS
Not done, but MEDS has cooperated with the pilot monitoring of ocean data on the GTS carried out by WMO.
q) France and USA to exchange data as necessary to have a 100% overlap in holdings and work out routine exchange of data between IFREMER and the CMD so as not to get out of step in the future. The routine exchange is to be worked out once the present exchange is complete and the causes of the present mismatch are understood.
Action: IFREMER, NODC
This is now happening regularly.
r) To prepare annual reports with a description of the flow of data through the WOCE UOT centres.
Action: WOCE Science Centres, WOCE IPO
This is being done in reports to DPC.
s) Arrange the location of the next meeting in consultation with IOC Secretariat.
Action: Chairman of the Steering Group
Not planned but expected to be in approximately 1 year.
t) Develop new Terms of Reference for the Steering Group as an end-to-end systems development steering group to be referred for comment and appropriate action to the next IODE officers and IGOSS Bureau meetings.
Action: Chairman of the Steering Group
Not done
u) Member States are to examine and make what commitments are appropriate and possible to the continuation of the SOOP operations. NODC and MEDS have already committed to continue the services now provided.
Action: IOC and WMO Secretariats to inform Member States of the Need
Done.
v) Since GTSPP deals only with observations that include sub-surface data, management of surface only observation data is to be referred to IGOSS and IODE to avoid a later rescue effort. (A possible solution would be a proposal for a GTSPP-like project for surface observations.)
Action: Chairman of the Steering Committee
See comments of this meeting.
w) To consider ways to attract new MS to GTSPP for various tasks and services (on-going).
Action: All Participants.
No discussion beyond trying to draw Japan in further.
x) There is a need to write to GOOS in regard to GTSPP. The letter is to inform the Chairman of J-GOOS and I-GOOS of the clients of GTSPP and their needs for data that might be available from the marine services module of G005, and of GTSPP data and services that might be of interest to GOOS.
Action: Chairman of the Steering Committee
Done and see comments from this meeting.
y) In the future, development and publication of long time-series and climatologies will be the responsibility of GODAR.
Action: GODAR Project Leader
Unknown.
z) There may be a requirement for a PC programme for key entry of data to the GTSPP format so that GODAR will have fewer formats to deal with and so that the data is then available to GTSPP centres in a format they can deal with. This requirement should be reviewed with GODAR.
Action: Chairman of the Steering Committee
Not done.
aa) A revised version of the GTSPP Project Plan is to be published.
Action: IOC Secretariat
Unknown.
ab) Demonstrate in published form the added value of QC as in the Scripps before and after QC maps.
Action: WOCE Science Centre
Bob Molinari carried out some preliminary work. Results showed that in looking at statistical results for broad space and time scales, added value was minimal. Work will continue looking at this problem.
ac) AODC to contact all sites and make available links from their GTSPP home pages from the AODC home page as an interim measure for publicizing GTSPP data and services.
Action: AODC
Unknown but Keeley will investigate.
ad) Republish the GTSPP brochure.
Action: NODC
Doug will check on this.
ae) Develop a series of linked Web Sites and Appropriate CD-ROMs to publish data and information on-line on overview of programme, contacting participants, GTSPP clients, how they use GTSPP, success stories, list of products and .services, complete set of publications, on-line data and data flow reports.
Action: Chairman of the Steering Committee
Not done.
af) Conduct demonstrations of GTSPP at all appropriate national and international meetings
Action: All Participants
Represented at SOOP, J-DIMP, ...
ag) Use IMS, scientific journals, other publications to demonstrate the benefits of GTSPP.
Action: All Participants, IOC and WMO Secretariats .
Not done.
ah) Use the series of WOCE Scientific Workshops as a forum and provide datasets for the Workshops.
Action: NODC
Done.
ai) Attempt to publish platform independent software for all aspects of GTSPP by following the specifications for common software as specified in the GTSPP Summary Report.
Action: All Participants.
Not done.
aj) Announce the availability of the GTSPP CD-ROM in the IMS and International Newsletters.
Action: Chairman of the Steering Committee
Unknown.
ak) Propose to GTSPP members simple electronic forms with information on data sets received and status of processing which may be circulated among GTSPP participants by mail or put on a common directory or the Web Site.
Action: IFREMER, NODC
Unknown
al) Distribute the datasets for 1993 to the Science Centres.
Action: NODC (Check with NODC)
Done and 1994 and 1995.
am) Data Flow Monitoring. MEDS will continue to produce the monthly GTS data flow analysis. NODC will continue the reporting of the TOGA/WOCE line occupations to the WOCE IPO. WOCE will continue to report on the data flow through WOCE UOT centres. WOCE and IODE will produce a report on the data flow through the system.
Action: MEDS, NODC, WOCE IPO
Done.
an) The WOCE Science Centres are to complete QC of delayed mode data in their possession for 1990-1992.
Action: WOCE Science Centres
See comments from meeting.
ao) Participants will in future advertise software development plans on Web pages, etc., so that interested parties can contact and perhaps co-operate to develop software that will run on multiple platforms.
Action: All Participants
Unknown.
ap) NODC to provide duplicates checking software to SIO for inclusion in SIO quality control application and to provide access to the system as a service for use.
Action: NODC, Scripps
Unknown
aq) CSIRO and Scripps visualization software for QC to be provided to data centres, in particular MEDS and NODC.
Action: CSIRO, Scripps, NODC, MEDS
Not requested by MEDS. Unknown for others..
9. Action Items from this Meeting
1. Bob Keeley will contact Warren White at Scripps to see if it would be possible to employ someone there to undertake QC of the 1995 data immediately.
2. Bob Molinari will develop and test some automated procedures to identify temperature profiles that need closer scrutiny. Rick Bailey has offered to collaborate. A report will be prepared to document the capabilities of the test procedures.
3. Bob Keeley will prepare a letter to be sent to Dr. Nagata. The letter will request MIRC to take on the scientific quality control of Pacific Ocean data and to extend the area of their climatology to the entire Pacific.
4. The chairman of GTSPP will send a letter to the Director of NODC to encourage him to support the goal of transfer of expertise to data centres.
5. Doug Hamilton agreed to prepare a report to be circulated to GTSPP members on how well P3 will accommodate GTSPP format. This report will be distributed by the end of March, 1998.
6. GTSPP representatives should attend a small meeting with representatives from WOCE DACs to discuss the functionality needed by P3 to meet WOCE requirements to be held at the Halifax WOCE meeting.
7. The chairman of GTSPP was asked to prepare a letter to go to NODC expressing the requirements for continued support of a CMD and preservation of metadata. This letter will be passed to the co-chairs of DPC for their comments and the letter sent with joint authorship of DPC and GTSPP.
8. Rick Bailey will send out a call for all 1997 data to be sent to NODC.
9. Doug Hamilton will prepare a report in April of 1998 for data received to the end of 1997 identifying which low resolution profiles are still not replaced by high resolution profiles. He will send the NODC report to SOOPIP chairman (Rick Bailey) for action. Rick will look into getting the ship operator's file to NODC so that they can connect the ships to the operators.
10. Yutaka Michida will look into sending the digital XBTs at 1 m resolution now rather than at standard depths to NODC.
11. Yutaka Michida agreed to look into improving the data transfer to MEDS. He also agreed to send the latest table that provides ship call signs to match the ship codes used in the data.
12. Bob Keeley will look into the floats deployed by Canada. Marie-Claire will look into the PIRATA floats of France and let Keeley know if the data from these could be expected to appear on the GTS.
13. Rick Bailey and Bob Molinari will cooperate to try to get XCTD data (as appropriate) on the GTS. Yutaka will look into the same problem for XCTDs deployed by Japan.
14. Keeley reported that modifications to the TESAC code to permit information about probe and recorder type (as now sent in BATHY code) have been sent to the appropriate WMO committee. He will check on the progress of this and report to the others.
15. Doug Hamilton agreed to check on practices at NODC to check for duplicates across the GTSPP and P3 archives.
16. Doug Hamilton agreed to generate a report for each year of GTSPP that shows the date that data entered the archives for data collected in each individual year.
17. Keeley will then pursue how to receive data in real-time from NEAR-GOOS. Yutaka Michida and Doug Hamilton agreed to begin discussions on making the necessary arrangements to accomplish exchanges of delayed mode data.
18. Doug Hamilton and Bob Molinari will discuss what steps need to be taken to ensure a smooth transfer of AOML GOOS Centre data to NODC.
19. MEDS offered to transfer the contents of its archive of TRACKOB (real-time TSG data) to NODC or AOML as appropriate when they were ready to receive it.
20. Keeley agreed to write a letter that discusses the terms under which GTSPP would feel comfortable as part of DPC and to circulate this to members. He will also discuss this idea with the chairman of GTSPP.