Report on the Second GOSIC Review

The second GOSIC Review was carried out by a review team appointed by the steering committees of GCOS, GOOS and GTOS. This report has been submitted by Dr. Matthew Howard, chairman of the review team.

GOSIC REVIEW 2003

Prepared by the 2003 GOSIC Review Team

Matthew Howard (Chair) representing GOOS / Lesley Rickards representing GOOS / David Legler representing GCOS / Stefan Rösner representing GCOS / Géraud Servin representing GTOS / Robert Cook representing GTOS

Submitted 28-October-2003

Summary

On October 7-8, 2003, a six-person team representing GTOS, GOOS, and GCOS (henceforth referred to collectively as G3OS) conducted a site review of the Global Observing Systems Information Center (GOSIC) for the chairs of the G3OS Steering Committees. This report documents the review and presents the findings and recommendations of the 2003 GOSIC Review Team (RT).

The RT concluded the review with a generally positive evaluation of GOSIC. The RT feels that GOSIC performs a necessary service and should be continued. The RT judged that GOSIC continues to meet its Terms of Reference (TOR) and has adequately addressed the recommendations made to them by the 2001 Review Team. Some of the recommendations made in 2001 were directed to the G3OS Steering Committees. We feel these were not adequately addressed and so they are repeated again in the 2003 recommendations. Foremost of these is a recommendation that the G3OS nominate and maintain data coordinators who will identify and select the G3OS data sets that should appear in the GOSIC data registry.

The RT found the GOSIC web site to be much improved relative to the previous review. The web site facilitates access to significant amounts of information about the G3OS programs and data sets. The GOSIC website provides an integrating overview of the three observing systems by presenting information in analogous and common views for each observing system and by collecting common information on a G3OS page. Other collective views may be useful, such as by IGOS Theme or by such topics that a data coordinator may deem useful.

New items were added to the web site following the last review. A cross-program search facility was added but needs refinement and/or simplification. A contact form was recently added which provides a convenient mechanism for obtaining user feedback. The GOSIC staff showed initiative by identifying and tagging more than 250 data sets as G3OS data sets in NASA’s Global Change Master Directory (GCMD). Many of these data sets were added to the GCMD by GOSIC staff and would not have been included in the GCMD otherwise. Additionally, GOSIC has been working with GCMD to add GTOS keywords and G3OS identifiers to data sets in the GCMD. A national commitments page was added under the GOOS section. A GTOS data matrix is under development. This matrix is an alternate view of the GTOS data holdings that facilitates rapid access to data. It is an excellent example of the value added through close interactions between GOSIC and a data coordinator.

GOSIC communicates well with GTOS through Géraud Servin, and to a lesser extent with GOOS through Bert Thompson. Bert Thompson may need to be replaced if he relinquishes his role in 2003. GCOS needs to identify a GCOS data coordinator. Identification of G3OS satellite data sets has been initiated by GOSIC but the input of an expert in remotely sensed data sets is needed to better select data sets for registry. John Bates (NCDC) was repeatedly mentioned as a possible candidate for satellite data coordinator. Apparently, communication between GOSIC and the G3OS Steering Committees is infrequent and indirect. The RT feels that regular communication between GOSIC and the G3OS Committees is important to the success of GOSIC.

It is not clear who uses GOSIC, what they seek, and if they are satisfied. The web site access log has not been processed in a way that reveals the affiliation of visitors. The recently added feedback form has received little use thus far although positive feedback was received from several regional GOOS. G3OS and GOSIC should work together to refine the definition of the GOSIC user community. GOSIC should then build more awareness within that community about the purpose and resources of GOSIC. The RT recommends that GOSIC publish articles about GOSIC in appropriate newsletters and bulletins, and develop communications materials for email, conference brochures and posters.

The RT feels GOSIC should not adopt any new IT technologies at this time, primarily because GOSIC doesn’t hold data. New technologies (e.g. XML, OPeNDAP, etc.) will facilitate machine interoperability, ease searches, and merge datasets. The holders of data and/or metadata should implement these new technologies. NASA/GCMD is currently working on modernizing the GCMD to incorporate developing standards and technologies. Because GOSIC is supplying critical metadata to GCMD, they are serving as an important link in the development of modern G3OS systems.

GOSIC expects to be transitioned to a permanent location within another organization and has submitted a proposal to NOAA that includes a plan for this transition. (The RT feels that input should have been sought from the individual G3OS organizations prior to the submittal of this proposal). The RT makes no recommendations about which agency should assume GOSIC, but given the international character of global observing systems, the RT feels the selected agency should be one that is acceptable to an international community. Howard Diamond (NOAA/NESDIS) and August Shumbera (NCDC/WDC-Meteorology) expressed some interest in operating GOSIC. The RT expects the selected organization will give equitable attention to each of the observing systems.

GOSIC REVIEW 2003

The 2003 review of GOSIC was held October 7-8 2003 at the University of Delaware’s College of Marine Studies in Lewes, Delaware, USA. The following supporting items were provided as PDF or MS word files along with this report.

The Draft Agenda for the 2003 Review of GOSIC List of Attendees* Terms of Reference for GOSIC Terms of Reference for the 2003 Review of GOSIC Report of the 2001 Review of GOSIC with imbedded responses made by GOSIC in 2001 Progress of GOSIC since last review (prepared by GOSIC for this review). The GOSIC Strategy for Selection of Data Sets for the Data Set Registry

* Matthew Howard and Howard Diamond were on the 2001 GOSIC Review Team. Géraud Servin was an observer during the 2001 Review of GOSIC.

Tuesday October 7, 2003

Following introductions, the RT adopted the draft agenda with a modification to include a brief talk the following morning on the GCMD by Mr. Gene Major who is the GCMD Task Leader. The Review Team then met for the first of three executive sessions; GOSIC staff and observers were excused from all executive sessions. The RT went over the Terms of Reference for the 2003 Review of GOSIC and the objectives (quoted below) given to the RT Chair by Colin Summerhayes.

“Objectives: To find out how well the programme is meeting the Terms of Reference and where GOSIC should be going in the future. To consider whether of not GOSIC is still needed. If it is needed, how should it be improved, and where should it ultimately be located? If it is not needed, how might GCOS, GTOS and GOOS manage its separate components?”

The RT discussed their pre-review impressions and concerns about GOSIC. Several members of the RT reported that their respective communities were unaware of the existence of GOSIC. Even after being made aware, those polled saw little value in using GOSIC, mostly because they were already knowledgeable about programs and data sets in their field and (apparently) don’t need information or data from other programs. The GCOS community appears to be more aware of GOSIC than the other two communities. The GCOS opinion of GOSIC is generally negative, especially within certain sub-groups of GCOS. All of the RT were initially concerned that information on the GOSIC web site might be a simple duplication of information available on the home pages of the individual observing systems, but the RT did not find this to be the case.

GOSIC staff and observers were invited back and the meeting continued with presentations by Ferris Webster, Ron Wilson, Christine Lief and Douglas White of GOSIC. Webster discussed the genesis of GOSIC. Wilson discussed the history of GOSIC, working relationships with the G3OS, and how candidate data sets were identified and screened for inclusion in the registry. While the inclusion criteria seems to be based on sound principles (see supporting documents provided in separate files along with this report), candidate data sets are initially identified through the tedious method of gleaning them from G3OS meeting reports and documents or from documents produced by organizations listed as G3OS members. In effect, GOSIC decides what is an important data set. The preferred approach is for data coordinators from each element of G3OS and remote sensing to designate data sets. This approach has been used to good effect for GTOS with the help of Géraud Servin and to a lesser degree for GOOS via Bert Thompson. Lief talked about progress made registering data sets with the GCMD and how much more was accomplished in a short time with the help of a motivated GTOS data coordinator. In fact, this approach has led to a backlog of GOSIC data sets submitted to GCMD; they have not been able to keep up with Lief. GOSIC has added information to their web site about GOOS national commitments. A search facility was added to the web site to search local pages and holdings at the GCMD. GOSIC has been working with regional elements of GOOS to develop web resources. Douglas White showed statistics for the web site visitors. The statistics showed most visitors were from the United States and GOSIC’s main page was the most often visited. The RT suggests web site usage be processed to show usage by domain and organization.

Following the lunch break there was a 3-hour question and answer period with GOSIC staff and observers followed by a 1-hour closed executive session.

Several RT members reported that resources are not available within their G3OS organizations to maintain their respective web sites and in some cases materials appearing on the GOSIC web site were more current than their own. Furthermore, the data information available on the G3OS web sites did not improve when additional resources were provided to G3OS secretariat offices. Given limited resources at the three observing system offices, perhaps the G3OS should view GOSIC as a shared resource.

While there will obviously be cross-linking and duplication of web site materials, the RT generally felt GOSIC’s one-stop organization should be of value to scientific/programmatic managers, steering committee members, new researchers in climates, engineering firms, and anyone desiring an overview of established and emerging observing systems. The RT feels these groups are likely to be the core of the GOSIC user community. There may be others.

While the RT applauds GOSIC for taking the initiative to identify and designate G3OS data, several of the RT wondered why some data were included while others were not. While it is likely that this question will always be asked, no matter who chooses the data sets, we feel data coordinators are best suited to designate G3OS data sets.

Two members of the RT were involved with the 2001 Review of GOSIC and so had browsed the GOSIC web site extensively in 2001. Thus, they were able to assess the degree of web site improvement since the last review.

Wednesday, October 8, 2003

The RT asked a few more questions of the GOSIC staff and observers. Then Gene Major talked about the GCMD, their efforts to develop program-specific portals, and future activities. The RT then went into the final executive session to prepare lists of achievements and recommendations. The afternoon session consisted of a summary presentation of the RT’s findings and recommendations to GOSIC staff and observers.

ACHIEVEMENTS

The RT feels the following items are significant achievements of GOSIC:

1. The GOSIC web site has been greatly improved since the last review. Common elements on each page make the web site more unified and easier to navigate. A wealth of information is available and presented in an organized way. A search capability, data set registry, and feedback facility have been added. The graphical representation of data flows and quality-control activities, etc. are valuable resources that may not be available elsewhere.

2. The hiring of Ms. Christina Lief as full-time staff in June 2002 has proven to be very beneficial to the GOSIC program and to the GCMD effort to design a GOSIC Portal.

3. Direct collaboration with the GTOS data coordinator provided a good example of how the G3OS and GOSIC should work together.

4. The use of a common format to present a systematic overview of the three observing systems is effective and informative. Aggregation of the G3OS programmatic information into a one-stop location is very convenient.

5. 250 G3OS data sets were identified by GOSIC and tagged as such in the GCMD. Many of these data sets were identified and registered by GOSIC. Registration of data sets in the GCMD should be required of all programs but often is not. Registration is tedious and involves the assembly of a lot of metadata. The fact that GOSIC has done this is of lasting benefit to the community.

6. Flagging G3OS datasets in GCMD using keywords such as G3OS/GCOS will allow G3OS data sets to be easily extracted from the GCMD web site.

7. The GTOS data matrix is an innovative presentation of data availability. Producing new views of data according to activities and keywords is encouraged.

8. GOSIC demonstrated initiative by preparing “The GOSIC Strategy for Selection of Data Sets for the Data Set Registry,” especially in the absence of G3OS guidance.

9. Contact Information & Feedback Form has been added recently to the web site.

RECOMMENDATIONS

A. GOSIC Web site content

We recommend that GOSIC:

1. Use G3OS experts when available to identify G3OS datasets for inclusion in GOSIC. We expect the G3OS to appoint data coordinators who may designate data sets themselves or may enlist the aid of data experts to designate data sets.

2. Continue to register G3OS datasets with GCMD.

3. Present G3OS programs and data in alternate views by organizing information into useful groupings e.g. IGOS themes or by climate, biodiversity, coastal zone, freshwater, carbon cycle, etc. or other groupings that data coordinators may make.

4. Have G3OS designees review information presented on the GOSIC web site for accuracy and approval. In some cases GOSIC has produced pages based on information gleaned from many sources. In effect this constitutes new material that should be checked by the organization that the new material purports to represent.

5. Complete all GCMD metadata fields when possible. Also, GCMD metadata should refer to GOSIC when GOSIC has additional information. For example, when GOSIC maintains information about a program and that program produces a G3OS data set, then the GCMD metadata for that data set should refer to GOSIC as a holder of additional information.

6. Display all G3OS calendars on GOSIC web site. Ideally, the calendars would be maintained by the individual G3OS web site administrators and linked-to by the GOSIC page. The calendars should not be merged, just displayed so that all calendars can be viewed together.

7. Link to pages on G3OS web sites that provide additional data services e.g. DODS. If a data source offers extended data services e.g. aggregation, sub-setting, overlays, then this information should be included on GOSIC pages. GOSIC Outreach We recommend that GOSIC:

8. Write articles for newsletters highlighting new and existing data, information and services available from GOSIC. Articles should be submitted to appropriate publications e.g. CLIVAR, TEMS distribution list, the WMO bulletin, EOS, BAMS, etc. Periodic GOSIC newsletters should be produced and sent to members of the GOSIC user community – see 22 below.

9. Provide outreach material e.g. conference posters, brochures, etc. for workshops and meetings. When possible, have a GOSIC representative attend these venues.

10. Prepare an annual report and submit it to G3OS. Make presentations before the G3OS steering committees and sub-groups (AOPC, OOPC, TOPC etc.) to update them on GOSIC and to solicit their input and guidance.

 B. GOSIC Web site functionality

We recommend that GOSIC:

1. Reduce the number of search options on the web site from four to two: 1) a free text search of GOSIC web pages and 2) a search of data sets through the GCMD portal. This would eliminate redundant and/or inconsistent results. For example, we noted inconsistent results were obtained when searching the GOSIC data sets and the GCMD data sets for the word “WOCE”.

2. Track web site visitation by organization and domain i.e. NOAA, USGS, etc. and .edu, .org, .gov, etc. Keep track of the most visited pages. Keep track of the number of visitors referred to GOSIC pages from the GCMD’s GOSIC Portal page. We think this will help G3OS/GOSIC to better define the GOSIC user.

3. Enable GOSIC web site visitors to indicate their desire to subscribe to receive GOSIC newsletters. This might be done through an option on the feedback form. These users would receive occasional newsletters via email – see 17 above.

4. Conduct an annual peer review of the navigational ease of use and ability to locate content on the web site. Enlist someone familiar with environmental sciences but not with GOSIC to exercise the web site from a fresh perspective.

5. Use drop-down menus on the navigation links on the left side of each GOSIC page to accelerate and facilitate navigation. Recommendations for G3OS We recommend that the G3OS Steering Committee:

6. Nominate and maintain a data coordinator for each of the observing systems and for remote sensing systems to identify G3OS datasets.

7. Provide annual guidance, review, and endorsement to GOSIC and provide input on the transition of GOSIC to a new location.

8. Work with GOSIC to define their user community; these users might be elements of: G3OS, scientific and research community, program managers, policy-makers, etc.

9. Define thematic structures for GOSIC. Programs and data sets can be organized on GOSIC pages in views other than GOOS, GTOS, and GCOS. For example by themes such as Harmful Algal Blooms, Coral Reefs, etc. Future of GOSIC After the transition, GOSIC should:

10. Seek annual guidance, review, and endorsement from G3OS.

11. Ensure a fair balance of treatments of GCOS, GOOS and GTOS.

12. Keep current with latest technologies in data information and management systems e.g. OPeNDAP, Live Access Server, metadata international standards, GIS, etc.