The GOSIC Strategy for Selection of Data Sets for the Data Registry

One of the main purposes of the GOSIC is to provide information on and links to the data sets that are being developed and managed by the Global observing systems (GTOS, GCOS and GOOS) by:

• Providing a directory of global observing systems global change data and information, including links to the datasets;
• Pointing to the centers and information systems that hold the major atmospheric, ocean, and terrestrial data and information from the operational observing programs of the global observing systems;
• Providing access to data sets, without holding datasets;
• Presenting information as held in the centers

To provide the data sets as described above, the first requirement is to develop a list of the relevant data sets. There are various ways to proceed to prepare such a list.

First of all, however, it is important to recognize that there are a large number of environmental data sets collected by various organizations for their own various purposes and that many of these data sets are not necessarily suitable for global observing systems purposes.

The science panels that are developing the observing systems have prepared lists of variables, sampling strategies, quality control and analysis procedures, and distribution schedules and archival procedures. In many cases at least some of these variables are already being observed by existing national or international programs for purposes that may or may not include detection, quantification, and prediction of global change. The data will be processed to various standards using algorithms that may or may not meet accuracy, resolution, and spatial or temporal distribution criteria appropriate to the needs of the global observing systems. Thus all data sets that are compiled for a given variable can not be used for global observing systems purposes.

The question is how does the GOSIC decide which data sets should be listed on the GOSIC site and referenced in the Data Registry? There are 3 ways that a dataset may be put on the list:

1) Data Sets from Specific Observing System Elements Developed by the Science Panels are Included. One of the science panels may develop the observing requirements for an element of an observing system including specifying the network, the quality control and processing to be applied to the observed data, the meta data to be attached, the data products to be produced, distribution schedules, and the short and long term archiving responsibilities. In this case the various data bases, data products, meta databases, and the archives should all be listed in GOSIC and links provided for access to the data sets themselves. Examples of this include the GCOS Surface and Upper Air Networks and the GOOS Argo Program.

2) Historical Data Sets Selected by the global observing systems Science Panels are Included. Some of the global observing systems Science Panels have reviewed existing data sets and identified a number of such that should be considered as relevant to global observing systems goals and meeting global observing systems standards. These data sets, often resulting from a particular study or analysis, are added to the Data Registry and links to them are provided.

3) Data Sets from Existing Observing Programs Identified as an Element of an Observing System are Included. All three observing systems are basing their implementation on cooperative arrangements with existing national and international observing systems already in the business of collecting data on variables of interest for matters related to global change. Once an existing program has been identified and has agreed that in addition to serving its present clients it can also be an observing element of the global observing systems, the GOSIC proceeds to list and link the databases developed and maintained by that program. Examples of this include the databases of the Global Sea Level Observing System (GLOSS) and the Global Atmospheric Watch (GAW) atmospheric constituents databases.

For the moment this action assumes that all data sets that belong to a program that has been identified as an element of the global observing systems are considered to be global observing systems data sets. This may be true in some cases but is certainly not true for all. The next step is for the GOSIC to liaise with the global observing systems steering committees and science panels for the purpose of eliminating those data sets that belong to an observing system element but are not considered relevant to the global observing systems.

This provides a starting point from which the designers and data managers for the global observing systems can further develop fully functional end-to-end data management systems to ensure the quality, standards, timeliness, and access needed for global observing systems data.

Updated December 30, 2009