The purpose of this paper is to stimulate discussion on the development of linked data centers that, together, would provide the broad user community with easy access to comprehensive sets of data and products for climate research in the Indian Ocean region. Our focus should be ocean data and products, but climate research requires the full scope of data and products including atmospheric, air-sea flux, and other air-land-sea products.
The greater Indian Ocean domain is subject to substantial short-term natural climate variability, and is both affected by and contributes to global environmental change. A major fraction of the world's population lives in the neighboring region, and climatic signals both bring prosperity and cause considerable stress on its inhabitants. Thus, understanding the nature and predictability of the regional climate system is of great societal importance, and warrants a concentration of effort and resources to accelerate progress.
The amount of oceanographic data available to study climate problems has increased dramatically in the last decade, owing to the successful outcomes of recent international observational programs such as WOCE, and to advances in satellite technology. When currently planned monitoring programs (e.g., Argo, time-series stations) are underway, this data stream will undergo another jump in magnitude. Despite its availability, however, the data are often underutilized, largely a consequence of its lack of easy accessibility. Thus, a primary motivation for regional data centers is to substantially increase the accessibility of climate data and products to a broad community.
In this paper we review the Argo regional data center concept, provide an overview of ongoing and planned data center activities for the Indian Ocean domain, and provide a draft vision and plan for a linked, regional data center for the Indian Ocean.
The possible role of regional data centers was discussed at the 4th Argo Science Team Meeting (AST-4) held at CSIRO, Hobart, Tasmania on 12-14 March 2002 (AST-4, 2002). The following quote from the report outlines the desired activities and possible organization of a regional data center.
Activities that are envisioned for regional data centers include:
Regional data centers may be contributed by a single national data center or may result from collaborations among two or more groups. Collaborative efforts might target different sub-regions or contribute different areas of expertise. Argo national programs (and institutions) interested in forming or participating in regional data centers are listed below. The first one listed for each ocean is designated as the lead institution, to work with the others in developing a regional data center.
The above list from the AST-4 Report is meant as a starting point for possible Argo specific needs. An overarching Argo objective is to encourage the rapid expoitation and broad use of Argo data. National institutes will have their own operational and scientific needs and responsibilities. A purpose for the present Workshop is to identify the specific institutional needs and plans for data in the Indian Ocean region that impact on Argo needs and on broader ocean climate research needs, and to develop a draft plan for the Indian Ocean "linked, regional data center" concept.
INCOIS acquires both in-situ (moored buoy, drifter buoy, XBT, current meter moorings, Argo floats, etc) and satellite data (SST, winds, ocean color parameters such as chlorophyll, etc) from all available sources to generate value-added data bases for the Indian Ocean. A dedicated dynamic web site and ocean portal with capability of data warehousing and data mining is under construction. Users will be able to access these data via the internet.
To start with, INCOIS plans to generate digital atlases, a monograph on the response of the tropical Indian Ocean to the meteorological forcings, and studies on the heat and salt budgets of the near-surface layers of the tropical Indian ocean on intraseasonal and interannnual time scales. The available algorithms on the heat and salt budgets of the near-surface layers would be carefully evaluated with these data bases to study the co-evolution of the near-surface thermohaline structure of the tropical Indian Ocean in all scales.
INCOIS plans to generate value-added data bases and digital atlases of the parameters such as surface wind stress and its curl, wind mixing, solar radiation, chlorophyll, suspended sediment, net heat flux, heat storage, rainfall, river discharge, sea surface temperature, sea surface salinity, mixed layer depth, thermocline depth and gradient, near surface circulation and sea surface height anomaly on intraseasonal and interannual time scales on CDs in digital, analysed and animation forms.
For the regional Argo data center activity for the Indian Ocean region, INCOIS has already gotten the PMEL delayed mode quality control software. In order to test this software in the Indian Ocean region, INCOIS plans to choose some test region, say a 10 degree square grid, and test how well the PMEL software functions in this region.
For preparing historical data sets, INCOIS already acquires all the available historic data sets. The data sets include oceanographic data collected by ships (research, commercial, fishery, and naval) in addition to WOCE and other climatological data sets.
During 2001 the International Pacific Research Center (IPRC) at the University of Hawaii established the Asia-Pacific Data-Research Center (APDRC). The IPRC's research effort requires easy access to the substantial atmospheric, oceanographic, air-sea flux and satellite data sets and products needed for climate research. The APDRC was established to provide this function. The vision of the APDRC is to link data management and preparation activities to research activities within a single center, and to provide one-stop shopping of climate data and products to local researchers and collaborators, the national climate research community, and the general public. Accordingly, the APDRC Mission Statement is: to increase understanding of climate variability in the Asia-Pacific region by developing the computational, data-management, and networking infrastructure necessary to make data resources readily accessible and usable by researchers, and by undertaking data-intensive research activities that will both advance knowledge and lead to improvements in data preparation and data products. The APDRC has four main activities: Data Server System (DSS) implementation and development; data management and archive building; value-added, data-intensive research projects; and coordination and collaboration.
The DSS is presently serving both gridded and in situ data. The Live Access Server (LAS) provides both restricted and unrestricted access to gridded products; the EPIC server provides access to in situ data and products. Both servers have been developed by our collaborators at NOAA/PMEL. Upgraded capabilities are being developed. The LAS allows users to visualize and download gridded products in various formats. Most of the datasets served by the LAS are also served through our DODS/OPeNDAP server, which allows users to directly access the data using their desktop client tools. The EPIC server is also a web-based, data search and display system suited for in situ (station versus gridded) data. The process of locating and selecting individual station data from large collections (millions of profiles or time series, etc.) of in situ data is a major challenge. We enhanced the capability of EPIC and configured OPeNDAP into EPIC to serve the numerous in situ data files and to export them to users.
In addition, the APDRC has an OPeNDAP (formerly DODS, Distributed Oceanographic Data System) Catalog/Aggregation Server (CAS) developed by Unidata at UCAR to serve climate data and products such as model output and satellite-derived products. These products are often large (> 2 GB) and are therefore stored as multiple files (stored separately in time or in parameters). The CAS remedies the inconvenience of multiple files and allows access to the whole dataset (or any subset that cuts across the multiple files) via a single request command from any DODS enabled client software. Any DODS (OPeNDAP) client (GrADS, Ferret, and ncBrowse) should be able to access this server.
The process of archive-building is underway and will be a central activity for the immediate future. The delayed-mode Argo data, WOCE data, and historical in situ data are top priorities. Providing user-friendly, web access to the WOCE Version 3 data is a top priority. The APDRC is currently serving the WOCE current meter DAC data via EPIC. In addition, EPIC is serving the (early Version 3) CTD and bottle data, and some "test" Argo float data. The CAS is serving (early Version 3) SST, satellite winds, and satellite sea level products. We plan to serve the final WOCE Version 3 data via DODS/OPeNDAP, EPIC, LAS and CAS servers as appropriate when the final data are public. In addition to the WOCE data, the APDRC will continue to serve global climate data and products in support of CLIVAR research activities.
Specific data management activities and value-added products are being identified in collaboration with NOAA/GFDL, GODAE, CLIVAR, Argo, DecVar, CSIRO and others.
The APDRC has recently downloaded the PMEL software for the delayed-mode quality control of Argo data (Wong et al, 2001). The software is now up and running using the climatology provided with the software. The APDRC plans to test the software in both the western Pacific and the Indian Ocean. We have downloaded some Argo data from the Pacific and Indian Oceans from the Argo GDACs, which are now supplying data. We plan to test the PMEL quality control procedures regionally, compare our results with other centers, and load the Argo data to our local EPIC server in order to provide value-added Argo regional products. A central element of providing regional quality control of Argo data is the development of improved historical data bases and climatologies. We plan to work with other institutions in the development of these global, improved-quality-control data bases along the lines on Macdonald et al. (2001).
At present we are entering into an agreement with CSIRO for the completing of a quality controled UOT data base for the Indian Ocean over the next two years.
The APDRC web site is at (http://apdrc.soest.hawaii.edu).
The Joint Australian Facility for Ocean Observing Systems (JAFOOS) conducts a broad range of activities including data assembly, analysis and modelling activities. JAFOOS is a joint effort between CSIRO and the Bureau of Meteorology Research Centre (BMRC). Specific activities are the scientific quality control, archiving and distribution of Upper Ocean Thermal (UOT) data; the development of data analysis and management systems, including data processing and quality control tools; and the production of routine data analyses and products which are available at http://www.marine.csiro.au/JAFOOS. JAFOOS participates in the Virtual Ocean Data Hub (VODHub) project using the Distributed Ocean Data System (DODS).
A current project is the Indian Ocean climatology project which aims to develop a new, fine resolution Indian Ocean climatology to investigate: seasonal cycle of storage and advection in the upper ocean; cycles of equatorial and boundary currents; and interannual and decadel changes. The climatology will be developed from the Indian Ocean Temperature Archive (IOTA). IOTA draws on all available UOT data (current, historical and previously unavailable data), and includes a semi-automated quality control procedure developed to reduce the number of profiles requiring visual inspection. Over the next two years, CSIRO in collaboration with the APDRC plans to complete the quality control of the Indian Ocean UOT data set.
In addition to developing quality controlled data bases for the Indian Ocean, it is also important to develop mapped or gridded products which can be easily used for analysis and model evaluation. Examples of specific techniques, and the enhanced value and spatial resolution of the resulting gridded products compared to standard climatologies, can be found in Dunn and Ridgway (2002) and Ridgway et al. (2001) for the waters around Australia.
CSIRO is participating in the deployment of Argo floats in the Indian Ocean, and may be able to help with the delayed-mode quality control of Argo data depending on the availability of resources. The focus of a CSIRO contribution to a regional Argo center would likely be in the Southern Ocean and the Indian Ocean region near Australia.
This paper has provided an overview of Argo needs for a regional data center and of ongoing data center activities which could contribute to the linked regional data center concept. An underlying assumption is that no single country has the necessary resources (funds) and expertise (manpower) to singlehandedly provide the required functions of an Argo (or climate research) regional center for the entire Indian Ocean domain. Especially, regional expertise is required for the regional quality control of data and the evaluation of data-based products and model products.
The task for the Workshop is to define a linked regional data center for Argo and, more generally, for climate research, to identify possible participants, and to identify specific tasks/timetables which the participating centers plan to achieve. The specific Argo tasks follow.
A possible strategy is for the contributing centers to conduct the above activities in sub-regions of the Indian Ocean. Another option is for centers to take on specific tasks for the whole Indian Ocean. Tasks which support the above Argo activities, but have more general application are:
Specific issues which need to be addressed include: ongoing coordination and communication mechanisms; infrastructure sharing and development; possible resource sharing, and others.
The Argo Science Team has identified desirable tasks for a regional data center for Argo. The IOGOOS Workshop provides an opportunity to plan a linked regional center for the Indian Ocean. The results from the Workshop should be made available to the Argo Science Team at their next scheduled meeting in China, 4-6 March 2003.