LTER Site Byte

LTER Site: Georgia Coastal Ecosystems

Contributor: Wade Sheldon (Jul 31, 2005)

Site Byte:

We are already busy preparing for our 2006 renewal proposal, so information management work at GCE has largely been focused on helping PIs with data synthesis projects, improving managment and access to GCE reprints and publications, and tightening integration between project databases through web application cross-references and content displays. The data synthesis efforts have been strongly aided by the development of the GCE Data Search Engine, a MATLAB application for metadata-based indexing, searching and integration of data sets stored as GCE data structures (http://gce-lter.marsci.uga.edu/lter/research/tools/toolbox_search.htm). This tool has been well recieved by PIs, and several use it to manage and analyze their own data in addition to GCE data sets. We also began using this indexing technology to generate detailed data summaries to enhance distribution of near-real-time and historic ancillary data on the GCE Data Portal web site (http://gce-lter.marsci.uga.edu/portal/). This is particularly useful beause we began providing full public access to the portal site this year to support our SLTER participants and UGA students as well as the broader LTER community. All portal data sets can be searched and retrieved alongside public data from the GCE Data Catalog using the GCE Data Search Engine client (i.e. via direct HTTP in MATLAB). All data access from the GCE Data Catalog and GCE Data Portal and all software downloads are now tracked in our data access database, requiring contact and affiliation information from downloaders, but we did add support for web browser cookies and stored logins in the Search Engine client to streamline the downloading process for regular data users.

We are holding off on implementing new data search applications on the GCE web site until the standardized LTER query interfaces in development are more mature, but we did significantly enhance the query capabilities of our existing data catalog interface this year (http://gce-lter.marsci.uga.edu/lter/asp/db/data_catalog.asp). Specifically, we added support for searching by LTER core area, searching by study dates, and also by text in individual or combined metadata sections. Metadata content searches are highly targeted and extremely fast compared to xml-based searches, and the searchable content can easily be tailored in the future by creating simple database views and adding them to the array of available metadata sections. These new query capabilities also allow search hyperlinks to be added anywhere on our web site to support improved cross-referencing between databases, such as searching for relevant data sets for individual species records in our taxonomic database (e.g. http://gce-lter.marsci.uga.edu/lter/asp/db/species_details.asp?id=Geukensia%20demissa)

This year we began acquiring and managing more full text reprints for GCE publications and we now have a searchable reprint archive linked to our bibliographic database on the private GCE web site for project participants. We also automatically generate downloadable reprint archives complete with HTML indices of all reprints (both with and without abstracts) to allow PIs and students to install up-to-date libraries of GCE publications on their local systems for research purposes. We also improved linkages between our bibliographic and personnel databases and began providing an automated reprint request service on the public web site for both GCE and UGA Marine Institute publications (http://gce-lter.marsci.uga.edu/lter/asp/db/biblio_query.asp). Reprint requests are emailed to first listed GCE-affiliated author and cc'd to the IM office, and if no authors are active GCE participants the request is directed to the IM or UGAMI librarian instead. Quite a few international researchers and students have been using this service to request reprints from journals that are not widely available overseas.

We continue to provide complete Level 5 EML metadata for all data sets added to the GCE Data Catalog, as well as customizable versions of species lists in EML format from the GCE taxonomic database (http://gce-lter.marsci.uga.edu/lter/asp/db/all_species_lists.asp). We improved our EML implementation to better match the current Best Practices recommendations this year (e.g. reorganized geographic coverage content to describe the overall bounding box under eml/dataset/coverage, with individual site descriptions moved to the eml/dataset/methods/sampling/studyExtent tree, as well as inclusion of standard "place" keywords). All dynamically-generated EML documents are harvested weekly for inclusion in the LNO and KNB metacats and also the NBII metadata clearinghouse.

On the IT front we experienced our first major hardware problem this year when a SCSI hard drive in our database server's RAID-5 array failed and an auxiliary drive developed some bad sectors. The RAID did its job and kept the system up until new drives could be obtained, but that process turned out to be very complicated (and expensive) due to the age of the drives, taking over six weeks. After this experience we decided it was time to acquire a new database server, so I was able to purchase a PowerEdge 2800 with dual Xeon cpus and a faster RAID-5 array and re-task the older server as a development box. We also tightened security for accessing all servers by generating SSL certificates and requiring HTTPS (or SSH/SFTP) for all authentication. Less secure transport protocols (e.g. FTP, WebDav) are now disabled on all our systems and all machines are protected by firewalls (in addition to the rule-based URL filter on the web server). This required some PI hand-holding, particularly installing and using SFTP software, but should help us minimize our "attack surface".

We also revised our project management structure this year by forming an executive committee to direct the project, and it was unanimously decided that the information manager should serve as a full voting member on this committee. This formal acknowledgement of the central role the IM program plays at our site is very encouraging, and should improve the already strong integration of our science and IM programs in long term planning. In addition to forming an executive committee we also established formal bylaws for GCE, and we developed a secure web application for voting on project referenda in accordance with these bylaws. This application allows our broadly distributed PIs to vote on important issues that arise between annual meetings and will also help us establish a formal record of major project decisions for posterity.

Web design by Wade Sheldon, Georgia Coastal Ecosystems LTER