This database of oceanographic time-series data is managed by an oceanographer at the USGS/WHSC. Historically, data stewards who work directly with the data have been trained in use of the processing software to assure consistency in data treatment and the process documentation. The data-processing and quality-control steps required to produce a BBV are usually adequate to assure the data have been properly handled. However, the database manager reviews the data before publication to ensure completeness and accuracy. Verification of the metadata associated with the various files is also performed to ensure the data are associated with the correct time and location.
Several backup methods are employed to assure the continuity of access to this data set. The primary method is bi-weekly local backups of the server. The backup media are stored at onsite and offsite locations. The computer server disks are also mirrored, so there are at least two "live" copies of the data files. The server is on an uninterruptible power supply with sufficient capacity to run the system for 30 minutes, with a generator providing backup power.
Some investigators publish a data report (a USGS Open-File Report or Digital Data Series) to describe the experiment and instrumentation, present plots of the data collected, and distribute the time-series data (see Appendix 3). These reports provide an additional backup of the data files. These reports may be updated (version 1.0, 2.0, etc.) if data are modified or corrected following initial publication. The data served by this online database is always the latest version, incorporating any necessary modifications. Periodic reviews are conducted by the database manager to assure that the files and links are current.
As new data are incorporated in the database, updates are sent to the several aggregation sites, such as the Geospatial One-Stop (GOS) and the Global Change Master Directory (GCMD). These sites function as libraries where a user can find data on a topic from a number of sources. Having the links to CMGP data on these sites is helpful for discoverability but because the data are measurements of one or more variables over time at a single location, it is not very well suited to the geospatial focus of many of the sites. Evaluation of new data clearinghouse sites is ongoing, and as suitable sites are identified, discussions regarding pointing to the time series data are initiated. USGS is currently (2008) working with NOAA to have this database accessed from one of their web sites that host historical data collections.
|