LTER metadata systems

A half page description of the current models and systems of data and metadata management at LTER sites.


Most of the sites (15 of 26 sites) use databases to manage their metadata. Most of these use relational databases, but not all. Some of these sites (9 of the 26 sites) use a hybrid model, where part of the metadata may be stored in flat text files in an structured file system. There are 5 sites that use the file system to manage their metadata (after all, a file system can be used as a simple database).

The preferred databases are mysql and SQL server, while some use postgres and there are a few instances of Oracle in the network. Two sites use eXist and DB2 respectively, which are XML oriented databases, especially eXist. Strictly speaking, only eXist and postgres are open source; although mysql follows an open source model, it has components that are proprietary.

The scripting languages or code use to make the data available to the users vary wildly, Cold Fusion, PHP, Perl, CGI, HTML, XSLT, Matlab, Java/Jsp, VB, SAS, Excel, Ajax, Javascript are technologies used, in many instances is an amalgam of such languages.

Most of the network Ecological Metadata Standard (EML) documents are very rich, with attribute level content or level 5 according to the
EML Best Practices, however, a fair amount of QA/QC needs to be conducted for these documents to be successfully used in a machine driven synthesis architecture such as PASTA or Kepler.

As for the data, many sites use a mix of database and flat text files (whether CSV or tab separated ASCII files) to manage the data. To see details per site, follow the links!