A DATA INFRASTRUCTURE TO SUPPORT EARTH HISTORY RESEARCH

One of the primary goals of the CHRONOS project is to bring together distributed repositories of earth history data and make them available to scientists in a coherent and consistent way. A software infrastructure created to achieve this goal includes a federated database to provide unified global-views of the diverse databases and the CHRONOS Query Description Format (QDF) service. Applications built on top of this infrastructure are supplied with a uniform access mechanism that can integrate the distributed data.

The foundation of the infrastructure is the data itself. At present the databases that are part of the CHRONOS federated system include Neptune, PaleoBiology, PaleoStrat, Janus, Timescale, MIOMAP and FAUNMAP. The federated database is built using IBM's DB2 Information Integrator which can process queries and send subqueries to the various database types, which in this case includes Oracle, PostgreSQL, MySQL and MS SQL Server. Most of the databases were developed independently with different schema and different naming systems. Information Integrator permits the construction of global-views which have a common unified schema supported by unique SQL queries to the individual databases. To support discovery and meta-data searches, these views are registered with GEON.

In conjunction with developing the infrastructure necessary to deliver integrated access to multiple geochronological and stratigraphic databases, we are working on developing tools to provide end users with intuitive access to those data resources along with data analysis capabilities. Age-Depth Plot (ADP) provides a GIS interface to display datum ages versus depth for a selected core hole, providing a palette of tools for fitting an age model to the data or modifying an existing age model. ADP interacts with the database through the CHRONOS Query Description Format (QDF) service, providing for greater security and robustness of the data connection. ConDoor is a graphical front end for the CHRONOS web service implementation of CONOP9, a program that uses constrained optimization to develop sequences of stratigraphic events that provide optimal simultaneous correlation of observed records in multiple sections. Input data may be extracted from the federated database that presents an integrated view of the data.