TWENTY-FOUR BITS WIDE; TWENTY TERABYTES DEEP

Since 1985, the IRIS Consortium has developed and operated a Data Management System for the archive and distribution of seismological data from the permanent stations of the Global Seismographic Network and portable instruments of the Program for Array Seismic Studies of the Continental Lithosphere (PASSCAL). With high dynamic range (24 bits), low noise and high sampling rate (> 20 samples per second), these data cover the full range of signals used by seismologists to study earthquakes and Earth's interior. With continuous recording at GSN stations and many of the PASSCAL sites, the archive has grown to 20 terabytes. A high-level database management system (DBMS) allows researchers to use a variety of tools and criteria to interactively select any subset of the data in the archive.

While the size of the IRIS archive is large, the primary data type (time-series of ground motion measurements) is relatively pure and simple. Even with associated metadata (instrument characteristics) and ancillary information (earthquake locations) the full data management system remains complex but straightforward. As the IRIS archive has grown, the seismological research community has made increasing use of the tools provided by a rich DBMS and other information technologies. This experience points to the need to develop a richer base of data from other geoscience disciplines.

The development of IRIS facilities, including the Data Management System, also provides experience on the care and feeding of community based resources. As NSF-funded facilities, the IRIS programs are guided by, and open to, the entire research community. The operations of the facilities involve a dedicated professional staff, with guidance and oversight from a community governance structure. This model for blending of talents and experience may be an appropriate starting point for the development of future applications in Earth science information systems.