Further Comments on Databases

Susan Y. Schwartz, University of California, Santa Cruz,
Chair, Data Management System Standing Committee

Julian and Foulger's article "Databases in the Field: A Dissenting View" is more than a criticism of a software package designed to assist PASSCAL investigators with data management in the field; it is also a condemnation of IRIS´s policy of open access to seismological data. IRIS´s enlightened attitude concerning the importance of open data exchange is responsible for its success in acquiring the impressive resources used to collect seismological data on both a global (GSN) and regional (PASSCAL) scale. Several other, less-generous, scientific communities still struggle with the concept of open data exchange and their attitudes have negatively impacted their ability to garner the support and funding required for significant advances in their science. Vigilance in maintaining open access to seismological data has resulted in the PASSCAL program's priority of ensuring timely delivery of data to the Data Management Center (DMC) for incorporation into the data archive as well as supporting experiments and investigators in the field. Data as a community resource was a design criteria of the PASSCAL program, is the hallmark of IRIS, and is non-negotiable.

Although PASSCAL data are in general of more variable quality than GSN data, it is not uncommon for PASSCAL broadband stations to have uptimes and noise levels that equal or exceed the GSN standards. Thus the importance of archiving PASSCAL broadband data should not be underestimated. Contrary to Julian and Foulger's characterization of PASSCAL temporary stations as "a rag-bag of short term field experiments", data from PASSCAL broadband stations archived in SEED format are a rich resource that over the last year has accounted for 10% of all SEED formatted data requests to the IRIS DMC. These approximate 1000 requests for SEED formatted PASSCAL data contrast with 43 requests for variably formatted assembled PASSCAL data sets in the same year. In addition, PASSCAL data have and will continue to provide unique, spectacular recordings of important earthquakes (1994 Fiji and Bolivian events) that would not exist if not for a fortuitously located PASSCAL experiment. PASSCAL experiments are funded based on the importance of the seismological data they are designed to collect. This importance, initially recognized by an individual PI is validated by the larger community in the proposal review process. Thus there should be little mystery about the significance of PASSCAL data to the seismological community and the increasing proportion of PASSCAL data requested from the DMC attests to its utility.

There is little question that PASSCAL data are important and of use to the seismological community. There is also little question that the DMC has a mandate and is well-suited to archive this important data. PASSCAL broadband data is archived in a similar manner as GSN data and all request mechanisms and data manipulation tools available for GSN data are applicable to PASSCAL data as well. This allows easy and seamless access to the data by the seismological community and also provides convenient data processing tools and large disk space to the PIs during the proprietary period. This year, ten PASSCAL experiments presently within the proprietary data period have used DMC tools and large data buffers to perform event, station and waveform extractions on their large data sets archived in SEED format at the DMC. The minority opinion expressed by Julian and Foulger regarding the importance of PASSCAL data standardization harkens back to the days of permanently proprietary data collection efforts that repeatedly under utilized the full potential of the seismic signals. The community will not return to those dark ages.


Return to: IRIS Newsletter Information
Return to: Title Page and Table of Contents
Continue to: Next Article