Test and calibration of the Digital World-Wide Standardized Seismograph

Abstract

During the past decade there has been steady progress in the modernization of the global seismograph network operated by the U.S. Geological Survey (USGS). The World-Wide Standardized Seismograph Network (WWSSN) has been augmented by new stations with advanced instrumentation, including the Seismic Research Observatories (SRO) and the modified High-Gain Long-Period (ASRO) stations. One goal in the modernization effort has been to improve signal resolution in the long-period band. A second goal has been to generate a global digital data base to support contemporary computer-based analysis and research. In 1976, a Panel on Seismograph Networks was established by the Committee on Seismology of the National Academy of Sciences to review progress in network seismology and recommend actions that would lead to an improved global data base for seismology. One recommendation in the Panel report (Engdahl, 1977) called for upgrading selected WWSSN stations by the installation of digital recorders. This was viewed as an economical way of expanding the digital network, which had proven itself to be a very promising new tool for earthquake and explosion research. Funds for the development and assembly of 15 digital recorders were provided to the USGS by the Defense Advanced Research Projects Agency and an ad Inoc panel of scientists was convened by the Committee on Seismology to advise the USGS on the selection of stations to be upgraded and on data recording requirements, A total of 19 digital World-Wide Standardized Seismograph (DWWSS) systems will be operational when all are installed. The additional systems were made available through purchase by the USGS and other organizations; for example, the University of Bergen purchased and installed a DWWSS-type recorder and agreed to furnish the USGS with the data. A list of operational and planned DWWSS network stations is given in Table 1.1. As one might expect, the digital recorder turned out to be somewhat more sophisticated than the original concept. It was decided to record three components of long-period data continuously, three components of intermediateperiod data in an event mode, and the vertical-component short-period data in and event mode (with the capability of adding short-period horizontal channels in the future). Special amplifiers were developed for use with the WWSS seismometers, and a 16-bit fixed-point analog-to-digital converter was chosen to provide increased resolution (as opposed to a 16-bit gain-ranged encoder). The microprocessor-based digital recording systems were developed and assembled at the USGS Albuquerque Seismological Laboratory (ASL) and ASL-based technicians began installation at WWSSN stations in 1980. The current and proposed locations of the DWWSSN stations, together with other stations in the Global Digital Seismograph Network (GDSN), are shown on the map in Figure 1.1. A system was operated at Albuquerque for about 18 months, serving as a test bed for evaluation studies. Although the network hardware has been available for some time, the installation of the DWWSSN has proceeded slowly. The National Science Foundation supported installation of six stations and the USGS is funding installation of most of the others; however, the network completion date is conjectural because of funding uncertainties. The DWWSSN stations are supported with supplies and technical assistance from ASL (subject to availability of funds). Data recorded on magnetic tapes are mailed to ASL where they are reviewed for quality, then merged with other GDSN station data on the network-day tapes. Hoffman (1980) provides a description of the network-day tape format. Zirbes and Buland (1981) have developed and published user software for reading and interpreting the day tapes. This report will serve several purposes. One is to provide nominal system transfer functions and calibration information that are needed in the analysis of DWWSSN data. A second purpose is to report on an evaluation of operating characteristics (calibration stability, noise levels, and linearity) that may limit the usefulness of the data and to determine if modifications may be needed to improve the data. It is not an exhaustive study in this respect. We continue to depend mostly on data user feedback to point out deficiencies and we solicit comments whenever anomalies are observed in the data.