The Seismic Research Observatory (SRO) network is generating an important new data base for seismological research. The SRO systems have extended both the range and resolution of seismic measurements beyond the limits of conventional seismographs and the data are recorded in digital format making it convenient to automate processing and analysis. The SRO network now comprises 12 stations and one additional station will be installed. The data produced by the SRO network, and other digital systems being installed to supplement the SRO network, will be collected by the U.S. Geological Survey's Albuquerque Seismological Laboratory, organized into network-day tapes, and made available to research organizations throughout the world. The network-day tapes are particularly unique and useful data files as they will contain not only seismic data but all of the information needed by an analyst to interpret the data, including station constants and calibration and other instrumental parameters. Hoffman (in preperation) describe these tapes and the format used in their compilation.

The global digital data base for seismology is still relatively limited in terms of both geographical and temporal coverage. For this reason there is a tendency, perhaps, to stretch the applicability of the available data beyond the intended purposes. The danger in this is compounded by the fact that computer processing and analysis often can be done without even looking at the waveforms, and it is done with deceptively high computational precision that belies the basic accuracy of the measurements. The day may be passing when seismologists have a hands-on familiarity with the data acquisition systems, but clearly it is as important as ever that they have a knowledge of the capabilities and limitations of the instruments.

The primary objective of the SRO network was to lower the threshold of event detection in the long-period band. This was achieved by using a newly-developed borehole seismometer that is capable of resolving both vertical and horizontal components of earth background noise at the quietest sites on earth. Another objective was to accommodate a large range of signla amplitudes. This was achieved by using an advanced digital recording systems which provides at least 110 dB of separation between system noise levels and clipping levels in the principal bands of interest. The broadband capability was not an initial program requirement and is not being fully utilized at present, but it adds significantly to the research potential of the network. The SRO systems have proven to be reliable in operation and the general quality of the data is excellent.

Every seismograph has functional limits of accuracy, and the SRO system is no exception. The chief purpose of this report is to define these limits for the SRO system and provide the most accurate calibration data currently available. At the same time we will describe the testing program in sufficient detail so that the data user can make his own judgments regarding the effectiveness and accuracy of the tests. This report is not a final statement; further work is needed in some areas to completely define the SRO system, in some cases purely for academic reasons. Refinements in calibration will appear on the network-day tapes as the new data become available.