Progress in long-period seismology has been considerable in the past few years, owing to the availability of digital data from well-calibrated worldwide instruments. The very long period International Deployment of Accelerometers (IDA) network [Agnew et al., 1976] has provided many new measurements concerning both earth structure and seismic source studies and demonstrated the usefulness of sparse global digital networks. The broadband Global Digital Seismographic Network (GDSN) network has given access to a large quantity of data whose exploitation can be readily automatized.
Both networks have their shortcomings, however, now expressed in the desire of many U.S. scientists to develop a new global digital network better adapted to present requirements of geophysical research. In the very long period domain (periods from about 100s to 1 hour), the IDA network only records the vertical component of ground motion, making information from horizontally excited modes of the earth unavailable. It also saturates the first Rayleigh wave trains from the largest earthquakes, causing a loss of data on direct source station paths, regrettable both for source and structure studies. The GDSN network suffers from some nonlinearity problems and, above all, the inadequacy of instrument responses for present needs of seismologists: The Seismic Research Observatories (SRO) network [Peterson and Orsini, 1976], which is the main constituent of GDSN, was designed mainly for discrimination between earthquakes and nuclear explosions.