ASDP: a PC-based program using a multi-algorithm approach for automatic detection and location of local earthquakes

Abstract

A few automated data acquisition and processing systems operate on mainframes, some are run on UNIX-based workstations and others on personal computer, equipped with either DOS or UNIX-derived operating systems. Several large and complex software packages for automatic and interactive analysis of seismic data have been developed in recent years mainly for UNIX-based systems, and some of these programs use a variety of artificial intelligence techniques. Here, the first operational version of a new software package, named PC-Seism, for analyzing seismic data from a local network is presented. This package, composed of three separate modules, provides an example of a new generation of visual object-oriented programs for interactive and automatic seismic data processing run on a personal computer. In particular, we discuss the automatic procedures implemented in the ASDP (Automatic Seismic Data Processing) module. A multi-algorithm approach to the on-line detection and location of local earthquakes is adopted in ASDP, and its operative mode is similar to that used in more complex systems, where the algorithms run on different processors and parallel computations are generally performed. Since highly complex computation routines may still be prohibitive for current PC when the number of analyzing traces becomes large, we have opted for simplicity and have planned three main routines (working in cascade mode) and a multi-station analysis (MSA) procedure in ASDP, to recognize phase picking, declare and locate earthquakes. Basically, signal detection on a single-component trace is obtained by a short-term average to long-term average ratio (STA/LTA) taken along a characteristic function (CF) envelope generated from the seismogram. To confirm and identify earthquake phase arrivals and to discard noise disturbances, two other sections of analysis are applied on short signal windows after the declared triggers. A spectral analysis is applied as detector of earthquake phase arrivals. Furthermore, for three-component data, the application of a polarization filter is used to identify P- and S-phases. Therefore in order to optimize the efficiency of a seismic monitoring system, the software is provided with full detection capability and makes use of a high degree of automation, aimed at minimizing the analyst's workload. PC-Seism and, in particular, the ASDP module are under constant review. In fact, the modular structure of the program will allow us to develop it with new and improved capabilities. Future developments improvements in ASDP will concern the application of other detection algorithms and routines to recognize secondary phases in regional and teleseismic earthquakes.