Institute Of Geophysics, Ruhr-University Bochum, Germany

The design of a global seismic system to monitor compliance with a ban on underground nuclear testing considerably deviates from previous concepts of international seismic data exchange. The new concept relies on centralized processing of continuous data from a fixed station network (“alpha” stations) which provides the primary detection and location capability. This alpha station network is augmented by additional stations (“beta” stations) which send data on request to refine the hypocentres of events which were detected by the alpha network. To test this concept we have used the GERESS array in Germany as a prototype alpha station and investigated its regional detection and location capability for events in France and surrounding areas. For this region, data from the national French network operated by LDG provide an excellent reference data base. Within a 5 degree distance, GERESS showed an excellent performance in terms of detection and location down to magnitude M(LDG) = 3. Between a 5 degree and 10 degree distance, the detection capability is still high but very often it is not sufficient to locate events below M(LDG) = 4. Generalizing these results, we can conclude that either the maximum distance between alpha stations should be 10 degrees or the contribution of beta stations has to play a significant role in a future monitoring system.

The technical concept of a future global seismic monitoring system includes 50 to 60 core stations, mostly arrays, which provide the primary detection and location capability. Due to the average station distance, these core (“alpha”) stations form a teleseismic network. Many of the proposed stations are to be newly installed and before the network can be regarded as fully operational, the stations have to be calibrated. As for traditional seismic networks, the station residuals - compared to a standard earth model - have to be determined. The standard earth model is defined in terms of travel-time tables and amplitude-distance curves. After recording a representative set of events, station residuals with respect to travel-time and magnitude can be calculated. In case of arrays, the determination of mislocation vectors (azimuth and slowness residuaIs) are of ulmost importance if array slowness vectors are used as starting solutions in a location procedure. Finally, in a monitoring context it is very important to estimate the station sensitivity for varying background noise conditíons and - in case of arrays - to know the frequency dependent improvement by beamforming. This paper uses the newly installed high-frequency GERESS array in Germany to demonstrate the calibration procedure.