Continuous seismic threshold monitoring

SUMMARY Continuous threshold monitoring is a technique for using a seismic network to monitor a geographical area continuously in time. The method provides, at a given confidence level, a continuous assessment of the upper magnitude limit of possible seismic events that might have occurred in the target area. Two approaches are presented in this paper. Site‐specific threshold monitoring: by ‘focusing’ a seismic network on a specific target site, continuous threshold monitoring of that site is achieved. We optimize the monitoring capability by tuning the frequency filters and array beams to known characteristics from previously recorded events at the site. We define the threshold trace for the network as the continuous time trace of computed upper magnitude limits of seismic events in the target area, at a 90 per cent confidence level. As an example, we have conducted a one‐week monitoring experiment of the northern Novaya Zemlya nuclear test site, using the Fennoscandian regional array network (NORESS, ARCESS, FINESA). We find that the threshold trace is below m b = 2.5 more than 99 per cent of the time. 34 peaks exceed m b = 2.5. All of these peaks correspond to interfering seismic events that have been independently located by a teleseismic or regional network. During the entire one‐week time period, the threshold trace exceeded m b = 2.5 only for 43 min. Regional threshold monitoring: this involves conducting site‐specific monitoring of a dense grid of geographical aiming points and requires the development of generic phase attenuation relationships for covering an extended geographical region. Using again the Fennoscandian regional array network, we illustrate the regional threshold monitoring approach by maps with colour contour displays. We demonstrate that the network thresholds in Fennoscandia and adjacent regions show strong regional dependence. The thresholds are below m b = 0.5 close to each array (<300 km distance) and range from m b = 2.0 to 2.5 in parts of the Norwegian Sea and Barents Sea. The thresholds also vary significantly under different background noise conditions, and an increase of about 1.0 m b units is observed during a large teleseismic earthquake. These regional threshold maps have advantages over standard network capability maps in being more accurate during time intervals when interfering seismic events occur. They can also more easily reflect special conditions such as particularly favourable source‐station propagation paths, and have the advantage of not being tied to specific event detection criteria. The paper concludes that continuous threshold monitoring offers a valuable supplement to traditional seismic techniques used in nuclear test ban monitoring. The method may also be useful for monitoring earthquake activity at low magnitudes for sites of special interest, as well as for monitoring earthquake aftershock sequences.