An evaluation of seismic decoupling and underground nuclear test monitoring using high‐frequency seismic data

Abstract

An effective solution to the problem of the detection and identification of low‐yield coupled and fully decoupled underground nuclear explosions appears available via use of high‐frequency seismic data ranging up to 30 or 40 Hz. In order to evaluate detection‐identification capabilities when using such data, it is necessary to estimate (1) spectral characteristics and relative amplitudes of both P and S waves from explosions and earthquakes over the frequency band from 5 to 40 Hz, (2) signal transmission characteristics over this band through pertinent types of earth structure, and (3) recording system and ground noise characteristics over this frequency band. In this study, each of these topics is considered in turn as they relate to detection and discrimination of the signals from low‐yield coupled and decoupled explosions in the regional and teleseismic distance ranges. Estimates of the capabilities of specific hypothetical networks to detect and identify (insofar as signal‐to‐noise ratio is an important factor in identification) explosions within the USSR are then considered. These estimates of signal detection capability provide the central focus for the study as they serve to translate diverse and rather complex sets of observational data and theory into concrete predictions of monitoring capability. Following the assessment of detection capabilities, the problem of identification of small events is considered, with particular emphasis on discrimination at regional distances where the network is calculated to provide signals of high signal‐to‐noise ratio. The principal results and conclusions of this study are as follows: (1) seismic system noise can be suppressed to levels well below ground noise at quiet sites up to frequencies at least as high as 30–40 Hz when using presently available hardware; (2) average amplitudes of high‐frequency noise in a variety of geological environments are very low and change little with time or season; (3) transmission of high‐frequency P and S wave signals in the regional distance range in stable continental areas and shields is nearly as efficient as at 1 Hz, with effective Q factors in shield areas being about 9000 and 4000 for high‐frequency Pn and Sn, respectively, while the effective Q for Pn waves in tectonic areas is about 1000; (4) a properly designed and deployed network of 25 simple three‐component nonarray stations internal to the USSR and 15 similar stations surrounding the USSR is predicted to be capable of multistation detection at high signal‐to‐noise ratio of fully decoupled 1‐kt explosions located at all potential decoupling sites within the USSR; (5) by inference from the quantitative agreement of empirical observations and theoretical predictions, when using lower‐frequency data over a great range of explosion yields, we conclude that procedures based on the use of both detectable P and S waves will serve to identify explosion‐generated seismic signals at least as small as those expected from a fully decoupled 1‐kt explosion.