The Complementary Nature of Seismic and Infrasound Technologies in Regional Monitoring (Invited)

Under current CTBTO event detection and location operating conditions, signal detection is a station-centric decision (was an event phase detected at this station?), rather than a global hypothesis test. Currently, infrasound and seismic detection use signal detectors run independently on each technology. It is only after event formation that the observations and inferences are merged. Development of this independent processing is a result of the vastly different signal and noise characteristics of these two waveform technologies. However, for specific signals there may be a utility to a joint seismic-infrasound detector. For example, noise estimates from one technology may help characterize or identify the noise on another technology (wind couples to both infrasound and seismic). Back-projection methods for both seismic and infrasound could easily be combined to produce a common seismo-acoustic detection and associated event location. The opportunity exists to integrate detection and location into a single multi-disciplinary approach. One such example is the ongoing infrasound detection and location procedure that utilizes an adaptive F-detector as input into the Bayesian Infrasonic Source Location (BISL, Modrak et al. 2010) procedure that provides an estimate of source location using assigned prior probabilities based on what is known of the propagation path and on the signal detector estimates (arrival time, phase velocity and azimuth). As the atmospheric model is better defined these priors may be changed, thus linking improved location estimates directly to improvements in atmospheric models. The final step following event location is identification. Seismic and infrasound observations and their interpretation for the recent set of North Korean nuclear explosions in 2006, 2009, and 2013 provide a motivation for multiple disciplinary approach to this step as well. Seismic analysis of these tests have documented that for existing parameterized source models, there are trade-offs between yield and depth (Mueller-Murphy, 1971, Koper et al., 2008, Chun et al., 2011, Murphy et al., 2013, and Park, 2013). An approach to integrating seismic and infrasound observations and models to constrain near-surface sources offers an opportunity to explore these events more fully. The procedure builds on regional and local seismic source models through moment tensors and uses these results to estimate ground motions directly above the source that can then be coupled to an atmospheric propagation code for investigating the complementary infrasound observations. The coupling from the seismic to the infrasound wavefield is done via the Rayleigh integral and the subsequent wave propagation can exercise any one of the existing atmospheric models. As documented in Arrowsmith et al., 2012, the technique has been successfully applied to the analyses and modeling of the seismic and infrasound data associated with the Circleville, Utah magnitude 4.3 earthquake on 3 January 2011. This exercise illustrates the importance of source mechanism, source depth and surface geology on strength of the subsequent infrasound signal as well as the importance of the atmospheric model at the time of the earthquake.