Using surface waves recorded by a large mesh of three-element arrays to detect and locate disparate seismic sources

Abstract

Surface waves recorded by global arrays have proven useful for locating tectonic earthquakes and in detecting slip events depleted in high frequency, such as glacial quakes. We develop a novel method using an aggregation of small-to continental-scale arrays to detect and locate seismic sources with Rayleigh waves at 20-50 s period. The proposed method is a hybrid approach including first dividing a large aperture aggregate array into Delaunay triangular subarrays for beamforming, and then using the resolved surface wave propagation directions and arrival times from the subarrays as data to formulate an inverse problem to locate the seismic sources and their origin times. The approach harnesses surface wave coherence and maximizes resolution of detections by combining measurements from stations spanning the whole U.S. continent. We tested the method with earthquakes, glacial quakes and landslides. The results show that the method can effectively resolve earthquakes as small as ∼M3 and exotic slip events in Greenland. We find that the resolution of the locations is non-uniform with respect to azimuth, and decays with increasing distance between the source and the array when no calibration events are available. The approach has a few advantages: the method is insensitive to seismic event type, it does not require a velocity model to locate seismic sources, and it is computationally efficient. The method can be adapted to real-time applications and can help in identifying new classes of seismic sources.

Stormquakes

“…The isolated Rayleigh wave signals are prominent and can be easily identified from the filtered seismographs (Figure c). They are excited by sources as strong as typical magnitude 3.5 or larger earthquakes in this period band (Figure S6; Fan et al, ). For instance, surface waves excited offshore New England around 23 August 11:27:06 were clearly recorded by stations 4,000 km away (Figure a).…”

Multiple Small Scale Landslides Triggered by Typhoon Talas 2011

“…Following (de Groot-Hedlin and Hedlin, 2015) and (Fan et al, 2018), we first divide the 103 stations into non-overlapping 68 triangular subarrays (triads) via Delaunay triangulation ( Fig. 1a) (Lee andSchachter, 1980, Thompson andShure, 2016).…”

“…For each triad, we measure relative travel times between station pairs to solve for a centroid arrival time and a propagation direction if the signals are coherent across the triad (average cross-correlation coefficient ≥ 0.5). We then invert the seismic source locations with aggregations of propagation directions and arrival times by grid-searching for possible source locations (Fan et al, 2018). To neutralize off-great-circle path propagation effects, we also apply empirical calibrations from detections of earthquakes in the Global Centroid Moment Tensor (GCMT) project (Dziewonski et al, 1981, Ekström et al, 2012 and landslides reported in a previous study (Yamada et al, 2012).…”

“…Here we apply a surface-wave detector that is based on the AELUMA method (Automated Event Location Using a Mesh of Arrays) (de Hedlin, 2015, Fan et al, 2018) to investigate landslide activities across Japan during the transit of Typhoon Talas. Our approach has been successfully applied to the USArray with over 400 stations and located various unconventional seismic sources, including glacial quakes, stormquakes and submarine landslides (Fan et al, 2018(Fan et al, , 2019(Fan et al, , 2020. In this study, we use to 50 s period Rayleigh waves from ∼40 stations to form 29 sparse triangular subarrays (triads), and locate three landslides, including a landslide in Tenryu, Shizuoka prefecture, which was over 400 km away from the track of Typhoon Talas.…”

Remote Detection and Location of Explosive Volcanism in Alaska With the EarthScope Transportable Array

“…Regional‐scale studies using seismoacoustic TA data have previously considered nonvolcanic events in the contiguous United States, with signals ranging from short and impulsive (e.g., de Groot‐Hedlin & Hedlin, ; Edwards et al, ; Walker et al, ), to long and emergent (e.g., de Groot‐Hedlin et al, ; Fan et al, ). Infrasound produced by explosive volcanism can be equally complex, from minutes to days in duration, with impulsive or emergent signals of time‐varying frequency (e.g., Fee & Matoza, ; Johnson & Ripepe, ; Lees et al, ; Matoza et al, ; Matoza et al, ; Petersen et al, ; Ruiz et al, ).…”