Nine Months of Seismic Network Records of Munition Disposal Reveal Multi-Scale Atmospheric Variability Controls on Near-Ground Explosion Monitoring Capabilities

Abstract:

Dense seismic networks routinely record infrasonic signatures of increasingly frequent near-ground explosions that locate in geographical regions, but that lack infrasound sensor network coverage. Development of transportable methods to analyze these sources remains important for global security applications, particularly in regions that lack infrasound assets, but where seismo-acoustic waveforms sample an atmosphere with multi-scale variability. To facilitate the development of such methods, we assemble nine-months of seismic data that record routine munition disposal operations conducted by the McAlester Army Ammunition Plant (McAAP) in Oklahoma state. The seismic stations that provide these data distribute over four neighboring US states and record sequences of repetitive, quasi-similar signals almost daily. These signals record small (100s of kg) to moderate (several ton) yield sources in the 5-15Hz band as wavetrains of short duration pulses (~1-3s widths) that separate by 20s intervals, and move near acoustic propagation speeds over regions instrumented by ~150 seismic sensors. Data collected from as far as 670km from source often recorded 52 such pulses per day, and over durations of 1200s. We apply noise-adaptive, constant-false alarm rate power detectors against these data to construct detection maps (see Figure) that reveal multi-scale spatial and temporal atmospheric variability over several seasons (note panels in Figure map to seasonal wind directions). We compare these data with infraGA modeling output to show that range-dependent raypath focusing and mechanical coupling significantly controls spatial detection variability. Our detection maps further reveal that concurrent, tropospheric and seasonally dependent stratospheric winds most reliably predict temporal detection variability. To conclude, we deliver our blasting records, seismic waveforms, videos, and processed data products as a comprehensive, ground-truth dataset to facilitate future atmospheric and seismo-acoustic studies.