The yields of the six declared underground nuclear tests at the North Korean test site are estimated using high‐frequency teleseismic P wave amplitude modeling and waveform equalization of short‐period teleseismic P waves and regional Pn signals. Average amplitudes of the first cycle of high‐frequency (>4 Hz) filtered P wave displacements for each event, adjusted for station sampling relative to the 3 September 2017 event, are modeled using Mueller‐Murphy explosion source models for granite and a constant‐Q attenuation operator with t* = 0.78 ± 0.03 s. The yield estimates range from 2.6 to 230 kt. Intercorrelation, a waveform equalization procedure that accounts for source function and depth‐phase variations between events, is applied to large sets of filtered (>0.8 Hz) teleseismic P and regional Pn seismograms. Searching over yield and burial depth for both events gives optimal parameters by simultaneous waveform equalization of multiple stations. Using specified burial depths spanning from 430 to 710 m for all events based on estimated locations in the source topography assuming tunneling with 4% grade, along with allowing for reduction in source region velocity due to weathering, rock layering, and damage zones, gives yield estimates ranging from 1.4 to 250 kt. Comparison of predicted and observed spectral ratios of Pn phases at station MDJ establishes that these source models are reasonable. Using the preferred yield estimates from intercorrelation, W IC, a yield‐calibrated relation of m bNEIC = 0.9 log10W IC + 4.13 is determined for the North Korean test site.

https://doi.org/10.1029/2019JB017418

Journal Article

Korea (North)

NOAA Cooperative Institutes