The Physical Basis of the Explosion Source and Generation of Regional Seismic Phases

The objective of this project is to determine the source physics and corresponding generation and evolution of local and regional seismic waves from nuclear explosions. In particular, we want to explain the source of the explosion- generated Lg phase. In a previous project we identified the following contributing sources to Lg: surface reflected pS that is trapped in the crust, S*, scattered Rg, and shear waves directly generated by non-spherical source elements. Our goal now is to quantify the contribution of each to Lg under different source conditions. To this end we have performed work in several complementary areas. The Russian Institute for the Dynamics of the Geospheres (IDG) has provided yield and depth information for 11 Degelen and 4 Balapan explosions. They have digitized records on their near regional stations (50-100 km distance) for all four Balapan explosions and one of the Degelen explosions. They have also provided near source peak particle velocities, rise times, and positive pulse duration measurements from ten Degelen nuclear explosions. These explosions were all conducted in high velocity media (>5 km/sec P velocity) and therefore place important constraints on the Lg generation problem. The Balapan near regional data provides a contrast with existing near regional Degelen explosion data. As the source media are similar while the near source topographies differ, these data may be useful in distinguishing the role of topography in near source scattering. The new near field measurements and near regional records provide an opportunity to track differences between Balapan and Degelen regional records back to their source. In a closely related analysis, to assess whether the amount of shear waves generated is affected by source depth and/or scaled depth, we examine the regional phase amplitudes of 13 Degelen explosions with known yields and source depths. The events range from 20% to 50% underburied. Preliminary analysis shows similar log10 amplitude vs. log10 yield curves for the initial Pn, the entire P wavetrain, Sn, Lg, and Lg coda. The slope of those curves varies with frequency, ranging from approximately 0.84 at 0.6 Hz to 0.65 at 6 Hz. We perform nonlinear source calculations to complement these observations, aimed at determining constraints on the relative size of CLVD and explosion sources. We have also examined recordings of historical co-located decoupled and tamped explosions at Azgir in the former Soviet Union, utilizing all 3 components of data and focusing on differences in shear wave generation between the explosions. The relative S to P wave amplitude appears similar at low frequencies, but is much greater above 8 Hz for the tamped explosion, although interpretation is complicated by the frequency dependence of the decoupling. The tamped explosion also has clear Sn at regional distances. No similar records were available for the decoupled explosion. The tamped explosion, at 64 Kt and 987 m depth in salt, was 2 times overburied. Even if it were tamped, the 10 Kt decoupled explosion would have been nearly 4 times overburied. Wavenumber synthetic seismograms show that generation of the observed shear waves by a spherical explosion source is implausible, even in very complex source structure. We plan to perform nonlinear source calculations to assess the possibility of non-spherical source terms, at the surface and at an interface between the salt and overlying sediments. Finally, we model scattering of Rg into Lg for known source areas and compare resulting synthetics with regional data. We use an approximation based on modal scattering of fundamental mode Rg into higher mode Lg, using as constraints estimates of Rg decay rates from Degelen and deep seismic sounding explosions. The goal of this analysis is to quantify the contribution of Rg to Lg in different areas with different earth structures.