Near-source effects on global wavefields: A forensic seismology study applied to nuclear monitoring

This CASE project is supported by AWE Blacknest

Key Questions

What is the influence of topography and 3D geology on nuclear explosions, as seen from seismograms across the globe?

Background

As a consequence of the large-scale tectonic processes, the Earth’s crust is constantly undergoing deformation and thus comprises a variety of topological and heterogeneous  geological structures. The primary means of probing earth's interior such as the crust is seismology. However, even though we understand wave propagation and its numerical solution, capturing the effect of such a broad range of complexities at large distances remains computationally prohibitive.

To improve our understanding of the crustal structures and processes, we must therefore incorporate numerical seismic methods that capture the relevant physics and provide high-frequency resolution. Modelling seismic observations of nuclear explosions provides an attractive starting point for a rigorous analysis of the interaction of near-source complexities and crustal structures.

Aims of the Project

Improve discrimination capabilities for identifying nuclear explosions in complex geological settings.

Project Description

To improve event detection, location, and identification, we need to better understand the complexities governing high frequency regional and teleseismic wavefields. Attempts at modelling various effects of near-source scattering at teleseismic distances started in the late 1970s. However, due to high computational costs limiting such long-range simulations at high frequencies of interest, the efforts till date have been limited. Despite rapid hardware and software developments, capturing a broad range of heterogeneities with conventional seismic wave propagation codes remains computationally prohibitive on the global scale. To bridge the gap between complexity and computational cost, we employ a global Instaseis-based (van Driel et al. 2015, www.instaseis.info) injection type hybrid method. The modified Instaseis interface couples the global wave-propagation solver, AxiSEM (Nissen-Meyer et al. 2014, seis.earth.ox.ac.uk/axisem), with an arbitrary three-dimensional solver of choice (WPP or SW4), and thus embeds a heterogeneous 3D domain within a spherically symmetric Earth model.

Complex structures can be accounted for in the source region to model specific seismic observations such as topography and 3D geology. Hybrid simulations provide insight by quantifying how structures contribute to waveform characteristics at teleseismic distances (such as amplitude, dominant frequency, onset form and pulse duration), and thus could refine detection, location and source identification capacities for nuclear test monitoring.

Data from historical nuclear explosions provide a starting point for rigorous analysis of the effects of topography and 3D crustal structures around the source. Seismic data from global seismic stations from hundreds of nuclear tests provide the necessary well-defined setting for a controlled experiment. The source location, near-source topography and geology of many nuclear test sites are reasonably well-constrained, and waveforms have been extensively analysed at AWE Blacknest. Hybrid numerical modelling at high-frequencies can provide insights into the often complex seismic waves observed at teleseismic distances recorded from explosions.

The aim of this project is to use the well-controlled synthetic and observed data to quantify how heterogeneities contribute to waveform characteristics – such as amplitude, dominant frequency, onset form, pulse duration, or the balance of SV- and SH- polarised motions, and how they can be used to characterise the source signatures. Such thorough analysis and fundamental understanding of the structure could improve event detection, location, and identification, as well as help refine the existing nuclear discrimination methods.

This project will initially concentrate on modelling waveform characteristics observed teleseismically from explosions conducted at three nuclear tests sites. The first test case will use waveforms observed from the DPRK nuclear tests that show variation due to source location and direction of the seismic station from the event. These observed teleseismic waveform variations are thought to be generated from the wavefield interaction with the steep near-source topography at the DPRK test site. The student will incorporate near-source topography and model its influence on teleseismic waveforms.

The second test case will involve nuclear tests that took place on the Yucca Flats region of the Nevada Test Site (NTS). The Yucca Flats region sits on a large sedimentary basin with relatively flat topography. Yucca Flats explosion waveforms have often been noted to possess unusual reverberant P coda (Barker et al., 1980). Also, regional phases such as Pg and Lg recorded in Yucca Flats have anomalously large ampltiudes and have extended coda  (Mclaughlin et al.,1987). By varying near-source velocity structure (basin geometry) this project will endeavour to model the observed waveform complexity.

Finally, the complexity of seismic waveforms observed from the 28th May 1998 Pakistan nuclear explosion show clear azimuthal variation. These waveform variations are thought to be generated by near-source topography, dipping near-source geology or a combination of both. The student will incorporate both near-source topography and a complex 3D near-source velocity model to try and replicate these observed waveform variations and understand whether topography or geological structure is having the larger influence on the outgoing seismic wavefield.

The collaboration between the seismology group at Oxford and Blacknest has a long and successful history, most recently by a NERC-funded internship of a DPhil student at Blacknest in 2019

Methods to be used

AxiSEM3D, AxiSEM, instaseis (all in-house), Wave Propagation Program (WPP) and Seismic Waves 4th Order (SW4

Specialised skills required

Strong mathematics, physics understanding. Ideally programming experience, interest in numerical methods, test ban treaty monitoring

Please contact Tarje Nissen-Meyer on tarje.nissen-meyer@earth.ox.ac.uk if you are interested in this project.