Constraining Earthquake Depth, Source Time Function and Focal Mechanism and Their Associated Uncertainties

Session: Earthquake Source Parameters: Theory, Observations and Interpretations
Type: Oral
Date: 4/30/2020
Time: 02:15 PM
Room: 120 + 130
Description: 

Many earthquake monitoring agencies from across the globe report earthquake mechanisms to the International Seismological Centre (ISC). For any given earthquake, there is significant variation not just in the location estimates reported to the ISC, but also in the source parameters reported. This likely reflects different methodologies and data availability of the reporting agencies, but also reflects the inherent range of plausible earthquake mechanises that would adequately fit the observed waveform data. Earthquake depth is often poorly resolved by the current methods, particularly for shallow earthquakes (<40 km depth) where distinct depth phase arrivals are not observable.

We invert for the earthquake source time function, moment tensor and depth, by fitting tele-seismic broadband body waves. Increased depth resolution is achieved by accounting for the surface-reflected depth phases. For intermediate depth earthquakes (70 - 300 km depth) it is well understood that these phases yield strong constraints on depth, but for shallow events they are often subsumed within the source time function and are therefore difficult to exploit. By jointly solving for the earthquake depth, focal mechanism and source time function, we are able to quantify the likelihoods of the observations fitting a range of trial depths.

The source time function is parameterised by an optimised set of base functions which are inferred from a highly quality controlled set of source time functions calculated manually using the method of Sigloch & Nolet (2006). The model space is sampled fully probabilistically using the neighbourhood algorithm and a set of likely earthquake mechanisms, source time functions and depths are produced by appraising this ensemble (Sambridge 1999a,b). A number of forward modelling methodologies are explored, including the established WKBJ method (Chapman, 1978) and the axiSEM based method Instaseis (van Driel et al., 2015). We outline the proposed methodology and show examples of the depth resolution and uncertainly estimates gained through these methods.

Presenting Author: Thomas I. M. Garth

Authors