Authors
John Townend, Yannik Behr, Kevin Buckley, Martha Savage, and John Hine (Victoria University of Wellington)
Abstract
As in many other data-intensive fields of science, geophysical research often involves a large number of incremental processing steps, during each of which decisions must be made regarding the particular choice of parameters or even algorithms to use. The research undertaken in this project is intended to combine an increasingly routine geophysical task, cross-correlating large data sets, with modern approaches to systematising and documenting the research process itself. Cross-correlation of seismic waveforms underpins much of modern seismology, particularly in the areas of differential earthquake location and ambient noise tomography. The cross-correlation operation computes the similarity between two signals and the optimal time-shift required to align one with the other: in broad terms, accurate earthquake location is based on measurements of this time-shift, and ambient noise tomography on the degree of similarity.
In recent years, it has become feasible to cross-correlate large data sets containing several tens of thousands of earthquakes or spanning many months of continuously recorded seismic noise. In particular, by comparing long records of ambient seismic noise recorded at two different locations, a small amount of coherent seismic energy propagating directly between them can be detected. This energy propagates as a seismic wave at a speed governed by the physical properties of the rocks it passes through. By measuring this speed, we can map the Earth’s deep structure in much the same way as ultrasound is used to look inside human bodies.
We focus here on performing the routine cross-correlation of long records of continuous ambient seismic noise using computational grid resources within a Sun Grid Engine environment. Our research is underpinned by high-quality continuous seismic data recorded by GeoNet, a New Zealand-wide network of seismographs (and other geophysical instruments) operated by GNS Science. In this presentation, we outline some of the lessons learned in matching geophysical demands to specific computational resources.