Matt Vaughan with Ben Holtzman, Felix Waldhauser, and Lapo Boschi
Can We Use Sound Perception to Identify Rupture Patterns in Earthquakes?
The processes by which earthquakes occur is poorly understood. In more detail, an earthquake results from the relative motion of blocks of crust, or pieces of tectonic plates, across a fault. The rupture nucleates at a point and propagates outward. The area and rupture time are reflected in the “Magnitude” of the earthquake, with larger rupture times and areas for larger magnitudes. However, in detail, the geometry of the rupture through time, or the “source function”, are not well known or understood. Progress is difficult to make because the problem is a difficult one: How do you uniquely identify the source function from ruptures and from surface displacement recorded on nearby and global seismometers? A better understanding of this problem will lead to insight into the problem of why earthquakes are difficult to predict, and how we may better predict them and assess seismic hazards. The aim of this project is to develop a different and completely novel approach to the inverse problem: we will create spatialized sound, created from multiple seismometers in a network around the fault, and see to what extent we can solve this inverse problem using our cognitive abilities to perceive motion in sound. Importantly, even if we are only provided with two ears, our hearing apparatus is not limited to two receivers, as we can count on (i) our anatomical transfer function, and (ii) our ability to move through the sound field we are experiencing. This idea comes directly from initial results from Vaughan’s internship with Holtzman at LDEO this summer (2013). Waldhauser provided us with a data set of repeating earthquakes from the Parkfield region of the San Andreas Fault, in Northern California. Over 25 years, one small part of the fault produced earthquakes of very similar magnitude and focal mechanisms with a <2 year recurrence time, recorded on multiple seismic stations. On our array of 16 speakers on a wall at LDEO, Vaughan found that there was a subtle sense that the waves were propagating across the array differently for several of the “repeating” earthquakes. This sparked the intriguing possibility that we variations in the source geometry from the wave propagation across the array are audible. Can we “hear” the answer, or at least a strong constraint, on the inverse problem for the source geometry?