Magnitude of repeating earthquakes due to slip over the same patch: implications on stress drop and rupture initiation.

Abstract:

The scaling of source parameters has been (and still is) a controversial aspect in our understanding of earthquake rupture. While some studies suggest that stress drop is constant as a function of magnitude, there are a number of studies that clearly show an increase in this parameter with magnitude. We use high-quality records of repeating intermediate-depth earthquakes in the Bucaramanga Nest, one of the most productive regions in the world. From 2010 to 2018 we have more than 160 M4+ earthquakes, many of them having very similar waveforms that we can assign to repeating families. After cross-correlation alignment, relocation, and clustering of the seismic waveforms of these repeating events, we find a robust increase of stress drop with magnitude. In particular, the corner frequency is almost constant for earthquakes with magnitude differences of up to 1.0. Using the corner frequency and seismic moment and a simple Brune source model, this would suggest that these different repeating events are occurring potentially on the same rupture patch, with very different slip but the same patch area. We test this model by simulating the ground motions due to ruptures along with a circular rupture patch, varying its area, rupture velocity, and slip. The observed ground motions cannot be explained by a constant stress drop model and it requires the rupture area to remain constant. Changes in rupture velocity can potentially allow for a varying rupture area but still does not explain the waveform similarity in a plausible way. Our model suggests a sharp increase of stress drop with magnitude for intermediate-depth earthquakes, as has been observed in the Bucaramanga Nest and other regions. Our particular interest is the observation that the waveforms are very similar throughout the entire record, so much that it should be possible to predict the eventual magnitude of the earthquakes within the early tenths of a second. That is assuming you know to which (repeating) family the earthquake belongs to.