Beauty and the Beast: Simple Brune source model applied to complex earthquakes

Abstract:

At first glance, the beauty of the Brune earthquake source model is its simplicity. It can be applied to earthquakes of variable sizes that occur in different tectonic settings. The simple relationship between corner frequency and earthquake source dimension allows a direct quantification of earthquake slip, stress drop and ground motions. I will present the applications of the Brune source model to a wide range of earthquakes, including microearthquakes, induced seismicity, transform earthquakes and deep-focus earthquakes. The consistent use of one source model facilitates the comparison of source parameters between different groups of earthquakes, which provides new constraints on their source mechanisms. I will show that stress drops of microearthquakes in an injection-induced experiment are comparable to those of induced earthquakes that are 8-9 units of magnitude larger, considering the dependency of stress drop on crustal shear strength. The median of Brune stress drops of deep-focus earthquakes is also found to be about a factor of 10 higher than that of shallow earthquakes.

However, a major obstacle hindering advances in earthquake source physics is the large variation of stress drop (over 3 orders of magnitude) estimated from a single corner frequency. One critical question arising from the observed variation is whether it results from the approaches and source models used to estimate corner frequency or from inherently different stress conditions on faults. I will give an overview of the recent progress in quantifying the uncertainties in stress drop estimation caused by different methodologies. I will show that both improved stress drop estimation methods and well-defined resolution limits of data can help reduce the uncertainties. Finally, I will discuss the limitation of the Brune source model when applied to large earthquakes with complex rupture processes. Our analysis of earthquakes with multiple subevents unveils a strong dependence of the Brune corner frequency on source parameters of the largest subevent and the delay time between subevents. I will highlight future directions that help improve the application of the Brune source model and aid in the interpretation of results.