Magnitude-Frequency Distribution for Potentially Induced Earthquakes in the Guy, Arkansas Sequence

Abstract:

The Gutenberg-Richter (G-R) distribution, logN=a-bN, is commonly used to describe the magnitude-frequency distribution (MFD) of earthquakes. The b-value, which characterizes the relative number of large vs. small earthquakes, is typically around 1 for tectonic earthquakes, but higher values have been reported for earthquake swarms and volcanic events. Both geometric and dynamic arguments have been used to explain why the b-value tends to be close to 1.

The earthquake catalog detection threshold is usually based on a break in b-value, but this has the important disadvantage of assuming the form of the underlying MFD a priori. Moreover, it’s difficult to discern spatial variations in the detection threshold or to document time-dependent changes, and it's silent about the detection threshold where no earthquakes are detected. We use the approach of Schorlemmer and Woessner [2008] to determine catalog completeness levels by calculating the detection probability as a function of magnitude and distance. We also develop a catalog by using template matching against known earthquakes. Template matching allows us to both lower the detection threshold, and to calculate more precise event magnitudes.

We apply this method to the Guy earthquake sequence in central Arkansas, which shows spatial and temporal correlation with injection wells. Our template matching technique leads to a remarkable increase of the number of earthquakes and significantly improves the magnitude of completeness. The earthquakes continuously migrated since late September, 2010 and eventually illuminated a ~ 13 km long fault in central Arkansas [Horton, 2012].

We have undertaken this study of earthquakes induced by fluid injection to determine whether their MFD is well described by G-R statistics, and if so, whether there are significant differences between induced and tectonic earthquakes. We will also explore whether there are temporal changes in b-value that might be diagnostic of changes in deformation patterns of induced seismicity.