Statistical Discrimination of Induced and Tectonic Earthquake Sequences in Central and Eastern US Based on Waveform Detected Catalogs

Abstract:

It is now well established that extraction of fossil fuels and/or waste water disposal do cause earthquakes in Central and Eastern United States (CEUS). However, the physics underneath of the nucleation of induced earthquakes still remain elusive. In particular, do induced and tectonic earthquake sequences in CEUS share the same statistics, for example the Omori’s law [Utsu et al., 1995] and the Gutenberg-Richter’s law? Some studies have show that most naturally occurring earthquake sequences are driven by cascading-type triggering. Hence, they would follow the typical Gutenberg-Richter relation and Omori’s aftershock decay and could be well described by multi-dimensional point-process models such as Epidemic Type Aftershock Sequence (ETAS) [Ogata, 1988; Zhuang et al., 2012]. However, induced earthquakes are likely driven by external forcing such as injected fluid pressure, and hence would not be well described by the ETAS model [Llenos and Michael, 2013]. Existing catalogs in CEUS (e.g. the ANSS catalog) have relatively high magnitude of completeness [e.g., Van Der Elst et al., 2013] and hence may not be ideal for a detailed ETAS modeling analysis. A waveform matched filter technique has been successfully applied to detect many missing earthquakes in CEUS with a sparse network in Illinois [Yang et al., 2009] and on single station in Texas, Oklahoma and Colorado [e.g., Van Der Elst et al., 2013]. In addition, the deployment of the USArray station in CEUS also helped to expand the station coverage. In this study, we systematically detect missing events during 14 moderate-size (M>=4) earthquake sequences since 2000 in CEUS and quantify their statistical parameters (e.g. b, a, K, and p values) and spatio-temporal evolutions. Then we compare the statistical parameters and the spatio-temporal evolution pattern between induced and naturally occurring earthquake sequences to see if one or more diagnostic parameters exist. Our comprehensive analysis of earthquake sequences in CEUS will hopefully shed new insight on whether intrinsic differences exist between induced and naturally occurring earthquakes.