Larger aftershocks happen farther away: Nonseparability of magnitude and spatial distributions of aftershocks

Abstract

Aftershocks may be driven by stress concentrations left by the main shock rupture or by elastic stress transfer to adjacent fault sections or strands. Aftershocks that occur within the initial rupture may be limited in size, because the scale of the stress concentrations should be smaller than the primary rupture itself. On the other hand, aftershocks that occur on adjacent fault segments outside the primary rupture may have no such size limitation. Here we use high‐precision double‐difference relocated earthquake catalogs to demonstrate that larger aftershocks occur farther away than smaller aftershocks, when measured from the centroid of early aftershock activity—a proxy for the initial rupture. Aftershocks as large as or larger than the initiating event nucleate almost exclusively in the outer regions of the aftershock zone. This observation is interpreted as a signature of elastic rebound in the earthquake catalog and can be used to improve forecasting of large aftershocks.

A Spatiotemporal Clustering Model for the Third Uniform California Earthquake Rupture Forecast (UCERF3‐ETAS): Toward an Operational Earthquake Forecast

“…We just do not see this kind of re-rupture occurring in nature, at least not anywhere near 77% of the time. Furthermore, and as discussed above, van der Elst and Shaw (2015) have presented observational evidence that aftershocks as large or larger than the parent nucleate almost exclusively in the outer regions of the parent aftershock zone, giving further evidence of some elastic-relaxation process.…”

“…There is also observational evidence that smaller subseismogenic ruptures exhibit such ERT effects. Using high-precision double-difference relocations for observed M 4-6.7 earthquakes, van der Elst and Shaw (2015) have shown that aftershocks that are as large or larger than the parent nucleate almost exclusively in the outer regions of the parent aftershock zone, which they interpret as evidence for elastic relaxation. Though applying such a recent finding here may seem premature, it does solve another potential problem with respect to triggering probabilities when transitioning between subseismogenic and supraseismogenic mainshocks.…”

Change in seismicity along the Japan trench, 1990–2011, and its relationship with seismic coupling

“…Triggering by previous earthquakes occurring at { x i , y i , t i } is modeled as i.e., the product of the Omori‐Utsu law with a power law spatial density. We define the characteristic length (or rupture radius) in kilometer [ Utsu and Seki , ; Van der Elst and Shaw , ]. We fix the model parameters to α = 2, p = 1 (since previous estimates of p performed for Japan using the JMA catalog range from p ≃0.9 [ Marsan et al , ] to ≃1.1 [ Zhuang et al , ]), c = 10 −3 days (sensitivity of our results on the c value is very weak for c in the 1 min (≃710 −4 days) to 1 h (≃0.04 days) interval), and γ = 2 (consistent with 1.7 < γ < 2.1 usually found for the decay of the linear density of aftershocks [cf.…”

Long‐Term Interactions Between Intermediate Depth and Shallow Seismicity in North Chile Subduction Zone

“…That is, the product of the Omori-Utsu law with a power spatial density, with c, , and p constants. L(m) = L 0 × 10 0.5(m−4.7) is the characteristic length in kilometer (Elst & Shaw, 2015;Utsu & Seki, 1955) and (m) is the productivity law, with constant (Ogata, 1988).…”