We thank Allmendinger et al. (2013) for their Comment and appreciate this opportunity to further discuss two main lines of arguments supporting reverse fault slip along the Santa María Fault Zone (SMFZ) during the 2010 Maule (Chile) earthquake: the spatial relation between megathrust slip and the SMFZ, and the relationship between coseismic and longer-term deformation patterns.
The upper-plate strain regime leading to reactivation of splay faults during megathrust earthquakes will mainly depend on the position of such faults with respect to the locus of plate-boundary slip. Because slip inversions are obtained using different fault geometries, regularization method, and data sets, results frequently differ. Thus, we believe that slip distributions should be interpreted in terms of their broad regional patterns, and not based on detailed features. We use the A-B-C down-dip zonation of Lay et al. (2012), where B is the most common large-slip region of megathrust earthquakes.
Because faults in northern Chile, where Allmendinger et al. worked extensively, are above zone C, they may be indeed reactivated as normal faults by earthquake slip in zones A and B. In contrast, the SMFZ is at the transition between domains A and B (∼80 km from the trench) where normal fault reactivation will be only favored by slip in zone A, while reverse motion will occur by slip in zones B and C. The Pichilemu fault discussed by Allmendinger et al. is at the B-C transition (∼125 km from the trench) and was reactivated as a normal fault, because here most of the slip occurred in zones B and A. In the SMFZ sector, models using the entire data set of GPS displacements show the most slip in zone B (e.g., Moreno et al., 2012) favoring reverse reactivation. The difference in slip zonation along the Maule rupture is empirically demonstrated by decimeters of coastal subsidence in the north and meters of coastal uplift in the south.
The inference of Allmendinger et al. that plate-boundary slip makes the SMFZ “almost ideally oriented for co/post-seismic reactivation as a normal fault” is based on Aron et al. (2013). But using coseismic Coulomb stress increments they concluded that “a negative stress field offshore suggests enhanced development of reverse faulting, which agrees with the observations of Melnick et al. (2012)” (paragraph 33 of Aron et al., 2013). Indeed, this inference is consistent with most slip in zone B and supports coseismic reverse motion along the SMFZ.
Aron et al. (2013) proposed that regional normal faulting is associated with permanent extension driven by megathrust-earthquake slip. Unfortunately, their compilation of normal faults contains major biases. For example, the NE-striking fault that best matches their conceptual model is located off Mocha Island and was mapped from bathymetry surveyed in the 1960s. A recent multibeam bathymetric survey shows exclusively NW-striking structures in this region (Geersen et al., 2011), but this study is not addressed by Aron et al. Indeed, the data of Geersen et al. pose a fundamental challenge to the model proposed by Aron et al. In addition, there is no evidence that normal faults selected by Aron et al. in the southern Maule rupture displace well-developed Quaternary marine terraces, and thus are likely associated with regional Eocene-early Pliocene extension (e.g., Kuhn et al., 2010).
Isla Santa María (ISM) was uplifted 2.2–1.6 m with a north-down tilt during the Maule earthquake. This tilt pattern closely mimics emerged Holocene beach ridges, which are progressively tilted with increasing age (Bookhagen et al., 2006). GPS surveys at ISM record overall subsidence before the earthquake (Melnick et al., 2012). Deformed marine strata imaged on seismic reflection lines surrounding the island show reverse-fault cored anticlines affecting the youngest units; and normal faults, solely affecting the older units. Small crustal normal faults such as those interpreted at ISM are below resolution. Thus, permanent deformation at ISM likely accumulates during megathrust earthquakes. Furthermore, the resemblance of coseismic and permanent tilt patterns, and the fact that permanent deformation is clearly associated with reverse faulting and related folding, support reverse reactivation during the 2010 earthquake. A seismic profile adjacent to the island (Melnick et al., their figure DR2) shows no evidence of normal faulting in the upper ∼1 km, suggesting such structures cannot be associated with permanent forearc extension, as argued by Aron et al (2013).
Allmendinger et al. seek to apply their conceptual model—a model developed to explain normal faulting in the forearc of northern Chile—to the entire Maule rupture. Although the model may apply to the northern Maule rupture, the southern sector is very different and a direct application of the model is problematic. The southern part (36.7–38.2°S) is characterized by higher topographic relief, the shortest coast-trench distance and fastest uplifting marine terraces along the Chile margin, and by metric coseismic uplift in 2010. Overall, our analyses of structures and terrace systems within the southern Maule rupture zone, at a wide variety of spatiotemporal scales, suggest that normal faulting has not played a significant role in Quaternary tectonic activity within this sector. We do agree, however, with the conclusion of Aron et al. (2013) that reverse faulting characterized the offshore region and the SMFZ during the Maule earthquake.