Comment on GSA Today article by Pavlis et al., 2019: “Subduction Polarity in Ancient Arcs: A Call to Integrate Geology and Geophysics to Decipher the Mesozoic Tectonic History of the Northern Cordillera of North America”

Pavlis et al. (2019) (abbreviated “P2019”) assert that “geologic and geophysical interpretations lead to fundamentally different conclusions regarding the polarity of subduction along the Cordilleran margin during late Mesozoic time” (p. 1). Their paper is a call to defend a model of uninterrupted eastward subduction beneath continental North America (which we refer to as an “alwaysAndean” style model) from purportedly contradictory geophysical observations. Our own work, critically cited 12 times on five pages, shows that no such contradiction exists. Neither geology nor geophysics supports always-Andean style subduction since 200 Ma (Sigloch and Mihalynuk 2013, 2017 [abbreviated “SM12013” and “SM2017”]). Instead, both record a Jura-Cretaceous period of simultaneous eastward and westward subduction under a vast archipelago in the northeastern proto-Pacific, analogous to today’s southwestern Pacific. Welded into westward-subducting lithosphere, North America (NAm) was pulled into the Archipelago and diachronously overrode it from ca. 155–50 Ma, accreting its arcs and microcontinents. P2019 portray our Archipelago model as featuring only westward subduction, then dismiss it by pointing to the Chugach subduction complex of Alaska, clearly derived from eastward subduction. In reality, our Archipelago model features as much eastward subduction as the Andean-style model, just located further west. Its very essence, missed by P2019, is long-lived subduction of two mature oceans beneath the Archipelago from opposite sides. Our Figure 1A corrects P2019’s rendering of our model (in their fig. 1 and discussion). Our Figure 1B develops today’s Southwest Pacific archipelago as a close tectonic analogue. The geologic crux of distinguishing between “always Andean” and “Archipelago” models is the fragmentary record of a suture between NAm and the Insular microcontinent (INS; ~“Wrangellia Composite Terrane”/ WCT of P2019). Heavy overprinting by the Coast Plutonic Complex results in observational gaps and interpretational ambiguities that are perilously discounted by P2019. Both inboard (continentward) and outboard of this overprinted suture, geologic records support Andean-style subduction over two periods— the “Native” Jurassic–Triassic arc on the cratonic southwestern United States (Dickinson, 2004; Barth et al., 2011; Riggs et al., 2016), and the Cenozoic Cascadia arc in the Pacific Northwest. The “always Andean” model originated as the simplest interpolation between these Jurassic and Cenozoic arcs, with no hypothesis-driven incentive to question Cretaceous times. Geophysical slab interpretations, including our own (Sigloch et al., 2008; Sigloch, 2011), followed geology’s lead in adopting the “always-Andean margin” interpretation up until SM2013. Geologists’ faith in the validity of this model crossed the community divide as a certainty cast in stone. Dissent among geologists in favor of Archipelago-like models (Moores, 1970, 1998; Schweickert and Cowan, 1975; Ingersoll and Schweickert, 1986; Cowan, 1994; Ingersoll, 2008) did not rise to audible levels for geophysicists. Also, as long as only the supermassive slab walls under NAm’s eastern seaboard were clearly visible to seismic tomography, they had to be interpreted as Farallon lithosphere in order to match the surviving seafloor isochron record (Atwater, 1989; Engebretson et al., 1985). Even a mantle reference frame was constructed to fix the locus of slab deposition to the continental margin (van der Meer et al., 2010; used by P2019 in circular reasoning to support their call). An easterly “Farallon slab” was only seriously questioned once a second set of lower-mantle slabs, more westerly and this time truly Farallon, came into focus under western NAm, thanks to USArray (Sigloch et al., 2008). P2019 exaggerate the geophysical uncertainties. Two decades after the iconic Farallon slab images of Grand et al. (1997) on the GSA Today cover, not many geophysicists will respond to P2019’s call to challenge “the hypothesis that the deep anomalies are indeed subduction zone remnants” (p. 5). P2019 claim serious problems with slab interpretation because, “These slabs are now in the mantle more than 3000 km from their presumed paleotrench. To restore the pathway over this distance requires multiple assumptions, including the nature of the mantle anomaly, uncertainties in slab sinking rates, and models of absolute plate motion” (p. 2). Confusingly they fail to mention that all these problems (serious indeed) arise only in the “always-Andean” slab interpretation. Our “tomotectonic” working hypothesis is that a slab indicates the former existence of a trench overhead, because accumulations of lithosphere sink essentially vertically at roughly the same rate everywhere— as expected intuitively and supported by simulations (e.g., Steinberger et al., 2012). Massive slab “walls” observed in the lower mantle under NAm must track paleo-trench lines of long-lived, stationary subduction zones (Fig. 1A). Mantle convection under North America, implied to be highly variable over time and space by the “alwaysAndean” model, becomes simple in the Archipelago framework. Quantitative plate