Effects of basal drag on subduction dynamics from 2D numerical models

Abstract

Abstract. Subducting slabs are an important driver of plate motions, yet the relative importance of different forces in governing subduction motions and styles remains incompletely understood. Basal drag has been proposed to be a minor contributor to subduction forcing because of the lack of correlation between plate size and velocity in observed and reconstructed plate motions. Furthermore, in single subduction system models, low basal drag leads to subduction behaviour most consistent with the observation that trench migration velocities are generally low compared to convergence velocities. By contrast, analytical calculations and global mantle flow models indicate basal drag can be substantial. In this study, we revisit this problem by examining the drag at the base of the lithosphere, for a single subduction system, in 2D models with a free trench and composite non-linear rheology. We compare the behaviour of short and long plates for a range of asthenospheric and lithospheric rheologies. We reproduce results from previous modelling studies, including low ratios of trench over plate motions. However, we also find that any combination of asthenosphere and lithosphere viscosity that produces Earth-like subduction behaviour leads to a correlation of velocities with plate size, due to the role of basal drag. By examining Cenozoic plate motion reconstructions, we find that slab age and plate size are positively correlated: higher slab pull for older plates tends to be offset by higher basal drag below these larger plates. This, in part, explains the lack of plate velocity–size correlation in observations, despite the important role of basal drag in the subduction force balance.

The Role of Subduction Interface and Upper Plate Strength on Back‐Arc Extension: Application to Mediterranean Back‐Arc Basins

“…Some previous numerical modeling studies suggest that viscous drag along the base of the overriding plate lithosphere controls the back‐arc stress field (e.g., Capitanio et al., 2011; Dal Zilio et al., 2018; Holt et al., 2015; Sleep & Toksoz, 1971; Suchoy et al., 2021). The basal drag force in our models is following Billen (2008) approximated by where v diff ∼ 5 cm year −1 is the velocity difference between overriding plate and the underlying mantle, μ ml ∼ 1 × 10 19 to 1 × 10 20 Pa s the viscosity of the upper mantle at the LAB, L op ∼ 1,500 km the width of the overriding plate, and the same length scale related to the velocity difference that was used for estimating ( = 100 km; after Billen, 2008).…”

“…Compared to the estimates of slab‐pull and interface resistance, viscous drag is thus unlikely to play a critical role influencing the back‐arc stress field in our models. It should be noted that with higher viscosities in the sublithospheric mantle and larger plates, viscous drag can become a significant actor (e.g., Dal Zilio et al., 2018; Suchoy et al., 2021).…”

“…Similarly, cases of reduced interface resistance lead to an increase of the net driving force available for overriding plate extension. Some previous numerical modeling studies suggest that viscous drag along the base of the overriding plate lithosphere controls the back-arc stress field (e.g., Capitanio et al, 2011;Dal Zilio et al, 2018;Holt et al, 2015;Sleep & Toksoz, 1971;Suchoy et al, 2021). The basal drag force in our models is following Billen (2008) approximated by…”

The Effect of a Weak Asthenospheric Layer on Surface Kinematics, Subduction Dynamics and Slab Morphology in the Lower Mantle

“…Some studies, which consider it to be a layer of partial melt, suggest that it is only 10-20 km thick (Sakamaki et al, 2013;Schmerr, 2012;Stern et al, 2015), whereas others advocate for a layer extending from the lithosphere-asthenosphere boundary up to 200-300 km depth (thus a thickness of approximately 100-200 km) (Barruol et al, 2019;Becker, 2017;Debayle et al, 2020;French et al, 2013;Kawakatsu et al, 2009;Paulson & Richards, 2009). Here, we simulate the presence of the WAL by imposing a viscosity reduction between the 1,100°C isotherm (a proxy for the LAB) and a depth of 220 km, similarly to Suchoy et al (2021). Note that models tested with a WAL extending up to 300 km depth showed little differences with the results reported below.…”

“…The presence of a WAL is predicted to affect large-scale dynamics of the underlying, convecting mantle (Lenardic et al, 2006), and to favor "plate-like" rather than "stagnant-lid" regimes (Höink et al, 2012). Since the sub-lithospheric mantle resists a plate's trenchward motion, the inclusion of a WAL in models of subduction dynamics yields faster subduction velocities v sp , as shown by Carluccio et al (2019) and Suchoy et al (2021). The latter authors also showed that increased v sp was coeval with reduced trench retreat v t , although they did not detail the implications for lower mantle slab morphologies.…”