10:15 AM - 10:30 AM
[SCG46-15] Numerical simulations of spontaneous subduction initiation in the oceanic lithosphere: Effects of rheology of hydrated fracture zones and slab dehydration
Keywords:hydrated fracture zone, spontaneous subduction initiation, numerical simulation, antigorite, slab dehydration, fault strength
Spontaneous subduction initiation is caused by lateral density contrasts between oceanic plates at fracture zones or transform faults. Recent numerical modeling studies have shown that such pre-existing fault zones are required to be mechanically very weak to allow subduction initiation (Leng and Gurnis, 2015). However, the values of the fault strength used in their models are much smaller than those suggested by laboratory deformation experiments using materials of the crust and mantle. Recent petrological analyses have identified two possible rheological weakening effects in which (1) hydrothermal alteration of mantle peridotites is caused by seawater infiltration into transform faults, resulting in the formation of hydrous minerals such as amphibole, chlorite, and serpentine, and (2) extremely fine-grained olivine aggregates deform by diffusion creep (Kohli et al., 2021). In addition, the above two phenomena are also expected to occur at the base of the immature mantle wedge where slab-derived fluids are added.
In this study, we conducted two-dimensional viscoelastic-plastic numerical simulation of spontaneous subduction initiation in the oceanic lithospheres to understand the effects of seawater infiltration into fracture zones and slab dehydration on spontaneous subduction initiation. In the numerical models, a mantle convection code (I2ELVIS) is used, with a 15 km-wide fracture zone between the younger plate (OP) on the left and the older plate (DP) on the right. The age of the OP and DP ranges from 1 to 11 Ma and from 71 to 91 Ma, respectively. In the hydrous fracture zones, serpentinization (power-law creep of antigorite) and diffusion creep of wet olivine were assumed to occur at lower and higher temperatures, respectively. We ran two types of models: one in which serpentinization occurs when fracture zone materials originally present in the high-temperature region intrudes into the low-temperature region (within the antigorite stability region) (termed model A), and the other one in which serpentinization does not occur (termed model B). In the brittle regime, the effective coefficient of friction (μ' = μ(1 - λ); where μ is the coefficient of friction and λ is the pore fluid pressure factor) was systematically varied. In addition, we considered the effect of slab dehydration, in which serpentinite or wet olivine forms when slab-derived fluids percolate into the mantle wedge in the OP side.
In model A, when the age offset between the two lithospheres was 90 Myr, μ' of the fracture zone had to be less than 0.029 (μ = 0.29, λ= 0.9) for subduction initiation to occur. On the other hand, in model B, μ' had to be less than 0.010 (μ = 0.10, λ = 0.9) for the same age offset. In both model cases, fracture zone materials formed a 5 km thick low-viscosity (< 1019 Pa s) shear zone at the slab-fracture zone boundary at depths of 5-15 km. To clarify the difference in mechanical conditions between the two models, strength profiles were constructed assuming a strain rate of 10-12 s-1 and a pressure of 2 GPa. The strength profiles show that for model A, power-law creep of antigorite contributes to lowering the strength (up to 40 MPa) at regions where temperatures are below 500 °C.
These results suggest that shallow serpentinization (< 15 km) of the high-temperature fracture-zone materials in the immature mantle wedge that has an important effect on subduction initiation. Cox et al. (2021) conducted direct shear experiments on natural fault rocks from sheeted dolerite layer exhumed South Troodos Transform Fault Zone, and found that the coefficient of friction ranging from 0.28 (fault gouge) to 0.48 (matrix-rich fault breccias), meaning that the range of the coefficient of friction measured in experiments is higher than that of our models (μ = 0.29). Therefore, we propose the need for more numerical experiments that incorporate additional weakening processes, such as strain-dependent friction coefficient weakening (Gerya et al., 2021; Liu and Gerya, 2023) and ductile damage processes such as grain-size reduction assisted by Zener pinning (Bercovici and Ricard, 2012).
References: Bercovici and Ricard, 2012, Phys. Earth Planet. Inter., 202, 27-55. Cox et al., 2021, Geophys. Res. Lett., 48, e2021GL096292. Gerya et al., 2021, Nature. 599, 245-250. Leng and Gurnis, 2015, Geophys. Res. Lett., 42, 7014–7021. Liu and Gerya, 2023, J. Geophys. Res. Solid Earth, 128, e2022JB024701. Kohli et al., 2021, Nat. Geosci., 14, 606-611. Maunder et al., 2020, Nat. Commun, 11, 1874.
In this study, we conducted two-dimensional viscoelastic-plastic numerical simulation of spontaneous subduction initiation in the oceanic lithospheres to understand the effects of seawater infiltration into fracture zones and slab dehydration on spontaneous subduction initiation. In the numerical models, a mantle convection code (I2ELVIS) is used, with a 15 km-wide fracture zone between the younger plate (OP) on the left and the older plate (DP) on the right. The age of the OP and DP ranges from 1 to 11 Ma and from 71 to 91 Ma, respectively. In the hydrous fracture zones, serpentinization (power-law creep of antigorite) and diffusion creep of wet olivine were assumed to occur at lower and higher temperatures, respectively. We ran two types of models: one in which serpentinization occurs when fracture zone materials originally present in the high-temperature region intrudes into the low-temperature region (within the antigorite stability region) (termed model A), and the other one in which serpentinization does not occur (termed model B). In the brittle regime, the effective coefficient of friction (μ' = μ(1 - λ); where μ is the coefficient of friction and λ is the pore fluid pressure factor) was systematically varied. In addition, we considered the effect of slab dehydration, in which serpentinite or wet olivine forms when slab-derived fluids percolate into the mantle wedge in the OP side.
In model A, when the age offset between the two lithospheres was 90 Myr, μ' of the fracture zone had to be less than 0.029 (μ = 0.29, λ= 0.9) for subduction initiation to occur. On the other hand, in model B, μ' had to be less than 0.010 (μ = 0.10, λ = 0.9) for the same age offset. In both model cases, fracture zone materials formed a 5 km thick low-viscosity (< 1019 Pa s) shear zone at the slab-fracture zone boundary at depths of 5-15 km. To clarify the difference in mechanical conditions between the two models, strength profiles were constructed assuming a strain rate of 10-12 s-1 and a pressure of 2 GPa. The strength profiles show that for model A, power-law creep of antigorite contributes to lowering the strength (up to 40 MPa) at regions where temperatures are below 500 °C.
These results suggest that shallow serpentinization (< 15 km) of the high-temperature fracture-zone materials in the immature mantle wedge that has an important effect on subduction initiation. Cox et al. (2021) conducted direct shear experiments on natural fault rocks from sheeted dolerite layer exhumed South Troodos Transform Fault Zone, and found that the coefficient of friction ranging from 0.28 (fault gouge) to 0.48 (matrix-rich fault breccias), meaning that the range of the coefficient of friction measured in experiments is higher than that of our models (μ = 0.29). Therefore, we propose the need for more numerical experiments that incorporate additional weakening processes, such as strain-dependent friction coefficient weakening (Gerya et al., 2021; Liu and Gerya, 2023) and ductile damage processes such as grain-size reduction assisted by Zener pinning (Bercovici and Ricard, 2012).
References: Bercovici and Ricard, 2012, Phys. Earth Planet. Inter., 202, 27-55. Cox et al., 2021, Geophys. Res. Lett., 48, e2021GL096292. Gerya et al., 2021, Nature. 599, 245-250. Leng and Gurnis, 2015, Geophys. Res. Lett., 42, 7014–7021. Liu and Gerya, 2023, J. Geophys. Res. Solid Earth, 128, e2022JB024701. Kohli et al., 2021, Nat. Geosci., 14, 606-611. Maunder et al., 2020, Nat. Commun, 11, 1874.