Water circulation and global mantle dynamics: Insight from numerical modeling

Abstract

We investigate water circulation and its dynamical effects on global-scale mantle dynamics in numerical thermochemical mantle convection simulations. Both dehydration-hydration processes and dehydration melting are included. We also assume the rheological properties of hydrous minerals and density reduction caused by hydrous minerals. Heat transfer due to mantle convection seems to be enhanced more effectively than water cycling in the mantle convection system when reasonable water dependence of viscosity is assumed, due to effective slab dehydration at shallow depths. Water still affects significantly the global dynamics by weakening the near-surface oceanic crust and lithosphere, enhancing the activity of surface plate motion compared to dry mantle case. As a result, including hydrous minerals, the more viscous mantle is expected with several orders of magnitude compared to the dry mantle. The average water content in the whole mantle is regulated by the dehydration-hydration process. The large-scale thermochemical anomalies, as is observed in the deep mantle, is found when a large density contrast between basaltic material and ambient mantle is assumed (4-5%), comparable to mineral physics measurements. Through this study, the effects of hydrous minerals in mantle dynamics are very important for interpreting the observational constraints on mantle convection.

Long‐Term Stability of Plate‐Like Behavior Caused by Hydrous Mantle Convection and Water Absorption in the Deep Mantle

“…This would allow stored water to be transported by subducted slabs in the mantle transition zone. This phenomenon is caused by high water solubility in the mantle transition zone minerals, as shown in the water solubility map of mantle minerals as functions of temperature and pressure (e.g., Iwamori, ) and the large amounts of water transported by subducting slabs with a smaller horizontal‐scale (Nakagawa et al, ). The major mechanism operating in a burst of mantle water content would efficiently cool the mantle such that the mantle transition zone could store the water transported from subducted slabs.…”

“…For the hydrous mantle, large‐scale thermo‐chemical structure is observed because the weak hydrous oceanic crust may lubricate the oceanic lithosphere more efficiently. This large‐scale thermo‐chemical structure in the deep mantle is similar to cases with weakly water‐dependent viscosity, i.e., r = 0.3 (Nakagawa et al, ). Again, r is the prefactor dependence of hydrous mantle rheology given in equation .…”

“…Most of the details of the numerical model used here are given in our previous study (Nakagawa et al, ). We assume a compressible and truncated anelastic approximation for the modeled mantle.…”

“…Water transport in the mantle can be expressed as where C w is the mantle water content, u is the convective velocity, and S w is the source‐sink term related to regassing; dehydration, such as nonmagmatic excess water migration; and degassing (including dehydration melting). The source‐sink effects of water migration processes are formulated as where ρ m is the mantle density, is the total mass of erupted material, C w is the water content; u z is the radial velocity, C w ,erupt is the water content of the erupted material equivalent to the water content of the molten material (see equation ), C w , ex is the excess water content relative to the water solubility, and u f is the conventional form of fluid migration velocity expressed numerically, which is computed as the ratio of radial grid spacing to the time step size (Nakagawa et al, ; Nakagawa & Spiegelman, ; Nakao et al, ). Note that computing source‐sink fluxes is done by post‐processing; therefore, they do not influence the numerical results directly.…”

“…The initial mantle water content may not appreciably influence the long‐term mantle water content evolution, as discussed in Nakagawa and Spiegelman (). In addition, Nakagawa et al () and Nakagawa and Spiegelman () have noted that the mantle water content is strongly regulated by excess water migration, which suggests that the initial mantle water content is unimportant for hydrous mantle convection simulations.…”