UC Santa Barbara

Abstract The interaction between aging oceanic plates and their underlying mantle is a crucial component of the plate tectonic cycle. Sub‐lithospheric small‐scale convection (SSC) explains why plates appear not to thicken after a certain age. Here, we link grain‐scale processes, dynamic models of asthenospheric flow, and seismic observations to gain new insights into the mechanisms of SSC. We present high‐resolution 3D geodynamic models of oceanic plate evolution with an Earth‐like rheology including coupled diffusion/dislocation creep and their interplay with evolving olivine grain size. Our models quantify how rheology affects the morphology and temporal stability of SSC, and we directly relate these quantities to geophysical observations from the Pacific OBS Research into Convecting Asthenosphere (ORCA) experiment. We convert variations in temperature, pressure, grain size, water content and stable melt fraction to seismic velocity and attenuation, seeking to match the wavelength and pattern of observed longitudinal convective rolls, the young SSC onset age, the large seismic velocity heterogeneity, low absolute seismic velocities, and high seismic attenuation. This requires low ( Pa s) asthenospheric viscosity, the contribution of both diffusion and dislocation creep to deformation, and the presence of volatiles and melt. Although SSC occurs at plate ages 60 Ma in our best‐fit model, the plate thermal structure approximately matches global observations of heat flux and bathymetry, indicating an important role of vigorous SSC in Earth's plate dynamics. However, reconciling all seismological observations is challenging, and additional mechanisms are required to explain the strong velocity heterogeneities suggested by body wave tomography. Plain Language Summary As oceanic tectonic plates age, they get colder and thicker. At some point, portions of the base of the plates may drip off into the hot mantle underneath them, a process known as “small‐scale convection” (SSC). This might explain diverse geophysical observations, including seismic imaging of cold blobs beneath the plates and the finding that plates seem to stop cooling with increasing age after they turn 70 million years old. We conducted computer simulations of aging oceanic plates that include novel components, such as more complex treatment of viscosity and its interplay with the size of crystal grains in rocks. Grain size turns out to be a key parameter in determining both how cold drips evolve and how seismic waves propagate through those rocks, which is how we identify structures beneath the surface. We link together the computer simulations with seismological studies by using simulated structure to predict synthetic earthquake data, for comparison with observations. In particular, we seek to match various attributes of apparent SSC seen recently beneath the central Pacific ocean plate. We find that small‐scale convection is indeed reproduced by our simulations, and determine that it manifests in a way that is consistent with several real‐Earth observations. Key Points 3D models with complex rheology and evolving grain size require low asthenospheric viscosity to match observed SSC in the central Pacific In our models, the seafloor depth–age curve flattens at a plate age (60 Ma) matching global observations, despite young SSC onset (35 Ma) A small amount of melt and volatiles is required to match the observed asthenospheric low seismic velocities and high attenuation