Abstract

The standard Earth of geodynamics and geochemistry is rationalized from assumptions that the mantle is compositionally inverted—still-unfractionated lower mantle beneath volatile-depleted upper mantle—and that material circulates easily from bottom to top. Multidisciplinary data better fit a less-volatile and less-radioactive planet wherein depleted lower mantle, fractionated early and irreversibly, is decoupled from upper mantle plus crust that evolve and circulate separately. Early Archean fractionation produced global(?) felsic crust and refractory upper mantle. Later Archean granite-and-greenstone upper crust formed atop this ancient crust, which remained hot and weak; distinct continents and oceans did not exist, and upper mantle was much hotter than now. Plate tectonics began ca. 2.0 Ga when continents could stand above oceans and oceanic lithosphere could cool to subduction-enabling density and thickness. Upper mantle has since become more fertile and new increments of continental crust more mafic as continental crust has been progressively diminished by recycling into cooling mantle. Plate circulation is driven by sub-duction, which is enabled by density inversion produced by sea-water cooling from the top of oceanic lithosphere, is self-organized, and is confined to upper mantle. The matching rates of hinge rollback and of advance of fronts of overriding plates are keys to dynamics. Slabs sinking broadside from retreating hinges drive both subducting and overriding plates and force seafloor spreading in both shrinking and expanding oceans. An Antarctica-fixed framework depicts prediction-confirming “absolute” plate motions that make kinematic sense, whereas hotspot and no-net-rotation frames do not. Plumes from deep mantle, subduction into deep mantle, and bottom-up convective drive do not exist.