The nonsense of convection
              (...or putting the 'con' in convection...)

Fig.1.  Plate tectonics  new convection model  - the left-hand turn at the crest of the cell, showing mantle rise, the fillet of partial melt intrusion in the crust (red sliver) and the ocean floor plate (blue).  Not to scale.

 

Mantle convection is supposedly the mechanism of plate tectonics, but no matter how much numerical sleight of hand, calculation, or new ("poloidal-to-toroidal")  jargon is invoked to support this convection, the nonsense is clear -  2000 km (or so) girth of rising mantle is transmitted through a knife-edged ridge fracture (and a double phase-change) before proceeding on its way.   This transmutation of convecting cell (twice) through a phase boundary  at the peak of its 'career' also implies a multiplier of 100 times in the rate of  lateral movement at the surface compared to upwards movement of the cell.   The problem is compounded at the subducting side of the cell (not shown), where exactly the same nonsense is invoked in reverse, where the downgoing slab is assimilated into the mantle.

Ye olde conveyor belt model, now eschewed by ridgie-didge platies (but still "Good enough for schools and colleges.")

The brand-spanking "the plates are the convection" bonzer beauty ( tarann!)  new model, which promises clarity of  vision absolute to the convinced, converted (conned)  'convection community'