Well it is precisely the 'surrounding rocks' which are cooling. Here is a 2D (average transect with age) tomographic inversion for upper mantle temperature over Pacific oceanic lithosphere as a function of lithospheric age:
http://ciei.colorado.edu/...piro_research_files/image003.jpgThis thermal structure is clearly a consequence of gradual conductive cooling of the upper mantle with age, corroborated by many other geophysical observations in addition to seismic tomography. No other physical mechanism can account for these observations.
TC, you are touching on an important point here. IMHO, the frictional heating is almost irrelevant. The reason is that all we have done on the oceanic lithosphere formation and conveyance, is convert the original convective magmatic heat from the mantle into conductive heat in the crust which ulltimately flows to the ocean and atmosphere and eventually radiates into space.
Simply put, the earth is cooling. Any frictional heat is minor and transient. The end result is the same: a cooling planet.
I suggest the way to attack the heat problem is to calculate the average current heat flow from the earth and project that 4.5 ga into the past (this would be a minimum number since the heat flows of the distant past were certainly much higher - so we are only getting an end-member here).
Okay, now compress that amount of heat into 6ky and see what happens.
By the way, I'm not so sure that your statement about where the heat-forming radioelements are found is correct. While many of the elements are found in continental crust, where we can extract them, there is some consensus that the core is another concentration of radionuclides such as thorium and uranium. After all, what causes heating of the mantle plumes?
™ Version 4.2
