Following recent develoments in geophysics (Schmidt et al. (2014), and a popular summary), we now know there to be a significant reservoir of water in the mantle-transition of planet Earth, now often quoted as "additional 3 oceans worth of water in the mantle". Here one ocean is being counted as $\sim 10^{23}\rm kg$ of water.

While a definitive answer to this is probably outstanding, I would be interested in knowing where this water was, prior to the onset of plate tectonics (e.g. see this question) at probably $\sim 3\rm \,Ga$.

Were the - in total - 4 oceans worth of water (ignoring loss of volatiles to space) delivered with the initially accreted solids, or did it come as a late veneer, cover the Earth first completely, before slowly diffusing into the mantle, starting tectonics and establishing re-/degassing equilibrium at 1 ocean on the surface? What geophysical data is out there to decide between those scenarios?

A recent review on the topic by Karaki et al. (2020), does not say much about the planet formation perspective on this, and the aim of my question points in the direction of deciding whether pebble accretion or planetesimal accretion would have been the main contributor for volatiles on our planet.

  • $\begingroup$ there was no liquid water on earth when plate tectonics started. $\endgroup$
    – John
    Jun 1, 2021 at 3:19
  • $\begingroup$ by "come as a late veneer", do you mean through comet/asteroid impact? $\endgroup$
    – f.thorpe
    Jun 1, 2021 at 4:08
  • 2
    $\begingroup$ @John Your comment would quickly be tagged with "citation required" had you written that on wikipedia. $\endgroup$ Jun 1, 2021 at 8:43
  • 1
    $\begingroup$ @John Furthermore the appearance of iron banded formations at 4.2 Ga requires liquid water and life to be present, so would you please clarify why you think there was no liquid water at 3 Ga? $\endgroup$ Jun 1, 2021 at 12:27
  • 1
    $\begingroup$ @AtmosphericPrisonEscape plate tectonics starts while the earth is forming, as it is molten and material begin to differentiate out due to buoyancy, literally as soon as you start having even semi solid rock, which is going to be at a much higher temperature than you can have liquid water. even the question you link indicates a start older than ~4Ga. there is a difference between modern plate tectonics and any plate tectonics. there is a continuum of processes not a discreet start time. $\endgroup$
    – John
    Jun 1, 2021 at 14:20

1 Answer 1


This all comes down to a single issue.

Define plate tectonics.

There are several ways it can be defined which is why estimates of when it started vary by more than half the age of the planet. At the broadest definition plate tectonics predates liquid water by a decent margin, by the narrowest it may not even predate multicellular life.

Like most process plate tectonics definition varies depending on what factors you consider essential to the process, because in many ways it is part of a continuum of processes, so it comes down to where you draw the artificial line. do you draw the line a the first mantle convection cell, the presence of layered convection cells, the first solid surface rock, A mantle convection speed similar to today, the formation of the moho? Each can give a drastically different date and some can be quite difficult to estimate a time for. Which is one reason the accepted answer to your linked question restates this problem.

Importantly I don't know if even answering the question posted will help you with your underlying question, as liquid water could not occur until after the accreting body would be cool enough.

I suggest asking your underlying question, "where did the water on earth come from", it will provide far more constructive answers to your question. one thing we do know is that water would have been very abundant in the suns accretion disk and thus the forming planets, since hydrogen and oxygen are two of of the three most abundant elements.

  • $\begingroup$ Well, I guess then according to my definition, plate tectonics would be "any process that is capable of burying 3 oceans worth of water in the mantle, under the assumption, those masses of water started out on the surface". I am aware of the research literature surrounding volatile delivery to Earth, and that we neither know what mass fraction Earth was at disc dispersal, nor do we know its volatile content, as the ratio of pebble-to-planetesimal accretion (and its time-dependence) is unknown. Hence I tried to approach this problem from a more geologic POV, hence my question. $\endgroup$ Jun 11, 2021 at 18:05
  • $\begingroup$ @AtmosphericPrisonEscape Only if assume the water did not start in the mantle and indeed the rest of the planet and got squeezed out to the outer layers via density stratification. $\endgroup$
    – John
    Jun 11, 2021 at 18:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.