Has Earth’s mantle always worked like it does today?

This seems to be the latest round in the eternal battle between the geophysics data which strongly suggests whole mantle convection (and for quite some time, if ideas about the origin of all the weird junk at the core-mantle boundary are correct) and the geochemistry data which suggests long-lived and distinct mantle reservoirs. The authors of this study in Nature are in latter camp.

I’m not against idea that this conundrum might be solved by changes to mantle convection regime over time. Over the 4.5 billion years of Earth history, the mantle has cooled substantially; and mantle convection is driven mostly by the stuff sinking into the mantle by above, which has changed – in nature and possibly also mechanism – too.

But in proposing a ‘recent’ transition between an impermeable 660 km mantle transition (isolating the upper and lower mantle from each other) and an impermeable one (as seems to be the case for the Earth of the present), the authors of this paper are frustratingly coy about exactly how ‘recent’ they are thinking. If recent is sometime in the late Neoproterozoic (700–800 million years ago)¹, this seems more plausible than say, in the last 100 million years.

Nonetheless, I’ve always found this particular debate a fascinating example of how scientists approach different datasets. Both geophysics and geochemistry data are complicated – you are trying to get to what you really want to see (structures and processes) indirectly, and you have to make assumptions and approximations about how what you can measure relates to those things.

When, like here, two datasets seem to be telling you different things, scientists tend to trust the interpretation of the one they are more familiar with and assume the problem is with the other one.

Everyone agrees we’re missing something – the disagreement is over where. And you trust what you know.

---

1. All the weird stuff was happening then, so why not one more?