On the use of $\mathrm{\delta^{18}O}$ as temperature proxy

Question

$\mathrm{\delta^{18}O}$ (i. e. the ratio of $\mathrm{^{18}O/^{16}O}$ compared to a standard) is massively used today in paleoclimatology as a global temperature proxy.
But how did it become a proxy for temperature? When and by whom was the correlation observed for the first time?
Was the model explaining why it followed global temperature found straight away or was this proxy purely empirical for a long time?

Answer 1

You're not going to believe this, but it turns out that the guy behind this was none other than Harold Urey (Yes, the Harold Urey of the Miller-Urey experiment!).

He made the discovery of the correlation sometime in the 1950s, and with his nuclear background, he was well aware of the properties of isotopes. He studied small ancient fossilized creatures called foraminifera, and used the ratio of $^{18}O/^{16}O$ in their shells to figure out ancient temperature patterns. Here's a quote of a passage from this site:

In the 1950s, the nuclear chemist Harold Urey devised another way to use the shells to measure ancient temperatures. He found he could take the temperature of an ancient ocean by measuring the oxygen that forams built into their shells. The rare isotope oxygen-18 is a bit heavier than normal oxygen-16, and biologists had shown that the amount of each isotope that a foram takes up varies with the temperature of the water. The isotopes were fossilized with the shells, and the ratio of isotopes (O18 to O16) could be determined with the new and exacting techniques of mass spectrometry. Urey and his team at the University of Chicago refined these tools, developed for nuclear studies, and applied them to calcite in fossils. They found plausible temperatures clear back to the Cretaceous era, more than 100 million years ago.

However, the page also says (a bit above this passage) that the person who realized that the foraminifera could give clues as to the temperature of the world at the time of their lives was Wolfgang Schott, working in the 1920s.

The page also cites two papers of Urey's, the first from 1947 and the second from 1950, making it seem as if his research was actually in the later 1940s.

Fun fact: The Wikipedia page on Urey hints that this work led him and one of his students to carry out the Miller-Urey experiment. I haven't been able to confirm this, though, so for now it's just a fun fact.