A record of Neogene seawater [delta].sup.11B reconstructed from paired [delta].sup.11B analyses on benthic and planktic foraminifera

The boron isotope composition ([delta].sup.11 B) of planktic foraminiferal calcite, which reflects seawater pH, is a well-established proxy for reconstructing palaeo-atmospheric CO.sub.2 and seawater carbonate chemistry. However, to translate [delta].sup.11 B measurements determined in calcareous fossils into pH we need to know the boron isotope composition of the parent seawater ([delta].sup.11 B.sub.sw). While a number of [delta].sup.11 B.sub.sw reconstructions exist, the discrepancies between them reveals uncertainties and deficiencies that need to be addressed. Here we present a new [delta].sup.11 B.sub.sw record based on the [delta].sup.11 B difference between planktic and benthic foraminifera and an estimate of the pH gradient between surface and deep water. We then calculate [delta].sup.11 B.sub.sw two different ways. One variant of our method assumes that the pH gradient between surface and deep has remained the same as today over the past 23 Ma; the other uses the [delta].sup.13 C gradient between surface and deep to represent change in the pH gradient through time. The results of these two methods of calculating [delta].sup.11 B.sub.sw are broadly consistency with each other, however, based on extensive carbon cycle modelling using CYCLOPS and GENIE we favour the [delta].sup.13 C gradient method. In our favoured [delta].sup.11 B.sub.sw reconstruction, [delta].sup.11 B.sub.sw is around 2 ‰ lower than today at ~37.5 ‰ during the early and middle Miocene and increases to the modern value (39.61 ‰) by ~5 Ma. A similar pattern of change is evident in the seawater composition of three other stable isotope systems, Mg, Li and Ca. Concurrent shifts in the seawater isotopic composition of all four of these elements during the late Miocene, suggest a common forcing mechanism. We hypothesise the most likely cause of these shifts is a change in the isotopic composition of the riverine input, potentially driven by an increase in secondary mineral formation since ~15 Ma.