[S-GC52] [EE] Volatile cycles in the Earth - from Surface to Deep Interior

[SGC52-08] Seawater cycled throughout Earth's mantle in partially serpentinised lithosphere

★Invited papers

*Mark Kendrick¹, Christophe Hemond², Vadim S Kamenetsky³, Leonid Danyushevsky³, Colin W Devey⁴, Thomas Rodemann⁵, Matt G Jackson⁶, Michael R Perfit⁷ (1.Research School of Earth Sciences, Australian National University, Australia, 2.Laboratoire Geosciences Ocean, IUEM-UBO, France, 3.CODES and Earth Sciences, University of Tasmania, Australia, 4.GEOMAR Helmholtz Centre for Ocean Research Kiel,Germany, 5.Central Science Laboratory, University of Tasmania, Australia, 6.University of California Santa Barbara, Department of Earth Science, USA, 7.Department of Geological Sciences, University of Florida, USA)

Keywords:halogens, water, HIMU, Ocean Island Basalt, Mid Ocean Ridge Basalt

Recent work has indicated that about 90% of non-radiogenic xenon in the Earth’s mantle has a subducted atmospheric origin, suggesting that other seawater-derived components may also have subducted origins in the mantle. In order to test this proposition, we investigated the concentrations of water and halogens (F, Cl, Br, I) in samples of magmatic glasses collected from globally distributed mid ocean ridges and oceanic islands including melts derived from all the mantle end-member reservoirs: Depleted MORB mantle (DMM), Enriched mantle (EM1 and EM2) and high-µ (HIMU). After eliminating samples that show evidence for late-stage assimilation of seawater components, our data show unexpected broad correlations between mantle H₂O/Ce, Cl/K, F/Pr, Nb/U and Ce/Pb. The data confirm that EM reservoirs with low Nb/U and Ce/Pb ratios have low H₂O/Ce, Cl/K and F/Pr that demonstrates a relative depletion in water and halogens, consistent with the presence of dehydrated sediments or continental crustal material in EM sources. In contrast, HIMU reservoirs, which are depleted in most fluid mobile trace elements and characterised by high Nb/U and Ce/Pb ratios, are substantially enriched in H₂O and halogens. The H₂O and halogen enrichment of these sources is at odds with HIMU reservoirs being derived solely from dehydrated ocean crust, but can be easily explained if subducted ocean crust is associated with variable quantities of serpentinised lithospheric mantle. Our data also show that the abundance ratios of the most incompatible halogens (Br/Cl and I/Cl) have overlapping and narrow ranges in MORB and all OIB. Furthermore, the median I/Cl of the mantle is estimated as 0.000063 ± 0.000005 which is substantially lower than the calculated primitive mantle value of 0.00027 ± 0.00012. Given that I and Cl have similar incompatibilities in the mantle, this cannot be explained by melting related extraction but can be explained by preferential subduction of Cl relative to I, which is suggested by the relative abundances of these halogens in amphibolites and antigorite-serpentinites. We calculate that subduction of 1-2 % serpentine in the uppermost 10 km of the lithospheric mantle is enough to generate the maximum H₂O/Ce (280-400), Cl/K (0.12-0.16), F/Pr (140-160) and Nb/U (65-70) ratios observed in a HIMU reservoir. Furthermore, a figure of this magnitude, equivalent to a subduction flux of 1-3×10¹¹ kg yr^-1 H₂O and 2-4×10⁹ kg yr^-1 Cl, is required to balance global input of H₂O into the mantle with global output of H₂O into the surface reservoirs, and maintain constant sea level through the Phanerozoic. This level of subduction suggests that the entire mantle inventory of H₂O and halogens could have been subducted in about 4 billion years and an equivalent amount transferred to the surface reservoirs. Therefore the Earth’s mantle is now dominated by subducted H₂O, Cl, Br and I and the exchange of volatiles between Earth’s surface reservoirs and mantle is much greater than previously anticipated.