Investigation of thallium fluxes from subaerial volcanism—Implications for the present and past mass balance of thallium in the oceans
Additional Document Info
A suite of 34 volcanic gas condensates and particulates from Kilauea (Hawaii), Mt. Etna and Vulcano (Italy), Mt. Merapi (Indonesia), White Island and Mt. Nguaruhoe (New Zealand) were analysed for both Tl isotope compositions and Tl/Pb ratios. When considered together with published Tl-Pb abundance data, the measurements provide globally representative best estimates of Tl/Pb = 0.46 +/- 0.25 and epsilon(205)Tl = -1.7 +/- 2.0 for the emissions of subaerial volcanism to the atmosphere and oceans (epsilon(205)Tl is the deviation of the (205)Tl/(203)Tl isotope ratio from NIST SRM 997 isotope standard in parts per 10,000). Compared to igneous rocks of the crust and mantle, volcanic gases were found to have (i) Tl/Pb ratios that are typically about an order of magnitude higher, and (ii) significantly more variable Tl isotope compositions but a mean epsilon(205)Tl value that is indistinguishable from estimates for the Earth’s mantle and continental crust. The first observation can be explained by the more volatile nature of Tl compared to Pb during the production of volcanic gases, whilst the second reflects the contrasting and approximately balanced isotope fractionation effects that are generated by partial evaporation of Tl during magma degassing and partial Tl condensation as a result of the cooling and differentiation of volcanic gases. Mass balance calculations, based on results from this and other recent Tl isotope studies, were carried out to investigate whether temporal changes in the volcanic Tl fluxes could be responsible for the dramatic shift in the epsilon(205) Tl value of the oceans at similar to 55 Ma, which has been inferred from Tl isotope time series data for ferromanganese crusts. The calculations demonstrate that even large changes in the marine Tl input fluxes from volcanism and other sources are unable to significantly alter the Tl isotope composition of the oceans. Based on modelling, it is shown that the large inferred change in the epsilon(205)Tl value of seawater is best explained if the oceans of the early Cenozoic featured significantly larger Tl output fluxes to oxic pelagic sediments, whilst the sink fluxes to altered ocean crust remained approximately constant. (C) 2009 Elsevier Ltd. All rights reserved.