Magmatic volatiles at rifts and arcs: Sources and fractionation effects

In Chapter 1, we show that nephelinite melts at Oldoinyo Lengai are the most carbon-rich natural silicate melts known to science. However, we argue that the mantle source for these melts is not unusually rich in carbon. Rather, extreme degrees of fractional crystallization and the high carbon solubility in these alkali-rich melts are responsible for the high observed carbon contents. We also show that the Oldoinyo Lengai magmatic system is rich in water, contrary to prior assumption, and that water degassing plays a fundamental role in the eruptive behaviour and magmatic evolution at Oldoinyo Lengai. The appendix to Chapter 1 contains additional data (including carbon, oxygen ,and sulfur isotope compositions and assessment of sulfur behaviour in the magma system) from Oldoinyo Lengai that have not been published.;Chapter 2 presents the first bulk gas and nitrogen isotope compositions from the Rungwe Volcanic Province, which is the southernmost volcanic manifestation of rifting in East Africa. In this paper, we show that the gases emitted at Rungwe are CO2-rich, presenting a hazard to the inhabitants of this fertile and thus heavily populated area. We find that the gas compositions record high temperatures in the deeper hydrothermal system, which could be a valuable geothermal resource for local economic development. The nitrogen isotope compositions and gas ratio tracers are consistent with an upper mantle source for the gases, and this signature is strongest in the central part of the province where the intersection of deep crustal structures provide direct conduits for mantle degassing. Interestingly, the mantle signatures are associated with the lowest temperature emissions in the region. The structures provide conduits for melts as well as gases, leading to the coincidence of mantle gas emission and volcanic edifice building in the central Rungwe Volcanic Province. These high elevation mountains are the recharge zones for the shallow aquifer and the mantle gases thus equilibrate with cold meteoric water close to the surface.;Chapter 3 examines S degassing during an explosive eruption at Anatahan (Mariana Arc) that emitted about 250 ktons of SO2 in the first ten days of the eruption. The results of this study show that the source of erupted sulfur was from ultimately from the mantle, with little addition from subducted seawater sulfate, or the pre-eruptive hydrothermal system. Sulfur isotopes fractionate during the degassing process, and the change in sulfur isotope compositions through the eruption are consistent with closed system degassing of a magma body.;Chapter 4 presents a detailed assessment of the S cycle at persistently degassing basaltic volcanoes. In this study, we constrain the conditions of degassing (oxygen fugacity, temperature) as rigorously as possible to utilize S isotope compositions of gases and melts to address equilibrium versus non-equilibrium degassing and Earth's sulfur cycle. We find that S degassing is not an equilibrium process and that S partitioning into the gas phase is associated with a kinetic fractionation effect reflected in the S isotope compositions. Erta Ale, a reduced magmatic system, convincingly demonstrates a kinetic effect because S2- (rather than SO42-) is the dominant S species in the melt. Equilibrium degassing should result in preferential partitioning of the heavy S isotope into the gas, however we observe that the gas is isotopically lighter than the melt. This is consistent with faster diffusion of the light isotope. At Masaya, an oxidized magmatic system, the equilibrium and kinetic isotope fractionation effects both favor isotopically light gas, which makes the contributions from equilibrium versus kinetic effects more ambiguous. However, under steady state conditions the gas phase is representative of the sulfur isotope composition of the source and the isotope composition of gas from Masaya indicates recycling of oxidized S through the subduction zone. Mass balance calculations show that only a small fraction of the S subducted at the Central American arc is returned to surface reservoirs. However, the flux of oxidized S from the subducted slab is high enough to rapidly oxidize iron in the mantle wedge, providing an explanation why arc magmas are more oxidized than those at rifts. Finally, the S retained in the subducted slab is isotopically light, potentially carrying the signature of microbial life into the deep mantle. (Abstract shortened by UMI.).