Knowledge of the provenance of crustal volatiles and the processes by which they are released is extremely important for the dynamics of magmatic systems. Presented here are the results of multiple investigations, which aim to understand magmatic volatile contamination from contrasting but complementary perspectives. The main methodologies used include He and C isotope values and CO2/3He ratios of volcanic gases and fluids; simulation of magma-carbonate interaction using high-pressure high-temperature experimental petrology; X-ray microtomography of vesiculated xenoliths and computer modeling. Findings show that the contribution from upper crustal volatiles can be substantial, and is dependant on the upper crustal lithology on which a volcano lies, as well as the composition of the magma supplied. Carbonate dissolution in particular is strongly controlled by the viscosity of the host magma. The details of the breakdown of vesiculated xenoliths is complex but has wide reaching implications, ranging from the dissemination of crustally derived materials through a magma body to highlighting that crustal volatiles are largely unaccounted for in both individual volcano and global volatile budgets. In synthesizing the conclusions from each of the individual perspectives presented, I propose the contribution of volatiles from crustal sources to play a significant role in many geological systems. This volatile component should be taken into consideration in future research efforts.

Xenoliths preserve evidence of magma-crust interactions in magmatic reservoirs and conduits. They reveal processes of partial melting of country rock, and disintegration into magma. Widespread evidence for frothy xenoliths in volcanic deposits exists, and these evidently indicate processes of gas liberation, bubble nucleation and bubble growth. This report focuses on textural analysis of frothy sandstone xenoliths from Krakatau in Indonesia, Cerro Negro in Nicaragua, Cerro Quemado in El Salvador and from Gran Canaria, Canary Islands, and involves attempts to experimentally reproduce xenolith textures. To achieve this, magmatic conditions acting upon country rock in volcanoes are simulated by subjecting sandstones to elevated temperature and pressure in closed system-autoclaves. Subsequent decompression imitates magma ascent following xenolith entrainment, and is largely responsible for the formation of frothy xenolith textures. The experiments show a range of successive features, such as partial melting, gas-pressure build up, bubble nucleation, growth and development of bubble networks. The experiments closely reproduced textures of natural xenoliths and help to assess the controlling P-T parameters that encourage efficient bubble growth. Conditions proved ideal between 850˚C and 870˚C and pressure release from 1 kbar. Such conditions limit bubble overprinting by secondary crystallization and melt infilling. Country rock lithology proved vital regarding gas pressure build-up and resulting bubble nucleation during decompression. In particular, increased water content and relict crystals in the melt produced appear to ease and promote gas liberation by enabling early and effective bubble nucleation. Moreover, experiments confirm a decisive role for bubble coalescence. These results attest to the great potential of country rock to develop interconnected bubble networks upon magma contact, exsolving large amounts of crustal volatiles into the magma. Volatile input involves a change in magma viscosity and thus an accompanied change in disruptive behaviour, and may hence be responsible for increased potential to cause explosive volcanic eruptions. Moreover, H2O and CO2 vapour are severe greenhouse gases, which seems to be added to the atmosphere from crustal rocks via recycling by volcanic activity, and may have yet underappreciated effects on Earth’s climate.

The focus of this work is magma-crust interaction processes and associated crustal volatile release in subduction zone volcanoes, drawing on rock, mineral, and gas geochemistry as well as experimental petrology. Understanding the multitude of differentiation processes that modify an original magma during ascent to the surface is vital to unravel the contributions of the various sources that contribute to the final magmas erupted at volcanoes. In particular, magma-crust interaction (MCI) processes have been investigated at a variety of scales, from a local scale in the Vesuvius, Merapi, and Kelut studies, to a regional scale, in the Java to Bali segment of the Sunda Arc.  The role of crustal influences is still not well constrained in subduction systems, particulary in terms of the compositional impact of direct magma crust interplay. To address this shortcoming, we studied marble and calc-silicate (skarn) xenoliths, and used high resolution short timescale experimental petrology at Vesuvius volcano. The marbles and calc-silicates help to identify different mechanisms of magma-carbonate and magma-xenolith interaction, and the subsequent effects of volatile release on potential eruptive behaviour, while sequential short-duration experiments simulate the actual processes of carbonate assimilation employing natural materials and controlled magmatic conditions. The experiments highlight the efficiency of carbonate assimilation and associated carbonate-derived CO2 liberated over short timescales. The findings at Merapi and Kelut demonstrate a complex magmatic plumbing system underneath these volcanoes with magma residing at different depths, spanning from the mantle-crust boundary to the upper crust. The erupted products and volcanic gas emissions enable us to shed light on MCI-processes and associated volatile release in these systems. The knowledge gained from studying individual volcanoes (e.g., Merapi and Kelut) is then tested on a regional scale and applied to the entire Java and Bali arc segment. An attempt is presented to distinguish the extent of source versus crustal influences and establish a quantitative model of late stage crustal influence in this arc segment. This thesis therefore hopes to contribute to our knowledge of magma genesis and magma-crust interaction (MCI) processes that likely operate in subduction zone systems worldwide.