This thesis focusses on shallow intrusions of rhyolite, in the context of heat resources for geothermal development. Rhyolite is the dominant sub-surface magma type at the roots of some of the most vigorous geothermal systems worldwide, including Krafla, in Iceland, and the Taupo Volcanic Zone, in New Zealand. Supercritical fluids located in deeper and hotter environments than those conventionally extracted provide more efficient energy production and greater power output. As potential reservoir, the margins of magma chambers constitute an enticing new target.

The Iceland Deep Drilling Project consortium accidentally drilled magma in 2009 at Krafla, at 2 km depth. This IDDP-1 well was the hottest geothermal well worldwide, with high productive capability (35 MWe). Based on this experience, a new consortium, the Krafla Magma Testbed (KMT), aims to drill back into this magma body for geothermal exploration, creation of research opportunities and monitoring. Improved scientific knowledge of the targeted reservoir is critical to the success of the project, and is the topic of this thesis.

The main objectives of this work are (1) to characterize rock properties at the margins of shallow rhyolite intrusions, specifically around the active Krafla rhyolite magma body, (2) to constrain the size and structure of this intrusion, and (3) to better understand magma response to drilling. Results are obtained via four complementary approaches: Firstly, fossil rhyolite intrusions in Iceland are used as a macro-scale analogy of rock properties around active intrusions. Four field case studies are compared to investigate the host rock response to intrusion, with field measurement of rock strength, permeability and fracture density, as well as textural and lithological information on the intrusions.

Secondly, chilled margins thickness and the extent of impacted host rocks of a further 14 intrusions that emplaced at <1 km depth and are <1 km thick are compared to a numerical model of heat transfer by thermal diffusion. Deviation from the heat diffusion model is discussed to determine the dominant processes of heat exchange at the magma/host rocks interface.

Thirdly, insights specific to the active rhyolite body at Krafla are gained through micro-scale characterisation of magmatic glass and host felsite drilling particles. These were retrieved from the IDDP-1 well as a time-series, after the drilling bit encountered magma. Analyses of texture and geochemistry are combined with previous results from the literature to examine possible scenarios of magma genesis, storage and reaction to drilling.

Fourthly, drilling logs recorded during the IDDP-1 drilling are examined to interpret lithology and rock mass conditions in the blind zone at the chamber margin, from which no cuttings were returned to the surface. Monitored drilling parameters are used to identify rock masses in terms of relative strength, fracture properties and permeability.

The field results reveal that the impact of shallow rhyolite intrusions on surrounding host rocks is dependent on the initial host rocks properties, which in turn influence the style of magma propagation. Initially weak, porous and highly-permeable conglomerate and hyaloclastite respond by pore occlusion, with permeability decreasing accordingly. Their ductile compaction accommodates magma propagation, allowing irregular intrusive geometries to form. In contrast, initially strong, low-permeability basaltic lava and welded ignimbrite respond by brittle deformation, with intense fracturing associated with weakening.

Rhyolitic intrusions in Iceland have variably sized and fractured chilled margins, with a maximum possible thickness of ~7 m. The aureole width in host rocks increases with increasing intrusion size, but aureole formation around < 10 m wide intrusions requires more rapid heat transfer than provided by conduction alone. For these small intrusions, intense fracture development in chilled margins could allow fluid advection and convection to significantly boost the heat transfer.

At Krafla, isotopic compositions and thermodynamic analysis suggest that the rhyolite melts formed by partial melting of felsite possibly triggered by the heat released from a shallow magma intrusion, and promoted by hydrothermal circulation. The presence of lenses of distinct magma composition and texture does not support intense magma convection. The main brown melt generated by high degree felsite partial melting is overlain by a clear melt formed by low- degree partial melting, and which is richer in silica, crystals, and vesicles. As the drill bit approached the magma, it encountered a ~14.5 m thick weak zone, where permeability have been enhanced by thermal cracking of hot rocks in contact with cold drilling fluids. Below this, a stronger ~8 m thick lithology, possibly damaged by the intrusion, separated the magma from the overlying hydrothermal system. Magma then responded to the encounters by limited bubble growth, deformation and nucleation, and caused systematic and consistent reaction of the drilling parameters, revealing it was approached nine times over three drilling attempts.