Low permeability reservoirs, such as tight gas, shale gas and geothermal reservoirs, must be stimulated by hydraulic fracture to achieve economic levels of production. Hydraulic fractures cause micro-earthquakes whose locations and rupture characteristics, expressed by the source moment tensor (MT), can be monitored for control of the fracture, evaluation of the stimulated volume and design of the next well. Moment tensor inversion (MTI) of direct compressional (P-)waves from such microseismic events can suffer both from lack of aperture of the 1D borehole seismic array usually employed in such observations and from the poorly-known properties of the medium between the event and the sensors. The use of shear (S-)waves as well as P-waves can constrain the MTI, though it is important to understand the effects of differences between P- and S-wave attenuations, and the greater sensitivity of shear travel-times and waveforms to medium anisotropy. The use of clusters of events, from similar locations on the same fracture zone and exhibiting similar waveform characteristics, allows the propagation uncertainty to be minimised by analysing relative MT solutions. Having solved for the MT in a local cluster and assuming similar mechanisms along an extended source region undergoing similar causative events under similar conditions of stress, it may then be possible to invert for the spatial properties of the medium between the sources and the sensors. Both these approaches are difficult with a microseismic dataset from just one or two borehole sensor arrays. Prof. White has high quality datasets from Iceland, covering magmatic- and geothermally-induced events, which are a close proxy for hydraulic fracture events. The datasets have good aperture control from surface seismometer networks and excellent signal to noise ratio. The magmatic events are closely constrained to a melt injection dyke. They are also in a normally ductile region of the crust, which enables simpler analysis than in the brittle shallow crust by removing some of the influence of pre-existing open cracks. There are two populations of events with opposing polarities which are worthy of further investigation in their own right but which can easily be distinguished and used to test the idea of obtaining a better understanding of the medium properties through relative MTI. Some extension of existing computer algorithms, developed in an earlier collaborative PhD project, would be required which would benefit both industry and academic applications. The code and workflow can be tested with accurate, 3D synthetic seismograms generated by 1D and 3D full elastic modelling codes developed by the project partner (SCR )before being used on the high quality Icelandic datasets. The approach would then be tested on a well-constrained industry dataset comprising a combination of many sensors located at the earth's surface, in shallow wells and in deep boreholes. Acquisition of this research dataset is planned by Schlumberger in conjunction with an oil company and would be made available to the student. This would allow the analysis of aperture limitations on the method and exploration of the tradeoff between measurement uncertainty, medium uncertainty and bounds on the volumetric and compensated linear-vector dipole (CLVD) components of the moment tensor. The student would benefit from interaction with both the SCR and university groups, drawing on state-of-the-art data, and modelling and inversion programs from both institutions. This would provide a powerful synergy for new advances both in data processing and in interpretation of the way fluids (be they magmatic, hydrofrac, geothermal or CO2) interact with cracks and fractures in the subsurface. As described in the Knowledge Transfer section, this research would have both economic and societal benefits, and would further strengthen industry-academic collaborations.