Geological applications of seismic full waveform inversion: Insight from a novel approach to evaluating the seismic properties of ice

Seismic surveys give insight into the physical properties of the englacial and subglacial properties of ice masses. Certain quantities, e.g. ice density can be related to climate models and therefore provide proxies for the climatic evolution of a glacier (Kuipers Munneke et al., 2014; Hubbard et al., 2016). Such information is critical for predicting the stability of ice masses in a warming climate; these ice masses include the large ice shelves which fringe the Antarctic continent, widely believed to underpin the long-term contribution of Antarctic ice streams to global sea-level rise (DeContro and Pollard, 2016).

A major limitation of deriving density from seismic velocities is that velocity:density conversions are almost entirely empirical and therefore of questionable accuracy (Booth et al., 2013). The widely-applied Kohnen (1974) conversion is based on lab analysis of a South Pole ice core hence ice is removed from its original temperature/pressure context which, in any case, is likely different for other ice masses. While the general trend of density is likely characterised, greater confidence in the accuracy of inverted parameters would be beneficial. Full wavefrom inversion (FWI) of a seismic dataset (Virieux and Operto, 2009) represents a promising approach to property estimation that deserves investigation for glaciological applications.
FWI methods circumvent the need for empirical methods by deriving the underlying physical properties (seismic velocity, density, etc.) from the seismic wavefield itself. FWI has seen significant development in the hydrocarbons field (Brittan et al., 2013; Bai and Yingst, 2014; Jones, 2015), but its applicability in the glaciological setting is not widely proven. The development of FWI in glaciology would represent a step-change in seismic capabilities, which could be widely adopted throughout the community.
This studentship aims to explore the transfer of FWI experience between the industrial and glaciological settings, in a three-phase research programme:
Phase 1) using existing glaciological archives of seismic data, review whether current seismic practice can produce data that are FWI-compatible;
Phase 2) develop an optimised acquisition strategy to enable FWI;
Phase 3) undertake field acquisitions to validate FWI performance.
Data available for Phase 1 include seismic and density measurements made on Antarctica's Larsen C Ice Shelf, during field campaigns of the NERC-funded MIDAS project (Kulessa et al., 2016; Hubbard et al., 2016; Ashmore et al., 2016); additionally, the British Antarctic Survey (BAS) will contribute data from their campaigns on the Antarctic Pine Island Glacier.
During Phase 2, BAS will liaise with the student for guidance on FWI-compliant acquisitions in forthcoming field campaigns, which include a further deployment on Thwaites Glacier and on Rutford Ice Stream. Here, seismic velocities have important implications for the thermal regime and internal crystal fabric.
In Phase 3, the student will undertake FWI-compliant seismic acquisitions to establish the evolving density of the Hardangerjokulen ice cap (Giesen and Oerlemans, 2010). There may also be the opportunity, via the Collaborative Antarctic Science Scheme (CASS) for acquisition in the environs of BAS's Antarctic Rothera Station. These new acquisitions will be accompanied by borehole measurements of sonic velocity and density to calibrate the inverted parameters, and compare their accuracy to standard interpretative approaches.
It is also expected that experiences of FWI in glaciology can benefit industrial approaches. For example, glaciers are structurally simple compared to the geology of a typical hydrocarbon province; furthermore, where borehole control is available, it extends from the ground surface to the target depth therefore offering depth-continuous constraint.