Investigating the basal and englacial properties of a West Antarctic ice rise with novel active-source seismic methods

Seismic observations of glacier beds are key to understanding processes of basal slip and incorporating these processes into ice sheet models, which in turn inform predictions of global sea level rise. Observations of seismic reflection amplitude are a powerful tool for identifying glacier bed materials. Amplitude versus-angle (AVA) analysis is a technique commonly used to identify glacier substrates whereby the amplitude of the basal reflection is measured as a function of its incidence angle at the ice base. Glaciological AVA experiments conventionally consider only the compressional (P) wave component of the wavefield, ignoring the shear wave (S) component; however, three-component seismic acquisitions are proliferating in the glaciological community. To harness the full potential of three-component recording, analysis of PS converted waves (incident P waves converted to S waves at the glacier bed) is necessary. This thesis presents an investigation into the glaciological application of joint PP and PS AVA inversion. Prior to inverting AVA data, amplitudes must be corrected for attenuation losses. The transition of snow to firn and glacial ice represents a challenge to seismic study due to the continuous transition in elastic properties with depth. I describe a method for measuring the seismic quality factor, Q, which enables more detailed characterisation of firn’s attenuative structure than previous approaches allow. Q increases from 56˘23 in the uppermost firn to 570˘450 between 55 and 77 m depth. This method offers a strategy of constraining attenuation in seismic reflection experiments which do not record multiples, also enabling improved constraint of source amplitude when compared with conventional methods. I present an inversion scheme which jointly inverts PP and PS AVA data for the properties of the icebed interface. Using synthetic AVA data, I investigate the improvement joint inversion of PP and PS amplitudes makes to constraint of bed properties when compared with PP inversion. In general, joint inversion improves upon PP inversion in both precision and accuracy over the same angular range. In many cases joint inversion of data over 0 ´ 300 performs favourably with single inversion of data over 0 ´ 600 . Joint inversion therefore has the potential to reduce ambiguity in substrate identification and reduce the logistical requirements of glaciological AVA surveys. The inversion scheme is applied to PP and PS data from Korff ice rise (KIR), in the Weddell Sea sector of West Antarctica. Analysis of PP and PS AVA responses at KIR shows the reflection to arise from a material with a P wave velocity of α “ 4.03 ˘ 0.05 km s´1 , an S wave velocity of β “ 2.16 ˘ 0.06 km s´1 and a density of ρ “ 1.44 ˘ 0.06 g cm´3 . The inverted properties are consistent with a reflection from a layer of basal debris overlying frozen sediments, with a poorly-defined boundary between the two. I propose that this results from a previous episode of flow as an ice rumple, followed by grounding on the lee side of the bathymetric high occupied by KIR and subsequent freezing of basal sediments. The indication of a reflection from a basal debris layer raises questions about whether conventionally-interpreted basal reflections can truly be considered as such, and whether these interpretations may mask the true nature of the underlying subglacial material.

Details