Thin-layer effects in glaciological seismic amplitude-versus-angle (AVA) analysis: implications for characterising a subglacial till unit, Russell Glacier, West Greenland

Abstract

Seismic amplitude-versus-angle (AVA) methods are a powerful means of interpreting the physical properties of subglacial material, although interpreting an AVA response is complicated in the case of a thinly-layered substrate. A layer thinner than one-quarter of the seismic wavelength is considered seismically "thin", and reflections from its bounding interfaces are perceived as a single event. Since a lodged (non-deforming) subglacial till can capped by a thin (metre-scale) cap of dilatant (deforming) till, serious misinterpretations can result if thin layer considerations are not honoured. AVA responses for layered subglacial tills are simulated: we model dilatant layers of thickness 0.1–3.0 m (up to a quarter-wavelength of our synthetic seismic pulse) overlying a lodged half-space, assigning typical acoustic impedance and Poisson's ratios to each. If thin layer effects are neglected, the AVA response to ultra-thin (<1.0 m) dilatant layers yields incompatible physical properties (acoustic impedance and Poisson's ratio indicating, respectively, a low- and high-porosity unit). We show an interpretative strategy that identifies thin layer effects and accurately quantifies the modelled acoustic impedance of lodged till from the composite AVA response. We apply this method to example seismic AVA data from the Russell Glacier outlet of the West Greenland Ice Sheet, in which characteristics of thin layer responses are evident. We interpret a stratified subglacial deposit, with upper and lower layers of high-porosity (<1.0 m thick, Poisson's ratio >0.492 ± 0.015) and low-porosity (acoustic impedance of 4.20–4.39 × 10<sup>6</sup> kg m<sup>−2</sup> s<sup>−1</sup>) material, respectively assumed to represent dilatant and lodged tills. Thin layer considerations are strongly advised wherever seismic AVA analyses are used to quantify subglacial material properties

Borehole measurements indicate hard bed conditions, Kangerlussuaq sector, western Greenland Ice Sheet

“…Our conceptualization is consistent with the thin layer of porous sediment proposed by Booth et al . [] but is not consistent with the existence of a thick underlying layer of lodged till. For example, we expect that our 1000 psi water jet would have easily penetrated into the top of a lodged till, as was apparently the case with the drilling conducted by Ryser et al .…”

“…Booth et al . [] used seismic amplitude versus angle methods to characterize bed reflections at a site south of our transect and ~70 km inland. The acoustic impedance was indicative of a lodged till deposit with low porosity, but Poisson's ratio suggested a highly water‐saturated substrate.…”

“…Borehole drilling appeared to confirm a soft bed in this area [ Ryser et al ., ]. Farther south in the Kangerlussuaq sector two seismic studies also suggested that till exists at the ice/bed interface [ Booth et al ., ; Dow et al ., ], although neither study employed methods capable of determining the sediment layer thickness.…”

Active seismic studies in valley glacier settings: strategies and limitations

“…We assume that the glacier ice has a uniform seismic velocity, that there is no firn and that the subglacial material is uniform and thicker than 1/4 of the seismic wavelength (~10 m at the center frequency of 100 Hz), so that thin-layer effects do not distort the reflection wavelets (Widess, 1973). In reality, dilatant till layers are often thinner (Iverson and others, 1995; Porter and Murray, 2001; Evans and others, 2006; Reinardy and others, 2011; Booth and others, 2012), but a consideration of thin layer effects is beyond the scope of this paper.…”

“…AVA analysis requires that the primary reflection from the bed have a sufficiently high signal-to-noise ratio (SNR) to allow estimation of the polarity of the wavelet and to measure its peak or root mean square (RMS) amplitude. In addition, AVA analysis is more robust if the bed reflection multiple can be identified and measured (Peters and others, 2006, 2007, 2008; Peters, 2009; King and others, 2008; Booth and others, 2012; Christianson and others, 2014; Luthra and others, 2016). Uncertainties in bed dip and strike at the reflecting point map to uncertainties in incidence angles.…”