Melt-Triggered Seismic Response in Hydraulically-Active Polar Ice: Observations and Methods

Glacier ice responds to environmental forcing through changes in its sliding speed and mass balance. While these changes often occur on daily time scales or longer, they are initiated by brittle deformation events that establish hydrological pathways in hours or seconds and allow meltwater access to englacial or subglacial depths to facilitate ice motion. In this thesis, we (various contributing authors including myself) use seismic monitoring to detect and locate the creation and growth of some of these hydraulic pathways by monitoring their seismic emissions, or icequakes. More specifically, we address (1) what seismic observables, unavailable from other sensing methods, indicate an initial glaciogenic response to melt- water input and (2) if these comprise evidence of feedbacks that may destabilize polar ice under a warming climate. Supplemental to our scientific contributions, we advance statistical processing methods that demonstrably improve the capability of digital detectors at discriminating icequakes from astationary noise.;We begin by interpreting geophysical observations collected from a dry-based, sub-freezing (--17 ° C), polar glacier environment (Taylor Glacier, ANT). By implementing a calibrated surface energy balance model, we estimate the timing and volume of surface meltwater generated during the collection of seismic data from a six-receiver geophone network.;We proceed by contrasting these response characteristics with geophysical observations following an early (spring) supraglacial lake drainage within the lake-forming ablation zone of the Western Greenland Ice Sheet. Using measurements from a ∼5km-aperture geophone network, we find that the anticipated post-drainage icequakes are diurnally responsive, largely surficial in origin, and indicative of tensile fracturing from shallow cracks in the ice. The creation of the lake-drainage moulin appears to coincide with a shift in mean icequake source locations, and an increase in icequake occurrence at night relative to that in the day. Contrary to our expectations, we find that the timing of GPS-derived surface speeds do not clearly indicate this seismic activity on any given day. Rather, these icequakes are best explained by peaks in localized strain gradients that develop at night when decreased subglacial water flux likely increases variability in basal traction. Additionally, our results appear comprise the first detailed seismic observations targeted at an actively draining lake.;Our last study addresses the apparent deficiency in observed basal icequakes detected from Greenland lake site. To explain the lack of deep icequakes, we compute thresholds on the magnitude of detectable basal events within the network and thereby illustrate that surficial icequakes with similar magnitudes and spectral content are more likely to be observed. By restricting our attention to seismic events that produce lower frequency waveforms, we find a population of nearly monochromatic, sub-1Hz, large magnitude ( M w ≤ 3) seismic events borne from remote glaciogenic sources. In contrast to surficial icequakes, these events occur without significant bias between day and/or night periods and are best explained as glacial earthquakes generated by sliding episodes or iceberg calving events in the vicinity of Jakobshavn Glacier. These events occur daily and not correlate with the presence of local, surficial seismicity.;We conclude with three general assertions regarding melt-triggered response characteristics of polar ice. First, hydraulic connections established by fracture events do not necessarily result in seismogenic basal stick slip, and therefore cannot necessarily be observed with conventional GPS monitoring. This was demonstrated at Taylor Glacier. Here, meltwater input to a hydraulic pathway led to fracture growth deep within a cold glacier without any change in surface speed. Second, the presence of melt-triggered basal sliding does not necessarily induce a clear seismogenic basal response in the lakes regions. This was demonstrated on the Greenland Ice Sheet. Seismogenesis may instead be more clearly reflected by surficial strain gradients established by variability in basal traction, suggesting these feedbacks are secondary rather than primary. The response is therefore not clearly indicated from day-to-day timing of GPS-observations. Third, the absence of an observed local response does not necessarily indicate the absence of a local physical response. This was also illustrated in Greenland. Here, deep local icequakes are likely muted by noise, waveform-attenuating ice, and viscous basal rheology. Magnitude thresholds suggest that M w ≤ 2 for consistent recording of local, basal sources. In contrast, remote, low frequency seismic events were clearly observed, and attributed to activity within ice catchments along the western edge of the ice sheet or Jacobshavn glacier.;Finally, we assert that early-indicators of melt-triggered glacial response include components of spatially localized, brittle deformation that is most suitable to seismic observation. Critically-stable regions along mass-balance equilibrium lines constitute potential sites for newly forming surface-to-bed hydraulic connections in a warming climate, and likewise, a potential target for future seismic experiments.