High‐Frequency (>2 Hz) Ambient Seismic Noise on High‐Melt Glaciers: Green's Function Estimation and Source Characterization

Using ambient noise to characterize subsurface structures has revolutionized solid earth seismology. In glacial applications, this technique could provide valuable information about ice thickness and bed properties, which to date are deduced from laborious and/or expensive active source geophysics or deep drilling. Despite challenging conditions such as minimal scattering and changing sources, we show that stacks of cross-correlation functions of several hours of ambient seismic noise can converge towards Green's functions under favorable conditions. These contain the sought-after information about subsurface seismic velocities and ice thickness. Applying this technique to two Alpine glaciers in Switzerland, we calculate signal-to-noise ratios of noise correlation functions. Faster Green's functions convergence during elevated ambient noise levels suggests that increased surface melt promotes spatially homogeneous excitation of noise sources. With the help of Green's functions amplitude asymmetry as well as standard beamforming, we locate dominant noise sources. In combination with spectrograms, glacier surface velocities from GPS and proglacial water discharge measurements, we find glacier noise in the range from2to8Hz to be a combination of tremor induced by flowing water, icequakes, and anthropogenic sources. We finally calculate dispersion curves from the cross-correlation functions to arrive at an estimate of the ice thickness. In combination with H/V spectral ratios, the method shown here allows sampling the depth and the properties of the glacier bed by deploying only two seismometers for several hours. Ambient noise correlations could therefore be a valuable ingredient for studies of glacier and ice sheet beds.