Quantifying iceberg calving fluxes with underwater noise

• Published in: The Cryosphere
• Main Authors: Glowacki, Oskar, Deane, Grant B.
• Format: Article in Journal/Newspaper
• Language: English
• Published: Copernicus Publications 2020
• Subjects: article; Verlagsveröffentlichung; Hansbreen; Svalbard; glacier; The Cryosphere
• Online Access: https://doi.org/10.5194/tc-14-1025-2020; https://noa.gwlb.de/receive/cop_mods_00050949; https://noa.gwlb.de/servlets/MCRFileNodeServlet/cop_derivate_00050606/tc-14-1025-2020.pdf; https://tc.copernicus.org/articles/14/1025/2020/tc-14-1025-2020.pdf

Accurate estimates of calving fluxes are essential in understanding small-scale glacier dynamics and quantifying the contribution of marine-terminating glaciers to both eustatic sea-level rise (SLR) and the freshwater budget of polar regions. Here we investigate the application of acoustical oceanography to measure calving flux using the underwater sounds of iceberg–water impact. A combination of time-lapse photography and passive acoustics is used to determine the relationship between the mass and impact noise of 169 icebergs generated by subaerial calving events from Hansbreen, Svalbard. The analysis includes three major factors affecting the observed noise: (1) time dependency of the thermohaline structure, (2) variability in the ocean depth along the waveguide and (3) reflection of impact noise from the glacier terminus. A correlation of 0.76 is found between the (log-transformed) kinetic energy of the falling iceberg and the corresponding measured acoustic energy corrected for these three factors. An error-in-variables linear regression is applied to estimate the coefficients of this relationship. Energy conversion coefficients for non-transformed variables are 8×10-7 and 0.92, respectively, for the multiplication factor and exponent of the power law. This simple model can be used to measure solid ice discharge from Hansbreen. Uncertainty in the estimate is a function of the number of calving events observed; 50 % uncertainty is expected for eight blocks dropping to 20 % and 10 %, respectively, for 40 and 135 calving events. It may be possible to lower these errors if the influence of different calving styles on the received noise spectra can be determined.