Acoustical characterization of glacierized fjords

The rapidly changing cryosphere motivates a better understanding of the physical processes governing ice-ocean boundaries. These processes are no more pronounced than in glacierized fjords where massive glaciers meet the ocean. This underwater acoustic environment is significantly louder than other ice-covered environments. The abundant acoustic information, coupled with the difficulty performing measurements on or near glacier termini, encourage the use of passive acoustic monitoring to observe the system. While progress has been made towards improved understanding of sound in glacierized fjords, there remain considerable gaps in the community's understanding of the temporal variability of the acoustic field and the influence of the acoustic propagation environment. To address these deficiencies, three field experiments were conducted in three glacierized fjords: LeConte Bay at the terminus of LeConte Glacier near Petersburg, Alaska; Hornsund Fjord at the terminus of Hansbreen Glacier in Svalbard, an archipelago in the Arctic Ocean; and Disenchantment Bay at the termini of Hubbard and Turner Glaciers near Yakutat, Alaska.

Acoustic data were collected between October 2016 and May 2017 from an underwater hydrophone array moored 500 m from LeConte Glacier. Ambient noise levels (ANL) from recordings were clustered, revealing relationships between ambient noise and the speed of icebergs above the mooring and with calving events. In particular, the local acoustic field demonstrated a uniquely consistent period between late February and early April 2017 where a single cluster dominated observations. The beginning and end of this period coincided with the formation and breakup of a dense pack of icebergs in the fjord.

Characterization of the underside of a brash ice surface was obtained using an inference procedure and time difference of arrival and transmission loss data from an acoustic propagation experiment performed in Hornsund Fjord in September 2017. The inferred surface was incorporated into a forward simulation of the environment using BELLHOP, a ray tracing code. The measured data and simulated results were compared, providing insight to the shape and reflection characteristics of brash ice.

Nearly continuous acoustic data was collected over fifty-two hours in June 2021 on two hydrophone arrays moored in Disenchantment Bay. The effect of recording duty cycle on sampling the spectral and temporal variability of the acoustic field was analyzed by comparing clustered spectral shapes and observations from full and reduced duty cycle recordings of varying length from 1% to 99% of the full 59-minute recordings. A relationship between duty cycle and relative error in hourly cluster observations was determined, which may inform the sampling procedure used in future deployments.