Travel time changes in a tomography array caused by a sea ice cover

Abstract

In September 1988 an acoustic tomography array was deployed in the Greenland Sea in the region 74–76°N, 1°E-5°W. We have modelled travel time changes which may be produced in this array by the presence, or growth, of a sea ice cover. Three types of icefield are considered: a typical East Greenland marginal ice zone (MIZ) thickness distribution, using as input data a sonar profile from a British submarine; a field of discrete floes without pressure ridging; and a pancake-frazils mix growing on the open sea. Firstly we calculated travel time changes caused by an ice sheet uniform thickness. The travel time decreases with increasing ice thickness, by 2–10 ms for 3 m thickness at a frequency of 250 Hz and a range of 200 km, but the magnitude depends on grazing angle and (strongly) on compressional and shear velocities in ice. With some values for velocity the relationship is no longer monotonic. In a real icefield we surmised that the modal ice thickness is responsible for the coherent reflected ray which is detected. Secondly we calculated the progressive travel time decrease resulting from the increase in mixed layer salinity caused by salt rejection from a growing ice cover. We conclude that this is negligible for MIZ-type ice covers, but is detectable for rapidly growing pancake fields and in the case of shallow rays may reach 3.5 ms per day. Finally we calculated the scattering loss associated with ice bottom roughness in MIZ-type icefields, finding that for steep rays it is about 15 dB on average at 200 km range, so that signal-to-noise ratios in such arrays need to be raised to ensure adequate timing precision. If this is achieved, we conclude that acoustic tomography can be used as a means of monitoring modal ice thickness in the Greenland Sea.