During the 1984 Marginal Ice Zone Experiment (MIZEX '84), a 224-Hz center frequency acoustic tomography sound source was deployed off Spitsbergen, Norway, in 1200 m of water. Over the course of 10 days it transmitted hourly signals that were recorded by receivers at two locations, one moving with the ice pack and the other stationary. The main purpose of the transmissions was to answer the following questions which are important to the feasibility of acoustic tomography in marginal ice zone (MIZ) and Arctic regions: (1) Can the signals be coherently averaged to improve the signal-to-noise ratio of the multipath arrival peaks, (2) are the surface-scattering and bottom reflection losses acceptably low, (3) are the arrivals identifiable, (4) are the arrivals stable, and (5) are the arrivals resolvable? Our data analysis provides generally positive answers, indicating that ocean acoustic tomography is feasible in MIZ regions. However, the shallow bathymetry in the MIZEX '84 area increases the difficulty of the problem, consistent with the observations of Palmer et. al (1985) and DeFerrari and Nguyen (1985) in the shallow water of the Straits of Florida.

In analyzing the data, it was noticed that the phase (or travel time) fluctuation spectra of surface interacting rays exhibited a shape similar to that of the surface wave frequency spectrum. This suggested that tomography could also measure the surface wave field, a previously unexplored possibility. Theoretical expressions are derived in this paper which relate the acoustic fluctuations to standard surface wave descriptors, e.g., the frequency-directional spectrum. These expressions are then used to compute the surface wave frequency spectra for 2 days of data. Though these spectra are uncorroborated by other wave spectrum measurements (e.g., wave rider buoys), the rms wave heights predicted by the acoustics agree well with rms wave heights estimated by wind force data. ¿ American Geophysical Union 1987