The main idea of Ocean Acoustic Tomography (OAT) is to use the data of (long-range) acoustic propagation to obtain information about the ocean interior. When being implemented in situ this very broad concept splits into many specific methods depending on the technical design of the tomographic system, type of acoustic data to be used, oceanographic value of interest and invesrion technique.
In the narrow sense of the term OAT is a method for estimating the temperature field from the travel times of acoustic rays propagating between moored sources and receivers [1, 2]. Later it was proposed to augment the input data set for inversion by incorporating into it the times-of-flight of acoustic modes which under certain conditions are resolvable in time domain at a single hydrophone . Further generalizations of this approach include peak inversion  and Matched Field Processing in Time Domain ([P. 2] and Sect. 2.2.4 of the present report). However, all these different methods fall into one general category because of common experiment design with broad-band signals travelling between single sources and receivers (or source-receiver pairs - transceivers) anchored on the sea floor. This group of tomographic methods will be further referred to as Traditional Tomographic Technique (TTT). In particular, the major THETIS-2 experiment described below in Sect. 2.1.2 pertains to TTT. Ray, modal and peak inversion are different theoretical methods for reconstructing the sound speed field and may be applied in TTT as well as in other approaches.
An essentially different technique consists in using ship-borne acoustic sources and receivers. Such measurements were utilized in Acoustic Oceanography and in OAT since very early days of these fields. One of the first at-sea experiments on acoustic tomography of oceanic currents was conducted in 1977 by P. Worcester  using the source-receiver pairs deployed from two R/S's. However, in that and some other trials the vessels were positioned at fixed locations throughout the experiment.
The idea to exploit the mobility of ship-borne sources and receivers for improved observations of ocean structure was put forward in . In OAT the vertical resolution of the sound speed field is mainly due to the acoustic multi-paths in each vertical slice, while the resolution in horizontal plane is acquired combining the data obtained from intersecting propagation paths between different moorings, similar to the medical computer-aided tomography (CAT). Though the number of crossing propagation paths grows as the number of moorings squared, it may be insufficient for adequate mapping of ocean mesoscale structure. Detailed numerical simulations of  have demonstrated that Moving Ship (or Dynamic) Tomography (MOST) may dramatically outperform the TTT if the mobile device is used together with a moored tomographic network. Operating in the same area they provide numerous acoustic multipaths in addition to those between the moorings and improve the horizontal resolution of the temperature field.
Both TTT and MOST have their inherent advantages and drawbacks when being compared to each other. At-sea experience in TTT and computer modeling suggest that the TTT:
On the other hand MOST:
The shortcomings of each approach are exactly the negations of the benefits of the other:
To summarize, our present knowledge suggests that traditional OAT suits better for long-term observations of seasonal, yearly, etc., temporal variability of space-averaged oceanographic values while dynamic tomography is more appropriate for mapping mesoscale ocean structure. However, for practical purposes one need quantitative estimates of TTT and MOST performance under the same environmental conditions. The first and most important goal of our investigation is to collect the experimental data necessary for such comparison.
One more method for the OAT implementation is the Matched Field Tomography (MFT), which assumes measurement of a complete sound field (amplitude and phase vs. frequency) at a long array. Possible methods for sound speed field reconstruction includes Matched Field Processing (MFP) and Modal Decomposition with subsequent Matched Mode Processing. Usually MFT uses CW signals and requires long receiving arrays spanning significant part of the water column. The number of array elements should be comparable to the number of high-energy acoustic modes. We will study the possibility to implement MFT, in particular, Modal Decomposition technique with broad-band tomographic signals. If successful, that will allow to use cheaper receiving arrays with fewer elements.
The array for MFT may be either installed on an autonomous mooring similar to TTT or deployed from a R/V as in dynamic tomography. Data acquisition at a long aperture would obviously increase the amount of information available for inversion and enhance capabilities of both techniques.
Mutually complementary properties of above listed approaches imply that they can be effectively combined in a unified observational system. Further analysis of the assembled data will allow to reveal a method of integrating data of these different techniques to improve resolution of the inversion and make recomendations on ways of using dynamic and matched field tomography in future long-term monitoring systems, e.g., recently proposed Global System of the Ocean Observation.