The MF Tomography experiment made use of 3 MIT/WHOI vertical  line arrays and the associated data acquisition system to acquire data in the 200-800 Hz  band for geoacoustic inversion. In the experiment, impulsive sound sources (light  bulbs) and a low frequency CW projector (for multi-tones and wid

(R. Chapman, M. McDonald, R. Kessel, L. Jaschke, V. Corre and P. Pignot)

     The MF Tomography project is based on a collaborative  experiment with Massachussetts Institute of Technology (MIT)/Woods Hole Oceanographic  Institution (WHOI) that was carried out in Haro Strait B.C. during June-July 96.   Partial support for this project was provided by a grant from the Office of Naval  Research. Geoacoustic matched field tomographic inversion is a relatively new approach  that is specifically designed for rapid, high resolution estimation of ocean bottom  properties. The new tomographic approach based on MF processing extends the MF  inversion method to include full 3-D anisotropic environments, i.e. to include  variability in depth, range and cross-range. This approach is especially well suited  for shallow water environments where conventional ocean acoustic tomography based  on travel times (i.e. phase) information is difficult to apply, because individual  ray paths cannot always be resolved at longer ranges where the signal is modal in character.

     The experiment was carried out successfully during 17-21  June 1996, using the research vessel CSS R.B. Young. An extensive set of low frequency  acoustic data were collected at all three VLAs for all the three types of sources.   The ground truth data for the bottom environment have been processed to obtain bathymetry  in 5-m contours, and a map of the surface sediment properties of the region.

     The use of light bulb implosions as sound sources in  this experiment has attracted considerable attention from researchers in ocean  acoustics. Although light bulb sound sources are not new, there is a current need  to find replacement for explosives as sound sources, due to environmental impact  regulations for the effect on marine mammals. The Group developed a simple apparatus  that was used in the experiment for controlling the breaking depth of the light  bulb. In a collaborative project with Dr. G. Heard at DREA, the light bulb data  were processed to obtain accurate measures of the acoustic source levels, and the  work was presented at the Oceans 97 conference in Halifax in October (Heard et al, 1997).

     Initial processing for geoacoustic inversion concentrated 
on the light bulb data set (Chapman et al, 1997). These data provide a means for 
estimating the experimental geometry as well as the bottom properties. In the  inversion method, waveform matching was applied to data from a single VLA to estimate  local geoacoustic properties. This project is the M.Sc. thesis of Mr. L. Jaschke (Dec  97). His thesis investigated a new method of global search in matched field inversion  that is based on simulated annealing, using the heat bath algorithm. Instead of cooling  to a final solution, the new approach, known as 'freeze bath inversion', cools only to  around the annealing temperature. At this temperature, it is possible to sample the  model parameter space to obtain a representation of the a posteriori distribution of  geoacoustic models that fit the measured acoustic field data well. The product of the  global search is thus an estimate with a statistically meaningful confidence limit. The  method also proposes a simple means of reparameterization to improve the efficiency of 
the search. A paper describing the method was published in the Journal of the  Acoustical Society of America (Jaschke and Chapman, 1999).