Nj. Bruguier et Ta. Minshull, ACCURATE MODELING OF SONOBUOY REFRACTION DATA TO DETERMINE VELOCITY VARIATIONS IN OCEANIC-CRUST, Marine geophysical researches, 19(1), 1997, pp. 25-36
Modern disposable sonobuoys can provide a simple and cost-effective al
ternative to ocean bottom seismometers for marine refraction experimen
ts over oceanic crust. Unfortunately, the fact that they are free to d
rift with the prevailing ocean currents can introduce significant trav
el-time errors into the modelling process if the seafloor topography i
s large. For sonobuoys recorded during and after turns the drift rate
and direction can be uniquely determined by inversion of the shot-rece
iver ranges derived from the water-wave arrival. The same method can b
e used to determine a best fitting average drift vector for the whole
dataset. A modification to conventional two-dimensional travel-time mo
delling techniques has been developed to account for this drift. Each
sonobuoy profile is divided into several subsets, typically of 100 sho
ts each, and each subset is then modelled as a separate common receive
r gather, significantly reducing the errors in the calculated travel-t
imes. For realistic bathymetry, the magnitude of these travel-time err
ors is up to 200 ms, significantly larger than the estimated picking u
ncertainty. Real data from a typical sonobuoy refraction experiment on
the Mid-Atlantic Ridge were modelled with and without the drift corre
ction applied. Much of the lateral variation in the velocity structure
was removed when the drift correction was applied, indicating that th
is structure was due to variations in the travel-times caused by sonob
uoy drift.