Fusion of Measured and Synthetic 
Sound Speed Profiles: A Viable Technique 
to Improve the Accuracy of Multibeam 
Bathymetry? 
Jean-Guy Nistad | Patrick Westfeld 
1 Introduction 
An important error source in multibeam echo sounder depth measurements in 
shallow and highly stratified water column environments is the sound speed er- 
ror. The impact of this error in hydrographic data acquisition and processing 
is potentially two-fold. First, an incorrectly measured sound speed value at the 
echo sounder’s transmit and receive arrays will affect their directivity pattern 
steering angles and lead to an erroneous array-relative beam pointing angle. 
Second, an incorrectly measured water column sound speed structure will not 
account for the true path travelled by the acoustic pulse, because it is affect- 
ed by a propagation error in distance and direction. These two impacts lead to 
similar yet distinct error signatures, which in both cases propagate to the final 
3D point solution referenced in a terrestrial reference frame. While a calibrated 
sound speed probe mounted at the antenna arrays should preclude the first error 
source, an undersampled water column relative to a changing sound speed struc- 
ture in both time and space, inevitably leads to so-called refraction errors. Given 
that in-situ knowledge of the water column dynamics is difficult to assess while 
surveying and in order to optimize survey time, a common approach is to use 
an underway profiling system to increase the number of measured sound speed 
profiles. However, excessive profile measurements leading to premature wear of 
the profiling system is often the consequence. As such, an optimal sampling pe- 
riod that balances cost with survey specifications should always be sought. In 
this work, we seek to determine the gain in accuracy achievable by combining 
measured with synthetic sound speed profiles to minimize the impact of sound 
speed errors in multibeam surveys conducted in the shallow waters of the south- 
western Baltic Sea. 
Previous work has demonstrated the added value of applying synthetic sound 
speed profiles in the absence of measured profiles for multibeam ray-tracing pur- 
poses. These synthetic profiles may be derived directly from regional hydrody- 
namic models, from spatiotemporal interpolation, or from a combination of the 
U Wilstier-Vella 
3and 102/207) a 
DVW-SCHRIFTENREIHE 
51