S.M. Jesus, C.Soares, N. Martins sjesus@ualg.pt , csoares@ualg.pt , nmartins@ualg.pt
SiPLAB-FCT, Universidade do Algarve
Campus de Gambelas, PT-8005-139 Faro, Portugal.
Abstract : Rapidly assessing the environmental conditions of
a given coastal area with the capability of being able to predict its
evolution in the next 24 or 48 hours has been the goal of many
initiatives since the end of the cold war and the shift of strategic
regions to shallow areas. Most efforts were carried out by
oceanographic teams feeding data of various nature (currents, SST,
temperature, altimetry, wave height, etc...) into small scale
circulation models (such as mini HOPS and NCOM). Testing has been going
on for several years on the validation of such models in various
scenarios. Among others, the goal of this testing is to decrease the
error variance of various environmental parameter predictions at 1, 2
or 3 days with a minimal model initialization. For tactical purposes it
was implicitly assumed that a decrease of the error variance on the
environmental parameter predictions would necessarily imply a reduction
of the mean square error of the acoustic detection probability
prediction. This can be accepted as a general trend but can not be
guaranteed since the relation between the temperature profile and the
acoustic output is non linear. With the appearance of Matched Field
Tomography (MFT) in the late 80's and its application in shallow water
scenarios in the early 90's, a new era on the usage of acoustics for
environmental purposes was opened. A large number of experimental
results obtained by various teams using MFT in shallow water regions
have shown the influence of internal tides and internal waves, the
strong influence of bottom properties on signal propagation and the
crucial role of source receiver geometry and bathymetry on the
inversion process. Since even the most sophisticated propagation model
existing today (and presumably in the years to come) can not take into
account all real world details, in many situations the outcome of the
inversion process that most closely matches the acoustic signal is an
environment often ``slightly shifted'' from the actual measured
environment, thus the concept of equivalent
acoustic model. At this point an acceptable goal would be to use
the present equivalent acoustic model together with the oceanographic
model predictions to obtain more accurate prediction of the acoustical
field at future times. As a matter of fact the problem is not so simple
since the environmental shift of the inverted data may not be solely
due to acoustic modeling issues but is also due to errors in the
inversion process and noise, which are effects often difficult to
separate. Using the acoustic and environmental data gathered during the
MREA'03 sea trial, this paper intends to put in evidence the
differences between the inverted and the measured environmental
quantities - mainly water column temperature profiles - and the usage
of the Acoustic Oceanographic Buoy (AOB) as a promising tool for
operational environmental assessment of small coastal areas.
ACKNOWLEDGMENT: this work was partially supported by FCT under programs POSI and POCTI.