|
P.F.J. Lermusiaux, P.J. Haley, Jr., HARVARD UNIVERSITY Division of Engineering and Applied Sciences Department of Earth and Planetary Sciences Cambridge, Massachusetts
|
|
|
P.F.J. Lermusiaux, P.J. Haley, Jr., HARVARD UNIVERSITY Division of Engineering and Applied Sciences Department of Earth and Planetary Sciences Cambridge, Massachusetts
|
|
The Focused Acoustic Forecasting-05 (FAF05) real-time at-sea field exercise was part of the Persistent Littoral Undersea Surveillance Network (PLUSNet) program, sponsored by the Office of Naval Research. The emphasis was on methodology development and engineering tests. Our joint Harvard-MIT objectives for FAF05 were to:
For the duration of the FAF05 experiment, ocean enviromental fields and uncertainties were predicted daily by the Harvard Ocean Prediction System and the Error Subspace Statististical Estimation approach. The data used for initialization and assimilation via Optimal Interpolation were satellite sea surface temperature snapshots and synoptic profiles of ocean temperature and salinity. To estimate environmental uncertainties, various scenarios were computed daily as a function of different initial condition estimates, assimilation procedures, modeling domains, numerical/physical model parameters and time of day. Results are illustrated here.
The various sound speed field predictions (in time and 3D space) were interpolated along several characteristic vertical sections and used for acoustic predictions with the Range-dependent Acoustic Model (RAM). The ensemble of sound-speed sections and the corresponding ensemble of acoustic transmission loss fields were utilized as input to an optimization algorithm that estimated the optimal parameters of the AUV sampling patterns for the next day(s). These optimal sampling parameter estimates and the corresponding environmental and acoustical predictions were emailed daily to the FAF05-MIT team at-sea aboard the R/V Leonardo. They provided the basis for coupled physical-acoustical adaptive sampling by the MIT AUVs, aiming for optimal surveillance in the region.
The ocean physical data-driven simulations and sound-speed estimations were carried out by Pat Haley and Pierre Lermusiaux. The ocean data and forcing were collected and prepared by Wayne Leslie. Acoustic simulations and sampling pattern optimizations were carried out by Ding Wang at Harvard. Henrik Schmidt led the FAF05 experiment, including the ocean sampling at sea. The modeling and real-time simulations at Harvard were under the guidance of Pierre Lermusiaux.
We thank Prof. Allan Robinson for inputs and support. We are grateful to E. Coelho, E. Nacini and A. Cavanna from NURC for the satellite and Alliance in situ data, Martina Tudor from the Cro. Met. Service for the Aladin atmospheric forcing and FNMOC for their real-time atmopsheric forcing. We thank Dr. Mitchell N. Shipley (ARL Penn State) and Dr. Emanuel Coelho for comments and suggestions on this web-site.
Ongoing HU-MIT collaborative PLUSNet research for FAF05 include: (i) Continue to improve the algorithms/software, and, (ii) Evaluate the quality of predictions and impact of adaptive sampling based on the available data. Some real-time evaluations are available from the 2 results links above.
