|
P.F.J. Lermusiaux, P.J. Haley, Jr., W.G. Leslie, O. Logutov Massachusetts Institute of Technology
|
Example PN06 and PN07 results |
|
PLUSNet Real-Time Sea Exercisxes Activities and Findings Presentations Ongoing MIT research |
|
|
P.F.J. Lermusiaux, P.J. Haley, Jr., W.G. Leslie, O. Logutov Massachusetts Institute of Technology
|
Example PN06 and PN07 results |
|
PLUSNet Real-Time Sea Exercisxes Activities and Findings Presentations Ongoing MIT research |
|
For the Focused Acoustic Forecasting-05 (FAF05) real-time at-sea field exercise (Wang et al, 2006, 2008), the emphasis was on methodology development and engineering tests. Our results are described in the FAF-05 web site. In preparation for the 2006 real-time experiment, virtual experiments were conducted in which model output fields were utilized, for one example, to determine which bearing should be chosen for the MIT-AUV and which yoyo pattern should it follow along that bearing, so as to best sample the environment and optimize acoustic performance, including reduction of acoustic uncertainties. Acoustic predictions and optimization of the AUV path along the 2D bearings were carried out by D. Wang, Prof. H. Schmidt (MIT) and the PI using RAM, with sound-speed field inputs from our HOPS predictions.
Environmental and acoustical nowcast and forecast products were distributed daily for the PLUSNet real-time exercise within the MB06 experiment during the period 17-28 Aug. 2006 via the exercise web site. The field and graphical products included: currents and sound speed at 0, 30 and 100m; currents averaged over 0-100m and 0-200m; and sound speed and transmission loss for a source at 5m and a receiver at 15, 45 and 75m along standard sections. In addition to the twice-daily plots, hourly model output data files were distributed. Transmission loss calculation data files were provided in Matlab format. Daily summary description emails to the PLUSNet team included: environment descriptions, adaptive sampling recommendations, estimates of transmission loss changes caused by the varying environment (e.g. relaxation conditions correspond to 3 to 8db more mean loss over 7 to 15km at 100Hz and 400Hz than upwelling conditions) and waypoints for adaptive yo-yo (AREA mission C2) and front tracking (AREA mission C4) missions. Several PLUSNet team members used our forecasts of surface currents in the planning of drifter missions. Subsequent to PLUSNet-MB06, novel research was carried out in acoustical-physical adaptive sampling (Yilmaz et al, 2006a; Lermusiaux et al, 2007a; Wang, 2008; Yilmaz et al, 2008) and uncertainty predictions and reduction by data assimilation (Lermusiaux, 2006; Lermusiaux et al, 2006b).
During PN07, twice-daily sets of real-time, tidally-forced nowcasts and forecasts of temperature, salinity, velocity and full-field acoustic transmission loss (TL) at 3 frequencies and 3 source depths were issued. Data was analyzed and quality controlled on a daily basis throughout the experiment. Our web-page, was used for distribution of graphics and all model data files. Forecasts included the: i) assimilation of ship and glider data; ii) external forcing by MIT-developed high-resolution barotropic tides and by NOGAPS atmospheric flux predictions; and, iii) the coupling of ocean-acoustic models: the free-surface HOPS and C-Snap. In collaboration with J. Curcio (MIT), kayak missions were adaptively planned to: i) examine the internal wave activity in Dabob Bay; ii) collect data for glider/ship CTD cross-calibration; iii) sample modes of upper layer variability; and, iv) sample the upper layer response to wind events. Importantly, one of the kayaks was directly controlled at sea from instructions downloaded on a web-page built here at MIT. A tidal model was developed which provided estimates from a generalized-inverse solution of the shallow water equations, in three nested multi-scale domains over high resolution bathymetry. Tidal estimates obtained from tide gauges were assimilated to correct for errors in the open boundary conditions of the three nested domains. Our PN07 results were summarized in the hotwash-up presentation (Lermusiaux, et al., 2007) and included oceanographic and acoustic findings (manuscript in prep) which are summarized next.
PN07 Physical Oceanography Findings: Dabob Bay was found to be more similar to a fjord than to the open-ocean. This was challenging but also interesting from a modeling viewpoint since as there is little or no shelf in Dabob Bay, the pycnocline immediately intersects steep topography and, therefore, the model sigma levels. This was significant because in Dabob Bay, salinity is the primary determining factor in density and temperature acts more as a passive tracer. Internal tides and atmospheric forcing were found (and forecast) to be important for cross-bay currents, sub-surface cold layer dynamics and the deep cold layer of Dabob Bay. Four layers of temperature were identified: the summer surface thermocline, a cold sub-surface (20m) layer, and intermediate layer, and a cold bottom layer.
PN07 Acoustics Findings: Seabed bottom properties were found to be a critical factor in estimating TL, especially for higher frequency. We observed that the HFEVA model has much more bottom attenuation (more acoustic loss) than the range-independent silt-clay bottom. This is more significant at 900Hz than at 100Hz, but it is apparent at both frequencies. Water column fluctuations, both from internal tides and atmospheric forcing, were predicted to impact TL performance. These TL predictions showed the most sensitivity to fluctuations in the near-surface thermocline and the 20m cold layer, at both 900Hz and 100Hz.
