Feature-oriented regional modeling and simulations for acoustic prediction in the Cabo Frio upwelling system: forecasting validation

AUTHORS: Codato, G. (1,2); Calado, L. (1); Martins, N. (3); Watanabe, W.B.(4); Domingues, R.M.(5)

1) IEAPM - Brazilian Navy
2) CEM - Federal University of Parana
3) UAlg - University of Algarve
4) IOUSP - University of São Paulo
5) NOAA - National Oceanic and Atmospheric Administration

Acoustic predictions usually suffer from uncertainties in ocean forecasts due to the sensitivity of acoustic propagation to the ocean mass field. For this reason, acoustic prediction systems require the best possible specification of initial conditions, demanding high accuracy and synopticity of the ocean circulation modeling. This study assesses the feasibility of applying a synoptic initialization scheme by a feature-oriented regional modeling system (FORMS) for acoustic prediction in the Cabo Frio coastal upwelling area. We employed a coupled oceanographic-acoustic modeling system using ROMS (4D-ocean model) and BELLHOP (2D-acoustic model) to forecast the acoustic field. ROMS is initialized with the output of FORMS, which consisted of a coastal upwelling parametric feature model with a background climatological thermohaline structure to create a non-dimensional 3-D field. This field is then re-scaled using high-resolution SST satellite data (GHRSST). This feature-oriented oceanographic forecast model is tested for acoustic applications. Two numerical acoustic simulations were performed using different initial conditions: (i) in situ hydrographic data from the OAEx10 cruise and (ii) ROMS output. The simulations were compared in terms of transmission loss (TL), detection probability (DP) and impulse response. The TL differences exhibit standard deviations ranging between 2.29 and 4.32 dB. These SDs measure the skill of the feature-oriented ocean model for sonar applications. An interesting result is that coastal upwelling may prevent the detection of submarine targets. The simulations using the ocean forecasts have produced a satisfactory spatial distribution of the DP zones, and agree well with the simulations initialized by the in situ data. However, the quality of the results decrease with distance, as observed in correlations between the impulse responses. This can be explained by an accumulation of forecast error during propagation. Results indicate that a realistic representation of the coastal upwelling on the sonar range is essential for tactical guidance. Inclusion of the upwelling feature in ROMS initial conditions generated a fitted oceanographic field, which agrees very well with the observed in situ structure. Finally, an accurate prediction of the acoustic field can be accomplished using a FORMS technique, considering that the feature-oriented ocean forecasts provided a realistic representation of the oceanic variability.