AUV survey design applied to oceanic deep convection

Oceanic processes are characterized by both temporal evolution and spatial variability. For surveys carried out using autonomous underwater vehicles (AUVs), compromises between resolution, total survey time, total survey area, and vehicle speed must be made. In this paper, quantitative tools for optimizing surveys are demonstrated in the context of mapping open-ocean deep convection. The survey performance is measured by a metric of total survey error, given by the squared difference between the true and reconstructed field. The convective ocean model is represented by a rectangular volume of water volume. Cooling at the top surface introduces convection which is characterized by spatially distributed convective plumes. The authors focus on reconstructing a synoptic map on the plume scale (1 km) utilizing single or multiple AUVs. It is shown that the most dramatic gain in error reduction come from using multiple vehicles.