Adaptive Tracking of the Barents Sea Polar Front Using an Autonomous Underwater Vehicle

Abstract

Abstract Exchanges and interactions between water masses are typically concentrated across ocean fronts, requiring targeted sampling. As fronts are dynamic in their spatiotemporal extent, they can be hard to map with limited sampling resources. In this paper, we use horizontal temperature gradient as the defining feature of a front to adapt the path of an autonomous underwater vehicle to follow these regions of scientific interest. We present results from simulations and successful field operations in the Barents Sea, where the vehicle equipped with a microstructure sensor crossed the polar front while adapting its trajectory to the front position. Using in situ measurements, we estimate the temperature gradient across the polar front and cross it at different depths. We show that the vehicle is able to estimate and detect the temperature gradient maximum and to adjust its path to provide measurements of temperature, salinity and microstructure along its path, on both sides of the front. This is a step toward the integration of turbulence measurements on autonomous underwater vehicles, showing promise of more targeted sampling of turbulence in regions of large spatial gradients and interactions between water masses, such as fronts. The horizontal vehicle path complements the more traditional vertical profile measurements, providing a finer horizontal spatial resolution. The adaptive behavior of autonomous agents contributes to increased accuracy in targeted measurements as well as expanded resource utilization.