High-Resolution Sampling of a River Plume Front with Uncrewed Underwater and Aerial Vehicles

Abstract Sampling fast-propagating oceanic features is inherently challenging and demands versatile instrumentation and innovative strategies. This paper introduces a novel sampling strategy designed to capture such phenomena, exemplified by a river plume front. Our method revolves around modifying the preprogrammed pathway of an uncrewed underwater vehicle (UUV) to dynamically track and three-dimensionally sample the evolution of the front. To enable the UUV to follow the feature, we adapt the use of a drifting gateway buoy to be positioned and trapped at the front’s convergence zone, allowing underway navigation relative to the buoy. In our demonstration, we showcase the effectiveness of this strategy by successfully conducting over 30 crossings of a river plume front within a 6-h window. The UUV sensors allowed a comprehensive assessment of key front characteristics, including density, velocity, and turbulence. Supplemental drone footage contributed to the overall picture and facilitated the transformation of the dataset into a front-following reference frame. This article provides an in-depth description of the deployment strategy and required postcollection data processing, including frontal crossing detection, the assessment of the frontal orientation from drone footage, and defining the plume bottom boundaries using backscatter intensity contours. Significance Statement Sampling fast-moving ocean features is challenging and requires flexible tools and creative approaches. In this paper, we present a new method for tracking and studying such features, using a river plume front as an example. Our approach involves using a buoy trapped in the plume front, guiding an uncrewed underwater vehicle (UUV) to follow the front in real time while gathering detailed data. In our test, the UUV crossed the river plume front more than 30 times in 6 h, collecting information on water density, speed, and turbulence. Drone footage added context and helped us analyze the data. This paper explains how we carried out the study and processed the data, including how we detected and mapped the front.