Evaluation of angle of attack of underwater glider using an electromagnetic current sensor and ADCP

Takahiro Tanaka1, Daisuke Hasegawa1, Takeshi Okunishi1, Ichiro Yasuda2 and Pat Welch3, (1)Tohoku National Fisheries Research Institute, Japan Fisheries Research and Education Agency, Shiogama, Japan, (2)Atmosphere Ocean Research Institute, University of Tokyo, Kashiwa, Japan, (3)Oregon State University, College of Earth, Ocean, and Atmospheric Sciences, Corvallis, OR, United States
Underwater Gliders have become a common oceanographic instrument to observe broadscale spatiotemporal variability of ocean phenomena such as internal waves, meso-scale eddies, and boundary currents. One of the recent advancements of oceanic technology has enabled gliders to measure turbulence intensity in the ocean interior. For accurate measurements of turbulence intensity, usually expressed as turbulent kinetic energy dissipation rate, 𝜀, it is necessary to obtain an accurate estimate of the glider’s axial velocity, Vg, while it is descending/ascending through the water column. The angle of attack (AOA) is the difference between the glider’s actual glide path angle and the glider’s measured pitch angle – an important parameter that is difficult to measure but greatly influences the estimation of Vg. To estimate the AOA, this study installed an electromagnetic current sensor (EM sensor) and acoustic Doppler current profiler (ADCP) on a Slocum G2 glider (Teledyne Webb Research). Based on the dataset obtained off the east coast of Japan, the mean AOA derived from the ADCP was 4.49 degrees during the descent, which was comparable to that derived from the steady-state flight model (4.05 deg.), while the AOA derived from the EM sensor was 0.83 deg. We found that by including a velocity offset for the EM sensor, the mean AOA derived from the EM sensor was comparable to those of the flight model and the ADCP. The flight model's drag and lift coefficients were tuned so the flight model's AOA agreed with the ADCP derived AOA. Using the resulting estimation of Vg, from either the flight model or ADCP, provides an unbiased estimation of epsilon.