Submesoscale Currents from UAV: An Experiment over Small-Scale Eddies in the Coastal Black Sea

A commercial unmanned aerial vehicle (UAV) is used for coastal submesoscale current estimation. The measurements were conducted in the Black Sea coastal area with a DJI Mavic quadcopter operated in self-stabilized mode at different look geometry (200-500-m altitude, 0-30 degrees incidence angle). The results of four flights during 2020-2021 are reported. Some scenes captured a train of or individual eddies, generated by a current flowing around a topographic obstacle (pier). The eddies were optically visible due to the mixing of clear and turbid waters in the experiment area. Wave dispersion analysis (WDA), based on dispersion shell signature recognition, is used to estimate the sea surface current in the upper 0.5-m-thick layer. The WDA-derived current maps are consistent with visible eddy manifestations. The alternative method, based on 4D-variational assimilation (4DVAR), agrees well with WDA and can complement it in calm wind conditions when waves are too short to be resolved by the UAV sensor. The error of reconstructed velocity due to the uncontrolled UAV motions is assessed from referencing to static land control points. At a 500-m altitude and 7-10 m s(-1) wind speed (reported by a local weather station for 10-m height), the UAV drift velocity, or the bias of the current velocity estimate, is about 0.1 m s(-1), but can be reduced to 0.05 m s(-1) if the first 10 s of the UAV self-stabilization period are excluded from the analysis. The observed anticyclonic eddies (200-400 m in diameter with 0.15-0.30 m s(-1) orbital velocity) have an unexpectedly high Rossby number, Ro similar to 15, suggesting the importance of nonlinear centrifugal force for such eddies and their significant role in coastal vertical transport.

Automated summary

This study uses a commercial unmanned aerial vehicle (UAV) to estimate coastal submesoscale currents and validates the consistency between the results obtained using wave dispersion analysis (WDA) and 4D-variational assimilation (4DVAR). The error caused by uncontrolled UAV motions is assessed by referencing to static land control points.