Revisiting the Two-Dimensional Estimate of Ocean Vertical Velocity Using Underwater Glider Fleet Observations

Two-dimensional (2D) estimates of the upper-ocean vertical velocity w have been commonly performed based on single hydrographic distance–depth sections. However, biases of these estimates have seldom been investigated. We conduct such an investigation employing a 2-month dataset (including temperature, salinity, and horizontal velocity) at a typical front, the Almeria–Oran Front in the Mediterranean Sea, which was collected by a glider fleet piloted in parallel across-front sections. Specifically, using daily objective maps constructed from the dataset, we perform three-dimensional (3D) and 2D estimates of the balanced w (w_3D and w_2D) through the quasigeostrophic omega equation and evaluate w_2D against w_3D justified previously. Results show a significantly biased w_2D that is estimated assuming a straight front without curvature. Generally, in the 400-m upper ocean, w_2D and w_3D have a weak spatial correlation of 0.4–0.6; w_2D also presents a notably different magnitude, less than 50% of w_3D (even less than 20% in many cases). We find a pronounced curvature-induced shearing deformation (of horizontal density gradients by geostrophic flows) effect destroying the geostrophic balance and so is the associated w to restore the balance; precluding this effect in w_2D leads to the biases. These biases are also analyzed using the potential vorticity conservation principle: As the curvature causes the across-section vorticity advection, water parcels advected by the across-section flow change their vorticity; they have to be vertically compressed/stretched, requiring w that is neglected in w_2D. Therefore, the biased w_2D may be insufficient for understanding the vertical heat transport and its impact on the climate system.