Submesoscale Dynamics in the Upper Ocean: Air--Sea Interactions and Energy Transfers

Submesoscale dynamics, with horizontal scales of O(1–10)~km, are ubiquitous in the surface ocean. Recent observations and simulations reveal enhanced sea surface temperature and velocity gradients at these scales, with sharpened buoyancy fronts attracting particular attention for their potential of energy transfer to link large-scale energy reservoirs to the small scales where dissipation occurs. This thesis addresses the following questions: How do the ocean and atmosphere interact at submesoscale? To what extent can balanced dynamics be applied to interpret submesoscale variability, and, given that, how well do balanced frameworks represent energy transfer at submesoscale fronts? Air--sea interaction has been extensively studied at mesoscale, O(100)~km, but remains less well quantified at submesoscale. Mesoscale sea surface temperature and vorticity imprint their own spatial structure on wind stress gradients through thermal and current feedbacks (TFB and CFB). These feedbacks have typically been assessed separately, with limited attention to their collocation and potential joint impacts. Atmospheric wind stress modifications feed back on the ocean by altering surface vertical transports through Ekman dynamics. In submesoscale-resolving air--sea coupled simulations, examining TFB and CFB jointly reveals their combined influence on surface wind stress. The coupled effect produces anomalous wind stress curls an order of magnitude stronger than those at mesoscale. The resulting nonlinear Ekman velocities from wind stress curl changes reach magnitudes roughly ten times larger than at mesoscale, underscoring the potential importance of submesoscale air--sea coupling for vertical exchange in the surface ocean. The submesoscale range covers a transition from balanced to unbalanced motion as spatial scales decrease. This transition scale is critical for understanding energy transfers and tracer transport in the upper ocean, yet remains poorly constrained. Because balanced motions evolve on time scales much longer than Earth’s rotation period, the temporal scale of submesoscale motions can indicate their dynamical regime. Spatio-temporal analyses in submesoscale-resolving simulations and surface mooring observations show that flow-following Lagrangian time scale of submesoscale flow is much longer than that registered by a fixed Eulerian observer, placing the transition to unbalanced dynamics at ~1 km, much shorter than suggested by Eulerian diagnostics. This implies that balanced theory remain applicable to submesoscale motions down to O(1)~km. Although balanced quasi-geostrophic (QG) and semi-geostrophic (SG) theories reproduce submesoscale frontal development, the associated energy transfers under balanced conditions are less well understood. We derive and assess the full kinetic energy budget in spectral space and in physical space using coarse-graining for an isolated front in QG and SG frameworks. A geostrophic strain field by itself is sufficient to produce forward transfer of kinetic energy in QG frontogenesis. In SG, ageostrophic advection generates a dipole in energy transfer linked to the convergent--divergent surface flow structure, but this balanced ageostrophic circulation is too weak to overcome the dominant forward transfer from the strain field. These results demonstrate that balanced dynamics can generate surface convergence and drive downscale transfers, and that ageostrophic circulation is essential to produce the asymmetry of energy fluxes at submesoscale fronts.