Response of the North Atlantic circulation to the small-scale surface wind perturbations

High-frequency and small-scale processes in the atmosphere have an important influence on the evolution of the underlying ocean. They can not only introduce variability to the coupled systems but also have long-term ramification effects on the sea surface temperature, thermocline structure and large-scale ocean general circulation via nonlinear interactions.

Comparisons between the newly-released ECMWF fifth-generation global climate reanalyses (ERA5) wind product and satellite/in-situ observations show that the latest reanalyses winds still considerably underestimate wind variability at high-frequencies and small-scales. A novel approach, Cellular Automata (CA), is used here to stochastically perturb the ERA5 wind field. CA, originally introduced into the weather forecast model to mimic the near-grid-scale variability associated with convective cloud clustering, generates spatially and temporally coherent perturbation patterns. Results show that the CA-perturbed ERA5 wind field enjoys an improved wavenumber spectrum, especially over high wavenumber bands (scales <400 km), when compared to the QuikSCAT measurements. In addition, including CA patterns also brings the level of wind variability at high-frequencies (>1 cpd) closer to in-situ mooring measurements. The local response of the upper ocean properties is investigated by a Multi-Column K-Profile Parameterization (MC_KPP) ocean mixed layer model over Atlantic section. It is found that overall the sea surface temperature (SST) decreases and oceanic boundary layer (OBL) deepens to respond to the enhanced surface turbulent heat loss and shear instability generated at the base of surface OBL caused by the small-scale wind perturbations. In particular, SST tends to decrease the most over the summer hemisphere by up to 1°C locally and 0.1°C averaging across the basin, in correlation with shallower background OBL and smaller OBL heat capacity. 

The ocean states under the forcing of stochastic wind perturbation as expressed by the local response are found rectified by a non-negilible magnitude. We argued that although the missed small-scale and high-frequency wind variability may be represented differently than our approach, our results highlighted the fact that these variabilities should take a singificant part in driving the current generation of coupled climate model. Furhtermore, the temproal and spatial variabilities of the local signals can pose significant influence on the large-scale ocean circulation in terms of their pathways, strengths and variabilities. The dynamical response of the ocean circulation is to be further understood with an eddy-resolving Massachusette Institute of Technology General Circulation Model (MITgcm).