An atmospheric boundary layer model to improve air-sea interactions in eddying ocean models

High-resolution ocean-atmosphere coupled models are able to realistically simulate air-sea interactions taking place at mesoscale between ocean eddies and fronts, and the lower atmosphere. These coupled processes have the potential to improve oceanic simulations by modulating wind work input and turbulent heat fluxes. However, the computational cost and the complexity of such coupled models is prohibitive and inadequate in the context of global eddying oceanic simulations.

Here, we propose an alternative approach based on a one-dimensional vertical atmospheric boundary layer (ABL) model driven by large-scale atmospheric data (forecasts or reanalysis). Its limited computational cost and intermediate complexity between a bulk parameterization and a full atmospheric model makes this approach well suited for applications ranging from process studies to global operational oceanography.

First, the ABL model is validated against a set of classic atmospheric testcases such as a SST front. The comparison with analytical or LES solutions indicates a good agreement with the ABL model results.

Then, two realistic configurations based on NEMO ocean model are presented to assess air-sea interactions: a global 1/4° configuration including sea-ice and a regional 1/36° configuration covering western Europe.

We show that the ocean-ABL coupled model produces negative correlations between surface current and wind stress mesoscale curl anomalies (oceanic eddy damping effect), and positive correlations between surface current and wind speed mesoscale curl anomalies (wind adjustment and ocean re-energization effects) in good agreement with literature. We also show that the simulated wind speed positively correlates with SST mesoscale anomalies, as observed with satellite data and classic coupled models.

To summarize, the ocean-ABL coupled model is able to realistically represent mesoscale dynamical and thermal feedbacks while keeping a good consistency with the atmospheric forcing, and with a very limited computational cost (10% of the ocean model). The ABL model will be released with the next NEMO version.