Sensitivity of ocean-atmosphere multiscale coupled model to oceanic parameterizations.

This work explores new pathways toward a better representation of the multi-scale physics that drive climate variability. We are targeting key upscaling processes by which small-scale localized errors have a knock-on effect onto global climate. We focus on the Peru-Chile coastal upwelling, an area known to hold among the strongest models biases in the Tropics.

Our approach is based on the development of a multiscale coupling interface allowing us to couple WRF with the NEMO oceanic model in a configuration including 2-way nested zooms in the oceanic and/or the atmospheric component of the coupled model. Upscalling processes are evidenced and quantified by comparing 20-year long simulations of a tropical channel (45°S-45°N), which differ by their horizontal resolution: 0.75° everywhere, 0.75°+0.25° zoom in the southeastern Pacific either in the ocean alone, in the atmosphere alone or in both, ocean and atmosphere. This group of 20-year long simulations was repeated with different sets of parameterizations to assess the robustness of our results.

In this presentation, we will underline the difficulty to disentangle the impact of the increase of resolution from the changes in the parameterizations required by each resolution. Our results show that adding an embedded zoom over the southeastern Pacific only in the atmosphere cools down the SST along the Peru-Chili coast, which is a clear improvement. However, increasing the resolution only in the oceanic component show contrasting results according to the different set parameterization used in the experiments. Some experiment shows a coastal cooling as expected, whereas, in other cases, we observe a counterintuitive response with a warming of the coastal SST. Using a zoom in both ocean and atmosphere mostly combines the results obtained with a zoom in only one component. In the best case, we archive by this means a reduction of the coastal SST of several degrees in agreement with the observations.