Group seminar on 24. January, 14:15 CET
Geographical dependence of multidecadal mean impacts of SST bias on the simulated atmospheric circulation and spatio-temporal variability
Dr. Yuan-Bing Zhao
The state-of-the-art climate models suffer from significant sea surface temperature (SST) bias, greatly damaging the climate prediction and projection. Using a dynamical approach, we investigate the multidecadal mean impacts of SST bias on the simulated atmospheric circulation and variability as well as their dependence on the location of SST bias. A set of century long simulations forced with idealized constant SST perturbations (with the maximal amplitude of 1.5 K), which mimick SST biases in the coupled climate models, are first conducted with an intermediate complexity atmospheric model. The impact of SST bias is then evaluated using the normal-mode function decomposition which can differentiate between balanced and unbalanced flow regimes. The results show that the strength of the SST-bias impact depends on the background SST. SST bias in regions with background SSTs ≥ 25°C (especially the tropical Indo-west Pacific warm pool) can result in strong atmospheric circulation biases worldwide, which have the Matsuno-Gill-type pattern in the tropics and Rossby wavetrain distribution in the extratropics. Otherwise, the impact is localized and weak. Likewise, SST bias in warm regions strongly affects the spatial (SV) and temporal variability (TV), and the impacts show distinct geographical dependence. Basically, SST bias in the tropical Indo-west Pacific region (elsewhere) weakens (strengthens) the unbalanced and the tropical (25°S-25°N) balanced SV, and that in the west Pacific (elsewhere) strengthens (weakens) the extratropical balanced SV. As for TV, it is generally weakened (strengthened) when SST bias is in the tropical Indian Ocean (Pacific), regardless of the flow regime. Finally, the physical causes of the geographic dependence are discussed.