Metric of the 2-6Â day sea-surface temperature response to wind stress in the Tropical Pacific and its sensitivity to the K-Profile Parameterization of vertical mixing

- We model the Tropical Pacific Ocean and compare to buoy observations.
- There is a negative correlation between wind stress and SST on short time scales.
- The model is biased to have stronger correlations than the observations.
- Perturbing certain mixing parameters affects the correlation significantly.
- Wind uncertainty still hampers model evaluation.

Uncertainty in wind forcing has long hampered direct tests of ocean model output against observations for the purpose of refining the boundary layer K-Profile Parameterization (KPP) of oceanic vertical mixing. Considered here is a short-term metric that could be sensitive to the ways in which the KPP directly affects the adjustment of sea surface temperatures for a given change in wind stress. In particular a metric is developed based on the lagged correlation between the 2-6 day filtered wind stress and sea surface temperature. The metric is normalized by estimated observational and model uncertainties such that the significance of differences may be assessed. For this purpose multiple wind reanalysis products and their blended combinations were used to represent the range of forcing uncertainty, while perturbed KPP parameter model runs explore the sensitivity of the metric to the parameterization of vertical mixing. The correlation metric is sensitive to perturbations to most KPP parameters, in ways that accord with expectations, although only a few parameters show a sensitivity on the same order as the sensitivity to switching between wind products. This suggests that uncertainties in wind forcing continue to be a significant limitation for applying direct observational tests of KPP physics. Moreover, model correlations are biased high, suggesting that the model lacks or does not resolve sources of variability on the 2-6 day time scale.