Process-Based Decomposition of the Global Surface Temperature Response to El Niño in Boreal Winter

This paper reports an attribution analysis that quantifies addible contributions to the observed temperature anomalies from radiative and nonradiative processes in terms of both amplitude and spatial pattern for the two most prominent surface temperature patterns in an El Niño winter. One is the El Niño SST pattern consisting of warming SST anomalies over the eastern equatorial Pacific basin surrounded by cooling SST anomalies in the western and subtropical Pacific, and the other is a tripole surface temperature anomaly characteristic of a positive Pacific–North American (PNA) teleconnection pattern. The decomposition of the observed temperature anomalies is achieved with the coupled atmosphere–surface climate feedback-responses analysis method (CFRAM), which is formulated utilizing energy balance in the atmosphere–surface columns and linearization of radiative energy perturbation. Out of the mean amplitude of 0.78 K of the El Niño SST pattern, the oceanic dynamics and heat storage term alone contributes to 2.34 K. Water vapor feedback adds another 1.6 K whereas both cloud and atmospheric dynamical feedbacks are negative, reducing the mean amplitude by 2.02 and 1.07 K, respectively. Atmospheric dynamical feedback contributes more than 50% (0.73 K) of the mean amplitude (1.32 K) of the PNA surface temperature pattern. Water vapor and surface albedo feedbacks contribute 0.34 and 0.13 K, respectively. The surface processes, including oceanic dynamics in the North Pacific, heat storage anomalies, and surface sensible/latent heat flux anomalies of ocean and land also contribute positively to the PNA surface temperature pattern (about 0.14 K). Cloud and ozone feedback, although very weak, act to oppose the PNA surface temperature anomaly.