Investigating North Pacific air-sea interaction with a hierarchy of empirical models

The null hypothesis of extratropical sea surface temperatures (SST) being driven by rapid, random atmospheric variability is tested during boreal winter in the North Pacific by employing two different coupled linear inverse models (LIM). The LIM is constructed using daily SST and several atmospheric variables such as air temperature, specific humidity and upper-tropospheric geopotential height. The two LIMs differ in complexity: Model 1 is strictly local, while the Model 2 is comprised of area-averaged grid-boxes spanning the North Pacific, allowing for intra-basin coupling. It is shown the simpler Model 1 shows comparable skill to Model 2 with few exceptions.

Decoupling the ocean from atmosphere in either model suggests large differences in coupling strength between the eastern and western portions of the extratropical Pacific. In the east, the atmosphere drives 75% of SST variability, while in the west this stands at 40%. The large difference is attributed to internal oceanic processes, which are more prevalent in the west due to the Kuroshio-Oyashio (KO) current system. Furthermore, it is shown that in the east, 80% of air temperature variability is intrinsic to the atmosphere and this increases to >90% in the west. Overall, results suggest that coupling is more efficient in the eastern North Pacific due to stronger similarity in time scales of atmosphere and ocean variability. In accord with previous studies, results imply that prescribing SST in modeling studies must be done with care due to the sizeable role of previous atmospheric forcing. A notable omission is the role of tropical forcing, which is highlighted in a separate presentation.