Do Radiative Feedbacks Depend on the Structure and Type of Climate Forcing, or Only on the Spatial Pattern of Surface Temperature Change?

Earth's climate sensitivity depends critically on the strength of radiative feedbacks linking surface warming to changes in top-of-atmosphere (TOA) radiation. Many studies use a simplistic idea of radiative feedbacks, either by treating them as global mean quantities, or by assuming they can be defined uniquely by geographic location and thus that TOA radiative response depends only on local surface warming. For example, a uniform increase in sea-surface temperature has been widely used as a surrogate for global warming (e.g., Cess et al 1990 and the CMIP 'aqua4k' simulations), with the assumption that this produces the same radiative feedbacks as those arising from a doubling of carbon dioxide -- even though the spatial patterns of warming differ. However, evidence suggests that these assumptions are not valid, and local feedbacks may be integrally dependent on the structure of warming or type of climate forcing applied (Rose et al 2014).

This study thus investigates the following questions: to what extent do local feedbacks depend on the structure and type of forcing applied? And, to what extent do they depend on the pattern of surface temperature change induced by that forcing? Using an idealized framework of an aquaplanet atmosphere-only model, we show that radiative feedbacks are indeed dependent on the large scale structure of warming and type of forcing applied. For example, the climate responds very differently to two forcings of equal global magnitude but applied in different global regions; the pattern of local feedbacks arising from uniform warming are not the same as that arising from polar amplified warming; and the same local feedbacks can be induced by distinct forcing patterns, provided that they produce the same pattern of surface temperature change. These findings suggest that the so-called ‘efficacies’ of climate forcings can be understood simply in terms of how local feedbacks depend on the temperature patterns they induce.