Tropical cloud feedbacks estimated from observed multidecadal trends

Tropical cloud feedbacks are an important source of uncertainty in estimates of climate sensitivity. The extent to which changes in atmospheric circulation contribute to these feedbacks remains an open question. Here, all-sky radiative flux observations and an atmospheric reanalysis are used to estimate tropical cloud feedbacks from multidecadal trends (1985–2020) in cloud radiative effect and surface temperature. We decompose the observed feedbacks into dynamic and nondynamic components to quantify the impact of circulation trends. Narrowing and strengthening of tropical ascent lead to substantial dynamic feedbacks on regional scales that are similar in magnitude to the nondynamic feedbacks. However, as previously shown for high- and low-resolution climate models, large dynamic feedbacks in different circulation regimes are connected by the atmospheric mass budget and approximately cancel when averaged across the tropics due to the quasi-linear relationship between cloud radiative effect and vertical velocity. This results in small dynamic contributions to the tropical-mean net, longwave, and shortwave feedbacks. We suggest that this result will hold in future and thus that isolating the nondynamic components associated with individual cloud types can provide important insights into the processes controlling the tropical-mean cloud feedback and its uncertainty. Additionally, we show that feedbacks estimated from multidecadal trends differ from those estimated from interannual variability. We demonstrate that, for dynamic feedbacks, this is because changes are controlled by different mechanisms and this leads to a differing spatial distribution of temperature sensitivity. Finally, we provide new estimates of the uncertain combined tropical anvil area and albedo feedback using both multidecadal trends and interannual variability.