Improved methods for estimating equilibrium climate sensitivity from transient warming simulations

Multivariate Estimations of Equilibrium Climate Sensitivity From Short Transient Warming Simulations

“…Hence, the equilibrium climate sensitivity of these models is instead estimated by extrapolating transient warming simulationsway before these models have reached equilibrium (Knutti & Hegerl, 2008;Rugenstein et al, 2020;Dai et al, 2020;Knutti et al, 2017). There are several techniques to perform such extrapolation that use different physical and mathematical properties of the system to give sensible estimates for the true equilibrium climate sensitivity of a model (Knutti & Hegerl, 2008;Dai et al, 2020;Gregory et al, 2004;Proistosescu & Huybers, 2017).…”

“…On top of that, recent results of the new generation of global climate models show even higher sensitivities, possibly due to better representation of cloud formation when using finer spatial grids (Andrews et al., 2019; Bacmeister et al., 2020; Bony et al., 2015; Duffy et al., 2003; Govindasamy et al., 2003; Haarsma et al., 2016; Zelinka et al., 2020). Still, even these state‐of‐the‐art climate models report significantly different climate sensitivities (Flynn & Mauritsen, 2020; Forster et al., 2020; Zelinka et al., 2020); moreover, estimates for a single model tend to have large uncertainties further hampering accurate pinpointing of the climate sensitivity (Dai et al., 2020; Rugenstein et al., 2020).…”

“…With the current trend—and need—to build models with higher temporal and spatial resolutions (Duffy et al., 2003; Eyring et al., 2016; Govindasamy et al., 2003; Haarsma et al., 2016), this is not expected to resolve itself in the near future. Hence, the equilibrium climate sensitivity of these models is instead estimated by extrapolating transient warming simulations—way before these models have reached equilibrium (Dai et al., 2020; Knutti & Hegerl, 2008; Knutti et al., 2017; Rugenstein et al., 2020). There are several techniques to perform such extrapolation that use different physical and mathematical properties of the system to give sensible estimates for the true equilibrium climate sensitivity of a model (Dai et al., 2020; Geoffroy et al., 2013a; Gregory et al., 2004; Knutti & Hegerl, 2008; Proistosescu & Huybers, 2017).…”

“…These feedbacks are quite diverse and span a wide range of spatial and temporal scales; these include, for example, the very fast Planck feedback, the slower ice‐albedo feedback and the even slower ocean circulation feedbacks. Estimation techniques deal differently with this problem, for instance by incorporating multiple time scales directly in the estimation method (Proistosescu & Huybers, 2017), by explicit modeling of long‐term (ocean) heat uptake (Geoffroy et al., 2013a) or more indirectly by ignoring initial fast warming behavior (Dai et al., 2020; Rugenstein et al., 2020).…”

Sea ice-air interactions amplify multidecadal variability in the North Atlantic and Arctic region

“…Superimposed on a rapid warming trend, Arctic surface air temperature (Tas) since the early 20 th century also exhibits large multidecadal variations [6][7][8] that cannot be explained by concurring monotonic increases in atmospheric greenhouse gases (GHGs). Model simulations 6,9,10 show that internal variability can generate similar low-frequency variations in Arctic Tas; and multidecadal variations in poleward energy transport associated with the Atlantic Multidecadal Oscillation (AMO) and Atlantic Meridional Overturning Circulation (AMOC) 11 have been identi ed as a leading cause of the Arctic Tas variations 9,12,13 . It is suggested 9,13 that above-normal oceanic heat transport from the North Atlantic into the Barents-Kara Seas (BKS) and other Arctic regions reduces sea-ice cover (SIC) there, which allows the Arctic Ocean to absorb more energy during the summer but release more heat to the air to cause warmer Tas during the winter.…”