Jump to Main Content
What can decadal variability tell us about climate feedbacks and sensitivity?
- Colman, Robert, Power, Scott B.
- Climate dynamics 2018 v.51 no.9-10 pp. 3815-3828
- climate change, models, oceans, surface temperature, tropics, uncertainty, water vapor
- Radiative feedbacks are known to determine climate sensitivity. Global top-of-atmosphere radiation correlations with surface temperature performed here show that decadal variability in surface temperature is also reinforced by strong positive feedbacks in models, both in the long wave (LW) and short wave (SW), offsetting much of the Planck radiative damping. Net top-of-atmosphere feedback is correlated with the magnitude of decadal temperature variability, particularly in the tropics. This indicates decadal-timescale radiative reinforcement of surface temperature variability. Assuming a simple global ocean mixed layer response, the reinforcement is found to be of a magnitude comparable to that required for typical decadal global scale anomalies. The magnitude of decadal variability in the tropics is uncorrelated with LW feedbacks, but it is correlated with total SW feedbacks, which are, in turn, correlated with tropical SW cloud feedback. Globally, water vapour/lapse rate, surface albedo and cloud feedbacks on decadal timescales are, on average, as strong as those operating under climate change. Together these results suggest that some of the physical processes responsible for setting the magnitude of global temperature change in the twenty-first century and climate sensitivity also help set the magnitude of the natural decadal variability. Furthermore, a statistically significant correlation exists between climate sensitivity and decadal variability in the tropics across CMIP5 models, although this is not apparent in the earlier generation of CMIP3 models. Thus although the link to sensitivity is not conclusive, this opens up potential paths to improve our understanding of climate feedbacks, climate sensitivity and decadal climate variability, and has the potential to reduce the associated uncertainty.