Estimating Climate Sensitivity from UV satellite observations and CMIP6 models since 1980

We construct a long-term record of Top of Atmosphere shortwave (SW) albedo of clouds and aerosols from 340 nm radiances observed by NASA and NOAA satellite instruments from 1980 to 2013. We compare our SW cloud+aerosol albedo with simulated cloud albedo from both AMIP and historical CMIP6 simulations from 47 climate models. While most historical runs did not simulate our observed spatial pattern of the trends in albedo over the Pacific Ocean, four models qualitatively simulate our observed patterns. Those historical models and the AMIP models collectively estimate an Equilibrium Climate Sensitivity (ECS) of ~3.5^oC, with an uncertainty from 2.7 to 5.1^oC. Our ECS estimates are sensitive to the instrument calibration which drives the wide range in ECS uncertainty. We force the calibrations to have a near neutral change in reflectivity over the Antarctic ice sheet. Our observations show no sign of dissipating marine stratocumulus clouds. Instead, they show increasing cloudiness over the eastern equatorial Pacific and off the coast of Peru as well as neutral cloud trends off the coast of Namibia and California.

 To produce our SW cloud+aerosol albedo we first retrieve a Black-sky Cloud Albedo and empirically correct the sampling bias from diurnal variations. Then we estimate the broadband proxy albedo using multiple non-linear regression along with several years of CERES cloud albedo to obtain the regression coefficients. We validate our product against CERES data from the years not used in the regression. Zonal mean trends of our SW cloud+aerosol albedo show reasonable agreement with CERES as well as the Extended Pathfinder Atmospheres (Patmos-x) observational dataset.