Fig. 3: Constrained SLF and its impact on top-of-the-atmosphere radiation.

a Left top “Original" NorESM2 (2009–2013) SLF (%) vs atmospheric (vertical) temperature (°C). Black lines indicate observed cloud interior (dotted) and cloud top (dash-dot) from CALIOP observations, orange and blue lines indicate NorESM2 SLF. Left bottom SLF versus temperature for the same NorESM2 setup but with observationally constrained SLF. The SLF vs. temperature figures show the SLF curves for the geographical areas indicated with the dark red rectangle on the map plot. All NorESM2 simulations used for matching CALIOP observed are with nudged ERA-Interim circulation. Map plot shows the difference in top-of-the-atmosphere (TOA) net radiation between NorESM2 Global with constrained SLF minus the unconstrained NorESM2 setup (blue indicates a cooling when SLFs are constrained). b Same as a but NorESM2 only constrained in the northern extratropics (≥30°N). c same as b but for the southern extratropics (≥30°S). d Absolute change in top-of-the-atmosphere radiation bias (W m⁻²) between NorESM2 Control and three NorESM2 simulations with constrained MPCs. We computed the bias for each simulation compared to CERES satellite observation first and then computed the difference between the NorESM2 Control and our three constrained simulations. A negative value indicates a lower bias in the constrained simulation than in NorESM2 Control when compared to CERES observations of TOA radiation.