Towards constraining climate sensitivity by linear analysis of feedback patterns in thousands of perturbed-physics GCM simulations

摘要: A linear analysis is applied to a multi-thousand member “perturbed physics" GCM ensemble identify the dominant physical processes responsible for variation in climate sensitivity across ensemble. Model simulations are provided by distributed computing project, prediction.net . principal component of model radiative response reveals two independent feedback processes, each largely controlled single parameter change. The leading EOF was well correlated with value entrainment coefficient—a model’s atmospheric convection scheme. Reducing this increases high vertical level moisture causing an enhanced clear sky greenhouse effect both control simulation and gas forcing. This compensated increase reflected solar radiation from low cloud upon warming. set ‘secondary’ formation parameters partly modulate degree shortwave compensation formation. second scaling ice fall speed clouds which affects extent cover simulation. most prominent feature longwave EOFs account 70% variance λ—the global parameter. Linear predictors strength climatology observational datasets estimate real world values overall found using correlations Differences between predictions due differences estimates top atmosphere fluxes. Our validation does not rule out all strong tropical convective feedbacks large sensitivity.