Long-term changes of the top-of-atmosphere radiation budget and tropical atmospheric circulation from satellite measurements and climate models

Satellite measurements of OLR and shortwave radiation, as well as surface wind speed have recently been made available for the past 25 years. This offers an opportunity to investigate changes and trends in these fields, as well as links between the observed radiation budget and atmospheric circulation. A hierarchy of simple statistical techniques is formulated: We first examine the linear trends in these variables, which show robust features over the whole tropics. A decomposition in EOFs and Hilbert EOFs is also performed to expose any known long-term variability like ENSO. Multivariate techniques like Maximum Covariance Analysis (MCA, also known as SVD analysis) and Canonical Correlation Analysis (CCA) are used to explore potential links between the radiative and circulation variables. We also use the corresponding variables from the latest runs of the GFDL model (GFDL CM2.1 AM2P13) to test whether this state-of-the-art model is capable to reproduce the observations. Additionally, examining AMIP-like and coupled model runs, we are able to gauge the extend at which these long-term changes arise as natural variability of the climate system, from prescribed SSTs, or other model forcings.