Shortwave Spectral Climate Change Signals of Earth Systems with Low, Medium, and High Climate Sensitivity

We present centennial-length shortwave hyperspectral simulations from low, medium and high equilibrium climate sensitivity (ECS) models that reported to the CMIP5 archive. The simulations, as part of an Observing System Simulation Experiment (OSSE) in support of the CLimate Absolute Radiance and Refractivity Observatory (CLARREO), run on CMIP5 archive results and are agnostic to the choice of model. We simulated spectra from the INM-CM4 model (ECS of 2.08 °K/2xCO₂), the MIROC5 model (ECS of 2.70 °K/2xCO₂), and the CSIRO Mk3-6-0 (ECS of 4.08 °K/2xCO₂) based on those models’ integrations of the RCP8.5 scenario. This approach allows us to explore which data records can exclude model results based on the climate sensitivity. We find that substantial differences exist in the distribution of shortwave radiant energy, both broadband and spectrally between the models, and we find that spectrally-resolved measurements can potentially exclude a low or high sensitivity model with under 15 years’ of absolutely-calibrated data. Shortwave spectral information is much faster at model exclusion than shortwave broadband measurements. More specifically, visible spectral channels are best at distinguishing models with different cloud radiative effect response to RCP8.5 forcing, while near-infrared water-vapor overtone bands are best at distinguishing models with different cloud height response. We also find that the complexity of spectral climate signals, as evidenced by the data record length required to distinguish different climate models, is much higher for regional analyses than for global analyses.