The Impacts of the Biases of Cloud-Radiative-Dynamics Interactions on Global Climate Models in CMIP3 and CMIP5

The coupled global climate model (CGCM) fidelity in representing cloud mass, cloud fraction and their radiative properties on global atmosphere-ocean coupled systems are evaluated.  It is found that, over Pacific oceans, the biases of radiation, surface wind stress, surface evaporation, sea surface temperature (SST), precipitation, and sea surface salinity (SSS) are fairly similar in the models in the Coupled Model Intercomparison Project (CMIP) contributed 20^th century simulations to the phase 3 (CMIP3) and phase 5 (CMIP5) relative to observations.

While all the CMIP3 and most of the CMIP5 models, do not consider precipitating ice (snow) radiative effects, this study examines the impacts of precipitating ice on radiation and explores their connections to the common biases surface wind stress, SST, SSS and Central Pacific El Nino (CP-ENSO) evolution in CMIP5 models using the NCAR Community Earth System Model version one (CESM1). This study indicates that most of these biases is attributable to erroneous cloud-radiative-dynamics feedback.  The feedback is associated with underestimated total ice water path (TIWP) due to exclusion of snow, producing excessive LW cooling in the upper troposphere with a radiatively unstable atmosphere producing low-level outflow anomalies from ITCZ/SPCZ. The outflow then generates weaker low-level trade-winds, warmer SSTs, and excessive precipitation, diluted SSS and failure of capturing the evolution (sustainability) of CP-ENSO events.  The inclusion of the snow radiative effects largely reduces most of the above-mentioned model biases of SSTs, surface winds, precipitation, SSS and the life cycle composite of anomalies associated with CP-ENSO events, compared with those without snow-radiative effects in CESM1 simulations.