Reproducibility by Climate Models of Cloud Radiative Forcing Associated with Tropical Convection

In this study, cloud radiative forcing (CRF) associated with convective activity over tropical oceans is analyzed for monthly mean data from twentieth-century simulations of 18 climate models participating in phase 3 of the Coupled Model Intercomparison Project (CMIP3) in comparison with observational and reanalysis data. The analysis is focused on the warm oceanic regions with sea surface temperatures (SSTs) above 27 degrees C to exclude the regions with cold SSTs typically covered by low stratus clouds. CRF is evaluated for different regimes sorted by pressure-coordinated vertical motion at 500 hPa (omega 500) as an index of large-scale circulation. The warm oceanic regions cover the regime of vertical motion ranging from strong ascent to weak descent. The most notable feature found in this study is a systematic underestimation by most models of the ratio of longwave cloud radiative forcing (LWCRF) to shortwave cloud radiative forcing (SWCRF) over the weak vertical motion regime defined as -10 < omega 500 < 20 hPa day(-1). The underestimation of the ratio corresponds to the underestimation of LWCRF and the overestimation of SWCRF. Clouds in models seem to be lower in the amount of high clouds but more reflective than those in the observations in this regime. In the weak vertical motion regime, the lower free troposphere is dry. In the large-scale environment condition, the reproducibility of LWCRF is high in models adopting the scheme where the relative humidity-based suppression for deep convection occurrence is implemented. Models adopting the Zhang and McFarlane scheme show good performance without such a suppression mechanism.