The Response of Trade-wind Clouds to a Changing Environment: Why Climate Model Cloud Feedbacks Might Differ from Nature?

Abstract:

The vertical distribution of clouds in the trades, and how this distribution is linked to the thermodynamic structure of the lower troposphere, is emerging as a key factor in the assessment of modeled cloud feedbacks. The spread in cloud feedbacks and climate sensitivity may be attributed to how efficiently models dry the lower troposphere as climate warms. This relates to how models mix moisture vertically, and thus to how their clouds are distributed.

A three year record of cloudiness and boundary layer structure from a ground-based remote sensing station at Barbados, situated in a typical trade-wind region, is used to gain insight into the major components of the cloud distribution, and how these vary with changes in atmospheric structure. These insights are then used to evaluate modeled cloud behavior in single time step (half-hourly) output of the ECMWF IFS and CMIP5 models, at a grid point nearby Barbados.

Cloudiness is found to be regulated by two components with a different sensitivity to the large-scale flow. Cloud near cloud base is on average the largest contributor to cloudiness, but is relatively constant on time scales much longer than a day. This is because turbulence and cumulus convection adjust fast to perturbations, and self-regulate the cumulus mass flux and cloudiness near cloud base. A secondary component consists mainly of stratiform outflow layers near the detrainment level of cumulus tops, which carry most of the variability on longer time scales. The organization of convection into larger and deeper clusters, which can vertically transport large amounts of moisture, appear key in regulating such outflow layers.

Several models fail at reproducing these smaller scale processes that control trade-wind cloudiness in the current climate. The stratiform component in modeled distributions of trade-wind cloudiness is seldom pronounced, and their distributions tend to be bottom-heavy. Whereas observations suggest that changes in clouds are more likely manifested in the component of cloud aloft, models show large variances in their cloud base component instead. Hence, changes in cloudiness into a warmer climate predicted by models may be largely carried by changes that are unlike nature.