Chidong Zhang — NOAA/PMEL
D Zermeno — University of Miami
Category
Shallow-Deep Convective Transition
Description
Composite profiles of fractions for shallow clouds at Manus at different time of the day from (a) the MMCR observation and (c) CAM5 simulations at model levels. Mean profiles of relative humidity for shallow cloud events from (b) the sounding observations near the launch times and (d) the CAM5 simulations six times per day.
Shallow convective clouds are defined as clouds with their bases within the boundary layer and tops below the freezing level. Shallow clouds at the Manus site were identified using data from the MMCR after errors due to attenuation by rain were corrected or avoided. Precipitating and non-precipitating shallow clouds were separated using rain rate estimates of a rain gauge. Shallow clouds were studied in two aspects.
(1) Moistening effect of shallow clouds on convective development of the Madden-Julian Oscillation (MJO). Fractions of precipitating and non-precipitating shallow clouds before and after local rainfall peaks of the MJO fluctuate on synoptic scales but do not show systematic increases leading to the rainfall peaks. Under an assumption that cloud liquid content linearly increases from the base to top within shallow clouds, water budget of shallow clouds were estimated using liquid water path measured by the microwave radiometer. Shallow clouds, either precipitating or non-precipitating, provide an important background moistening effect through their detrainment or dissipation, but do not explain the increases in low-level moisture leading to MJO rainfall peaks observed by soundings.
(2) Shallow cloud statistics at Manus were compared to those from CAM5 simulations. The model overproduces total clouds and their precipitation fraction, but underestimates total shallow clouds. However, the model produces excessive shallow clouds in a thin layer between 954-930 hPa. This layer coincides with excessive humidity near the top of the boundary layer. It is suggested that insufficient turbulence mixing at the top of the boundary layer is responsible for the excessive humidity, which leads to excessive shallow clouds produced by the microphysics scheme of the model. The model also fails to produce the observed diurnal cycle in shallow clouds. This is not caused by the model coarse grid spacing that misses the landmass of Manus Island, but is related to insufficient diurnal fluctuations in boundary-layer humidity. These results suggest that deficiencies in boundary layer parameterization can lead to unrealistic simulations in shallow clouds.
