Investigation of forced and active cumulus at the SGP. Inferences for the ARM SGP supersite configuration.

Description

Continental fair-weather cumuli are shallow, short-lived, non-precipitating clouds that form when moist thermals initiated at the surface reach the lifting condensation level. Despite the apparent simplicity of their formation process, numerical models of various scales struggle to adequately simulate their properties. A possible avenue to improve continental cumulus representation could involve segregating cloud fraction between forced and active clouds. This has been previously attempted using cloud thickness arguments and without considering the coexistence of these clouds within the same observing period. Here, in addition, we consider the avenue of using cloud base dynamical information to improve the cloud classification. Using 3 years (2011-2014) of summertime (months 05-09) daytime (08-17 LT) observations at the SGP site, we identify 747 hours of shallow fair-weather cumulus. Individual clouds are separated using continuous ceilometer first cloud bases height. Their thickness is determined using the KAZR moderate sensitivity mode when available and the general mode otherwise. The maximum Microwave radiometer liquid water path for each cloud is recorded and used to evaluate the validity of the estimated cloud thickness. The cloud base velocity is extracted from both the KAZR Doppler velocity and the Doppler lidar velocity for comparison. Initially, a thickness threshold of 300m is used to separate the forced and active clouds. Alternatively, a threshold of >70% cloud base updraft fraction is used to identify active cumulus. During our analysis, the issue of radar sensitivity to continental shallow cumuli emerged several times. As has been shown in the past, cloud radars are not sensitive enough to detect all clouds and very often are heavily affected by the presence of insects. On the other hand, lidar systems detect all cloud bases and their velocities. The presented results have implications on the configuration of the proposed SGP supersite and on the subsequent production of objective 4D data cubes suitable for routine evaluation of high-resolution models.