Simulation of subtropical precipitating shallow convection with single- and double-moment warm-rain microphysics

Details of the precipitation development and fallout in shallow convective clouds can have a significant impact on both the cloud layer and the boundary layer beneath. For example, as found in the LES model intercomparison based on RICO (Rain In Cumulus over the Ocean Experiment; Rauber et al. 2007) observations, precipitation development inhibits growth of the cloud layer and slows down the rise of the trade-wind inversion. At the same time, precipitation-laden downdrafts affect thermodynamic properties of the boundary layer and thus the magnitude and partitioning of the surface flux between sensible and latent components (i.e., the Bowen ratio). Grabowski and Morrison (2010) have found recently that differences in the formulation of the rain evaporation between single- and double-moment warm-rain schemes significantly affects the Bowen ratio. Entrainment/detrainment (i.e., the mixing of a cloud with adjacent cloud-free air) is yet another process through which cloud environment is impacted. The layer of air that surrounds the cloud is often described as the halo (e.g., Heus 2008). Some studies suggest that the halo is more humid and colder than the far environment (e.g., Gerber 2008, Heus 2008). This is due to detrainment and mixing, and details of this process do depend on the microphysical properties of the cloudy air (Heus et al. 2008, Jonker et al. 2008).

This paper considers effects of various modeling strategies for the warm-rain microphysics---the single-moment and double-moment scheme in particular---on selected aspects listed above. For that, we analyze our results from the LES intercomparison of RICO case presented at http://www.knmi.nl/samenw/rico/ and compare results obtained with single-moment (Grabowski 1998) and double-moment (Morrison and Grabowski 2007, 2008) schemes. The Eulerian version of the three-dimensional anelastic model EULAG (EUlerian/semi-LAGrangian; Smolarkiewicz and Margolin 1997, Grabowski and Smolarkiewicz 1996, and Margolin et al. 1999) is used. Simulations assume either the maritime (CCN concentration of 100/mg; the PRISTINE case) or the continental (1000/mg; the POLLUTED case) aerosol characteristics. We will report results concerning the impact of contrasting surface precipitation rates in PRISTINE and POLLUTED cases and different microphysical schemes on the boundary layer growth and on the mean properties of the cloud layer and the boundary layer.