Accounting for Vertical Subgrid‐Scale Heterogeneity in Low‐Level Cloud Fraction Parameterizations. (5th November 2018)

• Record Type: Journal Article
• Authors: Jouhaud, J.; Dufresne, J.‐L.; Madeleine, J.‐B.; Hourdin, F.; Couvreux, F.; Villefranque, N.; Jam, A.
• Is Part Of: Journal of advances in modeling earth systems. Volume 10:Number 11(2018)
• Page Start: 2686
• Page End: 2705
• Publication Date: 2018-11-05
• Subjects: cloud structure -- subgrid scale -- heterogeneity -- GCM -- LES; Geological modeling -- Periodicals; Climatology -- Periodicals; Geochemical modeling -- Periodicals; 551.5011
• DOI: 10.1029/2018MS001379
• Languages: English
• ISSNs: 1942-2466

Abstract

Abstract: Many general circulation models (GCMs) assume some heterogeneity of water amounts in their grid boxes and use probability density functions to parameterize cloud fractions C F and amounts of condensed water q c . Most GCM cloud schemes calculate the CF as the volume of the grid box that contains clouds ( C F vol ), whereas radiative fluxes primarily depend on the CF by surface ( C F surf ), that is, the surface of the grid box covered by clouds when looking from above. This discrepancy matters as previous findings suggest that C F surf is typically greater than C F vol by about 30%. In this paper we modify the single column model version of the LMDz GCM cloud scheme by introducing the vertical subgrid‐scale heterogeneity of water content. This allows to distinctly compute the two fractions, C F vol and C F surf, as well as the amount of condensed water q c . This study is one of the first to take into account such vertical subgrid‐scale heterogeneity in a GCM cloud scheme. Three large eddy simulation cases of cumuliform boundary layer clouds are used to test and calibrate two different parameterizations. These new developments increase cloud cover by about 10% for the oceanic cases RICO and Barbados Oceanographic Meteorological Experiment and by up to 50% for the continental case ARM. The change in condensed water reduces the liquid water path by 10–20% and therefore the cloud opacity by 5–50%. These results show the potential of the new scheme to reduce the too few, too bright bias by increasing low‐level CF and decreasing cloud reflectance. Key Points: Most cloud schemes assume vertically homogeneous cloud fractions in GCM grid boxes This approximation may induce an underestimation of cloud cover and overestimation of reflectance We parameterize 3‐D subgrid‐scale heterogeneities for low‐level clouds and reduce these biases