How finely do we need to represent the stratocumulus radiative effect?. (30th June 2016)

• Record Type: Journal Article
• Authors: Bellon, Gilles; Geoffroy, Olivier
• Is Part Of: Quarterly journal of the Royal Meteorological Society. Volume 142:Number 699(2016)
• Page Start: 2347
• Page End: 2358
• Publication Date: 2016-06-30
• Subjects: equilibrium boundary layer -- stratocumulus -- cloud radiative effect -- parametrization; Meteorology -- Periodicals; 551.5
• DOI: 10.1002/qj.2828
• Languages: English
• ISSNs: 0035-9009

Abstract

Abstract : Cloud radiative cooling at the top of stratocumulus is crucial to power turbulence in the well‐mixed, stratocumulus‐topped atmospheric boundary layer (ABL). The present work uses a large‐eddy simulation model to understand how finely the spatial structure of the cloud radiative effect (CRE) has to be represented in order to sustain a stratocumulus‐topped ABL in equilibrium. We first explore whether the vertical structure of the CRE can be simplified, and it appears that the cloud radiative cooling in the upper part of the cloud is essential to the existence of the equilibrium stratocumulus, and that the CRE has to be represented in some detail around the cloud top to sustain this equilibrium. Second, we investigate how the horizontal heterogeneities of CRE impact the stratocumulus equilibrium. Horizontal homogenization of the radiative cooling has a small thinning impact on the stratocumulus‐topped ABL. Because of the nonlinear dependence of the cloud radiative cooling on the liquid water path (LWP), computing the CRE from the mean liquid water profile overestimates the vertically integrated CRE and yields an unstable, ever‐growing ABL. Assuming a simple sub‐domain distribution of LWP and adjusting the cloud radiative cooling accordingly is sufficient to correct for this overestimation and sustain a stratocumulus equilibrium. Finally, we check that vertical and horizontal simplifications of the CRE are compatible, which opens perspectives for its representation in bulk models and larger‐scale models.