Subgrid-Scale Condensation in Models of Nonprecipitating Clouds

G. Sommeria National Center for Atmospheric Research, Boulder, Colo. 80303

Search for other papers by G. Sommeria in
Current site
Google Scholar
PubMed
Close
and
J. W. Deardorff National Center for Atmospheric Research, Boulder, Colo. 80303

Search for other papers by J. W. Deardorff in
Current site
Google Scholar
PubMed
Close
Restricted access

Abstract

One of the shortcomings of present condensation schemes is the assumption that a computational grid volume is either entirely saturated or entirely unsaturated, which is a crude approximation in some instances even with a relatively fine mesh. The concept of a statistical distribution of the points where condensation occurs inside a given grid volume is discussed and results are presented for diagnosing both the fraction of the grid volume containing saturated air and the liquid water content when that fraction is less than unity.

An abbreviated procedure is tested in a boundary layer model of 50 m mesh for a case of nonprecipitating tropical moist convection. Two main effects upon the simulated turbulent field are found: first, the cloud activity measured by the total liquid water content or by the vertical moisture transport is increased and extends farther up, at the expense of subcloud layer moisture; second, the temporal variations within the model domain of the calculated turbulent properties of the cloud layer are decreased, which allows a given time sample of the simulated atmosphere to be more representative of the ensemble mean.

Abstract

One of the shortcomings of present condensation schemes is the assumption that a computational grid volume is either entirely saturated or entirely unsaturated, which is a crude approximation in some instances even with a relatively fine mesh. The concept of a statistical distribution of the points where condensation occurs inside a given grid volume is discussed and results are presented for diagnosing both the fraction of the grid volume containing saturated air and the liquid water content when that fraction is less than unity.

An abbreviated procedure is tested in a boundary layer model of 50 m mesh for a case of nonprecipitating tropical moist convection. Two main effects upon the simulated turbulent field are found: first, the cloud activity measured by the total liquid water content or by the vertical moisture transport is increased and extends farther up, at the expense of subcloud layer moisture; second, the temporal variations within the model domain of the calculated turbulent properties of the cloud layer are decreased, which allows a given time sample of the simulated atmosphere to be more representative of the ensemble mean.

Save