Open AccessCharacteristic updrafts for computing distribution‐averaged cloud droplet number and stratocumulus cloud properties
Author(s) -
Morales R.,
Nenes A.
Publication year - 2010
Publication title -
journal of geophysical research: atmospheres
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.67
H-Index - 298
eISSN - 2156-2202
pISSN - 0148-0227
DOI - 10.1029/2009jd013233
Subject(s) - liquid water content , computation , mechanics , cloud computing , scale (ratio) , environmental science , function (biology) , probability density function , vertical velocity , statistical physics , gaussian , distribution function , computational physics , meteorology , mathematics , physics , thermodynamics , computer science , statistics , algorithm , quantum mechanics , evolutionary biology , biology , operating system
A computationally effective framework is presented that addresses the contribution of subgrid‐scale vertical velocity variations in predictions of cloud droplet number concentration (CDNC) in large‐scale models. Central to the framework is the concept of a “characteristic updraft velocity” , which yields CDNC value representative of integration over a probability density function (PDF) of updraft (i.e., positive vertical) velocity. Analytical formulations for are developed for computation of average CDNC over a Gaussian PDF using the Twomey droplet parameterization. The analytical relationship also agrees with numerical integrations using a state‐of‐the‐art droplet activation parameterization. For situations where the variabilities of vertical velocity and liquid water content can be decoupled, the concept of is extended to the calculation of cloud properties and process rates that complements existing treatments for subgrid variability of liquid water content. It is shown that using the average updraft velocity (instead of ) for calculations of N d , r e , and A (a common practice in atmospheric models) can overestimate PDF‐averaged N d by 10%, underestimate r e by 10%–15%, and significantly underpredict autoconversion rate between a factor of 2–10. The simple expressions of presented here can account for an important source of parameterization “tuning” in a physically based manner.