Direct numerical simulation of transient cumulus cloud flow

Abstract

Clouds play a major role in climate change, and the ability to simulate moist convection patterns is crictical for prediction of tropical weather and climate. Cumulus clouds in particular can play a significant role in transportig heat across the whole extent of the atmosphere. Recent laboratory experiments (Narasimha et al. (2011) have successfully reproduced a variety of naturally occurring clouds, and suggest that the transient diabatic plume, subjected to off-source diabatic heating is the appropriate cumulus flow model. In the present work we report the first direct numerical simulation of a transient diabatic plume as a fluid-dynamical model for understanding cumulus flows. The simulation solves the 3D Navier-Stokes-Boussinesq equuations for an axisymmetric transient diabatic plume. The equations were solved using a fractional step method within the finite volume frame work. The solver developed has been validated against three bechmark cases - (i) lid driven cavity flow; (ii) Rayleigh-Benard convection and (iii) Turbulent-jet simulations. The visualisations of the cloud flow were carried out using a coarse grid of around 4 million grid points, The final simulation was performed using 128 million grid points at a Reynolds number of 2000. The computations were carried out in the ICE cluster housed at CSIR Fourth Paradigm Institute, Bangalore.