DYNAMICS OF THE SMALL-SCALE TURBULENT MIXING IN CLOUDS: NUMERICAL EXPERIMENT. Mirosław Andrejczuk, Szymon P. Malinowski Warsaw University, Poland. Wojciech W. Grabowski, and Piotr Smolarkiewicz National Center for Atmospheric Research, USA. 1. INTRODUCTION In recent years a growing attention is focused on small-scale properties of clouds (review in Valllancourt and Yau, 2000), especially on interaction of cloud particles with turbulence. In this research, however, turbulence is assumed as “given”, mostly by transport in spectral sense from large scales (according to the Kolmogorov theory). Little attention is being paid to production of turbulence in small scales by evaporative cooling of cloud liquid water in process of cloud-clear air mixing. There are, however, indications, that this production may affect smallest scales of turbulence resulting in e.g. anisotropy of cloudy filaments (Banat and Malinowski, 1999, later BM). In this paper we investigate such effects studying very small scales of turbulent mixing in clouds by detailed numerical modeling of dynamics, thermodynamics and microphysics with centimeter resolution. We assume that at these scales no subgrid TKE parameterization is necessary, thus dynamical setup of the model is similar to DNS simulations with decaying turbulence (Herring and Kerr, 1993). We address the following questions: 1) When effects of LWC evaporation are important for TKE evolution in small scales and when can be neglected? 2) Does sedimentation of cloud droplets play an important role in smallest scales of turbulent mixing? 3) Is small-scale turbulence in clouds really isotropic? The model used in these simulations is nonhydrostatic anelastic model by Smolarkiewicz and Margolin (1997) with moist thermodynamics by Grabowski and Smolarkiewicz (1996). Governing equations applied in the simulations can be written as follows: v k v ∆ + + −∇ = µ π B Dt D ; 0 = ∇v ; θ µ θ θ θ ∆ + = d e p e C T c L Dt D ; d v C Dt Dq − = ; where v-velocity vector, π-pressure perturbation with respect to a hydrostatically balanced environment profile normalized by the density, k-vertical unit vector, L, c p - latent heat of condensation and specific heat at constant pressure, C d - condensation rate; θ - potential temperature; q v , q c - water vapor and cloud water mixing ratios, µ µ θ - viscosity and thermal diffusivity of the air. Index "e" denotes environmental undisturbed value and B is buoyancy defined as:         − − − − = c ve v e q q q g B ) ( 0 ε θ θ θ ; where ε=R v /R d -1, g- acceleration of gravity, θ e - environmental temperature profile. In the simulations we use two alternative parameterizations of microphysical processes: 1) Bulk parameterization, described by: d c C Dt Dq = ; 2) Detailed parameterization, closely following Grabowski (1989), where we solve conservation equation for the number density function of cloud droplets f(x,r,t) accounting for droplet sedimentation velocity. Here f(x,r,t)dr is the concentration of droplets of radius between r and r+dr at a given point x in space and at given time t evolving according the equation:       ∂ ∂ − ∂ ∂ = dt dr t r f r t Dt t r Df ) , , ( ) , , ( x x η . Here D/Dt=∂/∂t + (v-v t (r i )∇ , v t (r i ) is sedimentation velocity for the droplets of the radius r i . dr/dt describes changes of the number density function due to diffusional growth of cloud droplets dr/dt=AS/r, A=10 -10 m 2 /s, S=q v /q vs , is supersaturation, and ∂η/∂t is the nucleation rate. For the finite number of droplet size bins the condensation rate is given by: dt dm f C i i i d ∑ = . In the simulation presented here we use 16 classes of droplets (radius in range 0.78-24 µm). Sedimentation velocity is prescribed for each class according to Stokes law: v t (r)=Cr 2 , where C gives 1 cm/sec for 10 µm droplet. Corresponding author's address: Szymon P Malinowski, Institute of Geophysics, Warsaw University, ul. Pasteura 7, 02-093 Warszawa, Poland. e-mail: malina@fuw.edu.pl