Statistics of temperature fluctuations in a buoyancy dominated boundary layer flow simulated by a Large-eddy simulation model

Temperature fluctuations in an atmospheric convective boundary layer are investigated by means of Large Eddy Simulations (LES). A statistical characterization for both weak temperature fluctuations and strong temperature fluctuations has been found. Despite the nontriviality of the dynamics of temperature fluctuations, our data support the idea that the most relevant statistical properties can be captured solely in terms of two scaling exponents, characterizing the entire mixed layer. Such exponents control asymptotic (i.e. core and tails) rescaling properties of the probability density functions of equal-time temperature differences, $\Delta_r \theta$, between points separated by a distance ${\bm r}$. A link between statistical properties of large temperature fluctuations and geometrical properties of the set hosting such fluctuations is also provided. Finally, a possible application of our new findings to the problem of subgrid-scale parameterizations for the temperature field in a convective boundary layer is discussed.