Analysis of Scale Dependent Statistics of Quasi-Steady-State Large Eddy Simulations in the Time Domain

Abstract:

Recently, the authors have developed a novel Large Eddy Simulation (LES) infrastructure that has been designed from the outset to address problems of climate relevance. Addressing climate problems requires conducting LESs that reach realistic statistical quasi-equilibria when subject to realistic fixed and time varying large scale forcing. Reaching quasi-equilibria demands simulations of long duration.

A consequence of conducting long duration LES is that the domain must be sufficiently large to contain mesoscale variability that emerges during the simulation itself. In a pioneering paper, de Roode and Duynkerke (JAS 2004) sought to address this question in fixed forcing LES, and found large domains are required to capture the largest scales of variability for all but the simplest boundary layer configurations. However, in that paper, which considered analysis of spectra in the spatial domain, the scales of analysis are limited by the spatial extent of the LES domain.

In this presentation, we report the results of an analysis of scale dependent statistics in the time domain. Unlike analysis in the spatial domain, which is limited by the extent of the LES domain, analysis in the time domain is bound by no constraint other than the temporal length of the simulation. Moreover, in the context of fixed and time varying large scale forcing, analysis in the time domain allows investigation of the impacts of time varying forcing across all temporal scales independent of domain size constraints.