Large eddy simulation of land surface-atmosphere interactions with Boltzman Lattice
A Large-Eddy Simulation - Land Surface Model (LES-ALM) has been developed, which couples the WRF (Weather Research and Forecasting) large-eddy simulation model with a new land-surface model. The new land-surface model includes a multi-layer vegetation canopy, a new method for flux computation and a fine soil-layer configuration. To investigate heterogeneous land-surface processes, a number of numerical experiments are carried out for the FLUXPAT experiment site located in the Rur River catchment (50°53', 6°27'). Using the numerical results, we study how surface heterogeneity propagates in the atmospheric boundary layer and the dependence of the propagation on the dominant scale of land-surface heterogeneity. Fig. 1 shows the time-averaged sensible heat fluxes on levels of 0, 10, 40 and 160 m for two land-use patterns: (a) the original land-use pattern at the FLUXPAT experiment site and (b) the corresponding low-pass filtered land-use pattern. Close to the surface, the flux and land-use patterns are closely correlated. The correlations between the flux and land-use patterns are clearly identifiable at 160 m and visible at even higher levels. This suggests that land-surface heterogeneity is rather persistent in the atmospheric boundary layer, because certain types of large eddies (e.g. warm updrafts) preferably develop over particular land-use types (e.g. settlement) and propagate the land-surface signals to high levels in the atmospheric boundary layer.
The Boltzmann Lattice method is now being implemented in the framework of LES-ALM. With this, the air mass adjacent to the land-surface will be represented with a larger number of air parcels with sizes obeying specified probability distribution density functions. The exchanges between the land-surface and the atmosphere will be simulated by determining the motion of the air parcels. The numerical simulation will be carried out by means of parallel computing in cooperation with the Computing Center of the University of Cologne and Forschungszentrum Jülich.
Project: Modeling and Theoretical Investigation of multi-scale Interactions between Convection and Land-Surface Heterogeneity
An unsolved problem in weather and climate models is the parameterization of convection and meso-scale convective systems (MKS), which involve multi-scale interactions and land-surface heterogeneity. In this project, a new theory for the interactions between atmospheric convection and land-surface heterogeneity will be developed. A technique is proposed to quantify land-surface patterns in terms of content, scale and heterogeneity. The concept of information entropy is introduced to measure pattern content, i.e., the amount of information a pattern consists. This quantity is however insufficient to characterize how information is aggregated. It is thus proposed to conbine the wavelet and information entropy analyses to quantify patterns. We express a 2-D variable f(x,y) as,
f(x,y) = wn(x,y) + wn-1(x,y) + ... + w0(x,y) + f0(x,y)
where wn(x,y) is the nth wavelet component of f(x,y) and f0(x,y) a basis function. Fig. 2 shows an example of the wavelets decomposition of a near-surface temperature field. Assumed in this case is n = 1. The coarse-scale information of f(x,y) is represented by f0(x,y), the fine-scale information by w0(x,y) and w1(x,y). Locations, where convective cells develop, can be identified from the fine-scale information contained in w0(x,y). Our research is now focusing on (a) how convection is related to land-surface patterns by carrying out advanced simulations with LES-ALM and (b) the multi-scale interactions of the convective systems.
