Parametrization of turbulent fluxes over inhomogeneous landscapes

Authors:
Publisher website
Google Scholar
Abstract
Reasons for the nonclosure of the heat balance in the atmospheric boundary layers over natural land surfaces are analyzed. Results of measuring the heat-balance components over different land surfaces are used. The Cabauw (Netherlands) data (obtained throughout 1996 over a grass surface with intermittent shrubs and single trees) and the data from the Anchor station in Germany (measured over coniferous forest in 2000–2001) are analyzed. In all, the analysis involves about fifty thousand independent values of the heat-balance components measured in the experiments, which should be indicative of the reliability of the results obtained in the paper. The data have shown that the heat balance is not closed and the imbalance is 50–250 W/m2. The sum of the latent and sensible heat fluxes λE + H = STF is found to be systematically smaller than the difference between the net radiation and the heat flux into the ground R n − G. It is shown that the main cause of a systematic heat imbalance in the atmospheric boundary layers over inhomogeneous land surfaces is that the methods of surface-flux measurement and estimation are based on the theory that requires the hypothesis of stationarity and horizontal homogeneity. Direct data analysis has shown that the heat imbalance increases with landscape inhomogeneity. In the paper, a parametrization of the heat imbalance is carried out and the coefficient k f (z 0 ef /L ef ) is introduced as a measure of inhomogeneity. For this, data from the experiments FIFE, KUREX, TARTEX, SADE, etc., are also used. Empirical formulas are presented to refine the results of direct measurements and calculations of surface fluxes over natural (inhomogeneous) land surfaces from profile and standard (using bulk parametrizations) data. These formulas can also be used to determine surface fluxes over inhomogeneous underlying land surfaces in order to take into account so-called subgrid-scale effects in constructing prediction models.

This publication has 27 references indexed in Scilit: