The effect of chessboard variability of the surface fluxes on the aggregated turbulence fields in a convective atmospheric surface layer

To what degree the variability of surface features can be identified in the turbulent signals observed in the atmospheric boundary layer is still an u nresolved problem. This was investigated by conducting an analytical experi ment for a one-dimensional 'chessboard'-type surface-flux distribution on t he basis of local free convection scaling. The results showed that, due to their nonlinear dependency on the surface fluxes, the dimensionless gradien ts of the mean quantities and the dimensionless standard deviations are alt ered by the surface-flux variability. Furthermore, passive scalars, such as humidity, are considerably more sensitive to surface variability than the main active scalar, temperature. However, the response of the gradients of the mean quantities is fairly negligible in the range of variability studie d herein as compared to that of the standard deviations, which were found t o be more sensitive to the surface-flux variability. In addition, the phase difference between the active and the passive scalar flux distribution str ongly affects the passive scalar turbulence. This dissimilarity between pas sive and active scalars, or between passive scalars when their source distr ibutions are different, brings into question the use of variance methods fo r the measurement of a scalar flux, such as evaporation, over variable surf aces. The classical Bowen ratio method, which depends on the validity of th e Reynolds analogy for the vertical gradients of the mean quantities, was s hown to be relatively more robust. However, under conditions of strong surf ace variability, it can also be expected to fail.