Available methods to determine the sensible and latent heat fluxes from the Earth's land surfaces are still relatively imprecise; this is due primarily to the inherent irregularity of the turbulent transport mechanisms and also to the pronounced variability of most natural land surfaces. While great progress has been made in the study of turbulence, until recently, surface variability has received relatively little attention in this context. Some thoughts are provided on approaching this issue by focusing on the stochastic structure of the land surface within the framework of homogeneous, i.e., stationary, random functions. In addition, suggestions are made regarding the likely applicability of classical concepts by relating the relevant scales of surface variability structure to those characterizing turbulent transport in the atmospheric boundary layer. ¿ 1998 American Geophysical Union