A framework is developed to explore the coupling between the land and the atmospheric boundary layer using three-dimensional turbulence simulation over remotely sensed land surface images. The coupled set of equations is integrated with boundary conditions from day 221 of the Monsoon '90 experiment, and analysis is conducted to quantify the transmission of surface heterogeneity information into the atmospheric boundary layer (ABL). The large-eddy simulation (LES) model incorporates radiant energy availability; spatial fields of remotely observed surface cover, temperature, and moisture; and the ability to account for the separate contributions of soil and vegetation (i.e., two sources) to the mass and energy exchanges. This effort reflects a merging of active lines of research: the use of remotely sensed land surface properties to study water and energy fluxes and the use of LES to study impacts of surface variability on ABL processes. Analysis of the results reveals (1) that the combination of remotely sensed data and LES (in the absence of free parameters) yields regionally averaged land surface fluxes and ranges of spatial variability in the fluxes that compare well to similar measures from a network of flux measuring stations, (2) that the correlation between time-averaged surface and air temperatures is dependent on the length scale of the surface features, (3) that the horizontal standard deviation of mean air temperature decreased logarithmically with height in the atmospheric surface layer; and (4) that the mean air temperature contains spatial variability induced preferentially from variations in surface temperature occurring at scales >500--1000 m. Hence the feedback strength between the land and the atmosphere is shown to be scale-dependent for the range of length scales (i.e., ≤O(10 km)) studied here. ¿ 2001 American Geophysical Union