The surface character of a suburban area is far from the uniform, smooth and flat planes over which current surface-layer theory is valid and where vertical eddy-fluxes can be assumed to be almost constant horizontally and vertically. The complexity of the surface introduces considerable variability into the atmosphere at small spatial scales. This variability is partly reduced and spatially-averaged by turbulent mixing but still leaves the concerns about the spatial representativeness of sensible heat flux measurements over a suburban area. The spatial scales of sensible heat flux variability are discussed in terms of the distribution of surface temperature and roughness elements. It is shown that : (1) an eddy-correlation measurement can be considered spatially representative, if its surface zone of influence (source area) is large enough to include a spatially representative sample of surface temperature and roughness elements. (2) a quantitative measure of spatial representativeness can be estimated by use of the two-dimensional Fourier transform of the surface temperature and roughness element distributions (i.e. by the normalized integrated variance spectrum). (3) the source area of an eddy correlation measurement may be evaluated by a numerical model based on a probability density function plume diffusion model. The source area model developed herein can also be used to estimate the relative influence of specific surface sources or sinks upon an eddy-flux measurement in the surface layer. These concepts are tested in a suburban residential area in Vancouver, B.C., Canada. Remotely sensed surface temperatures and a digitized roughness element inventory are used as data-bases for the Fourier transforms to develop representativeness criteria for eddy-flux measurements. A set of sensible heat flux measurements at six sites and the corresponding source area calculations are used to formulate recommendations for the objective evaluation of the spatial representativeness of sensible heat flux measurements over a suburban area. The validity of the suggested evaluation methods is confirmed by the observations. Internal boundary layer growth, estimated by the source area model, compares well with existing work. Some consequences of complex surfaces on the surface layer structure are briefly discussed. / Arts, Faculty of / Geography, Department of / Graduate