The importance of incorporating landscape heterogeneity into climate models has been recognized by a number of researchers. However, attempts to relax the assumption of uniform, homogeneous and complete vegetation cover have generally been limited to modeling each component of the landscape (e.g., trees, grasses, shrubs) as a homogeneous unit and then combining the results for each type, weighted by the percentage of ground area it covers (i.e., an area‐weighted average is computed).The purpose of this paper is to demonstrate that substantial errors in the representation of vegetation‐atmosphere interactions can arise from the assumption of landscape homogeneity. Even when weighted averages are employed as a first attempt to consider landscape heterogeneity, the underlying assumption of homogeneity of the vegetation patches can lean to significant errors. To demonstrate an example of how this can occur, vegetation canopy albedo derived from weighted averages is compared to that computed with explicit consideration of canopy heterogeneity.