The land surface is characterized by pronounced spatial heterogeneity that spans a wide range of scales. This heterogeneity affects the surface energy and water budgets, as well as the land-atmosphere exchanges of momentum, heat, water and other constituents, through a number of highly nonlinear processes. The resolution of present-day Earth (or climate) system models is still too coarse to explicitly capture the effects of surface heterogeneity, which therefore needs to be parameterized within the framework of complex and nonlinear land surface process schemes. The effects of surface heterogeneity are here grouped in two categories, which we define as ''aggregation'' and ''dynamical'' effects. Models of aggregation effects attempt to calculate the contribution of different subgrid scale surface types to the grid box average energy and water budgets and surface-atmosphere exchanges. Such models have been based on discrete approaches, whereby heterogeneity is described in terms of a finite number of subgrid ''tiles'' or ''patches,'' and on continuous approaches, in which heterogeneity is described in terms of probability density functions. Subgrid scale aggregation has been shown to especially affect the surface latent and sensible heat fluxes, the simulation of snow, and the dynamics of soil moisture and runoff. Dynamical heterogeneity effects are associated with microscale and mesoscale circulations induced by heterogeneous surfaces. These circulations can influence boundary layer structure, cloud formation, precipitation, and vertical transfer of momentum, energy, and water up to the midtroposphere. In the last decade or so, the importance of land surface heterogeneity representation has been increasingly recognized in a large number of new studies. This paper reviews and critically discusses different approaches that have been proposed to represent aggregation and dynamical effects of surface heterogeneity and their incorporation in land surface process schemes. Some of the methodologies discussed in this paper are of general nature and therefore can be of interest for problems of subgrid scale process description in other geophysical disciplines.
