| This dissertation focuses on the question: "How does land surface heterogeneity affect exchanges of heat and moisture between the atmosphere and land surface?" The proper representation of heat and moisture transfer has been shown to be important in both numerical weather prediction and climate modeling. In the context of this question, we investigate an existing water and energy balance model and find that errors in the representation of ground heat flux are responsible for significant errors both in sensible heat flux and in surface temperature spatial variability. We recommend a new parameterization of thermal conductivity designed for easy implementation by land surface modelers. Incorporation of this parameterization as well as environmental controls on stomatal conductance, stability corrections, and a new moisture diffusion approximation into the model provides a good representation of heat and moisture dynamics on time scales of hours to months.; The second model application involves data collected during an August, 1994 Oklahoma field experiment. This experiment is described along with analyses of the distributions, variability and spatial scales associated with important parameters affecting land-atmosphere interactions. In general, the underlying parameter distributions are found to be skewed and represent multiple spatial scales. In addition, the range of dominant scales found for topography and vegetation parameters suggest that these processes operate on smaller spatial scales (less than 1 km) than what the range of scales found for soil parameters suggests (1-4 km). Topography, soils and vegetation are all found to exert significant controls on soil temperatures, while they are found to be insignificant for air temperature and humidity. This implies that land surface heterogeneity is the dominant control on spatial variability of energy fluxes.; Finally, modeled fluxes and states are analyzed to determine their spatial structure and scaling properties. The dominant scales are shown to vary with moisture condition, and the 5 cm soil moisture and latent heat flux are shown to exhibit multiscaling behavior. Empirical relationships for this behavior as a function of soil moisture are presented, which should be tested against future remotely sensed and in situ datsets. |
