Impact of connectivity on the modeling of overland flow within semiarid shrubland environments

The objective of this study is the evaluation of the spatial variability and intrinsic connectivity features of model input parameters for the parameterization of process-based, spatially distributed overland flow models. Parameter scaling tools based on the statistical and geostatistical properties of an extensive field data set were developed. These allowed the reproduction of the spatial heterogeneity of model parameters associated with the soil- and vegetation-related properties of semiarid shrubland environments to a varying degree. The outcome of the study emphasizes that connectivity plays a fundamental role in the modeling of water fluxes within semiarid catchments. The larger the degree to which connected features are represented, the better the model performance. In contrast, the parameterization approaches that did not contain connected patterns of parameter values performed comparatively poorly. A spatially connected overland flow model therefore enabled the generation of realistic overland flow patterns that qualitatively resembles field surveys of overland flow generation not only at the outlet of the model domains but also within the catchments without the need of calibration.