On depression storage, its modeling and scale

Thumbnail Image



Abedini, Mohammad Javad

Journal Title

Journal ISSN

Volume Title


University of Guelph


Depression storage can be a significant storage element on a watershed surface, accounting for the retention of much water. This research was undertaken to investigate effects of surface storage elements on catchment response at a range of spatial scales and experimental settings. The spatial scales included small-scale laboratory experiments, small-scale field experiments and large-scale field experiments. Interactions of surface treatments, slope orientations, rainfall patterns, initial soil-water conditions and surface storage elements, and resulting effects on runoff response including the timing of outflow hydrographs, were considered at these spatial scales. Surface topography was measured with the aid of a laser scanner down to a 3-mm grid spacing. After the spatial location of depressional storages was delineated using digital elevation data, the results were linked to the GRID module of ARC/INFO via an indicator variable to derive polygon coverage of depressional areas versus nondepressional areas in a spatial context. From the pond analysis and associated spatial mapping, it became clear that most estimates and/or geometric characteristics relating to size and spatial location of depression storage, including area, volume and depth, are scale dependent. These geometric objects may best be described by resorting to fractal geometry, a popular tool for quantifying variability across scales. Indirect characterization of surface storage elements was achieved via observation of rainfall and the corresponding surface runoff at different spatial scales. From the analysis of runoff response data, it was found that when there is no infiltration, depression effects can be detected in response, i.e. effects due to size of depressions and spatial location of different sizes. In the presence of infiltration, separation of depression storage effects from infiltration effects on catchment response was found to be extremely difficult if not impossible. A simple holistic type approach was suggested to model depression storage. Application of the modeling approach showed that for simple depressional cases with no infiltration, differences in response due to various spatial patterns of depressions could be delineated well as measured by a coefficient of determination statistic, with 'R'2 values above 0.90 for the majority of cases. In such situations, the parameters took on values which appeared to have physical meaning in terms of physical characteristics such as mean time of travel and mean depth of depressional storage. For simple and more complex situations involving infiltration at a range of scales (from small plots to small watersheds), modeling applications showed that response could still be estimated well with a simple holistic model. However, the parameters can take on values which may or may not have any physical meaning at all. In all such modeling exercises, the mean time of travel was found to be the most stable; the mean depth of storage was quite stable and the recession constant was the least stable parameter.



Depression storage, Watersheds, Modeling, Water retention, Fractal geometry