Scale dependence of hypsometric integrals: An analysis of southeast African basins

Previous studies have shown that the hypsometry of individual drainage basins is sensitive to both tectonic uplift rates and variations in erosional resistance of different lithological units. However, basin hypsometry has also been found to be related to the size, the shape and relief of the sampled area as well as factors such as the type of dominant erosional process. This study extends the evaluation of the controls on a basin's hypsometry by assessing factors that might influence the spatial distribution of hypsometric integrals at different scales. The method is applied to catchments draining the southeast margin of southern Africa (KwaZulu-Natal region) because of the region's rather uniform runoff, a rather well-constrained denudational history and a relatively simple coastward sequence of bedrock geology. Several sets of basins were extracted from an SRTM-derived 90 m-resolution DEM: one set of main basins (those that have outlets at sea level) and six sets of sub-basins of different size based on Strahler order. The mean area of the sub-basins sampled ranged from 9 km(2) for Strahler order 2 through to 3215 km(2) for Strahler order 6. In contrast to previous studies data from this study fail to show a correspondence between hypsometric integral and indices of basin dimension, such as basin area and basin relief Rather basin hypsometry in SE Africa varies spatially and with scale. In particular, a mountain-range scale, positive hypsometric integral anomaly situated between 10 and similar to 100 km of the coast line was observed but only in basins of Strahler order 5 and less. This anomaly is compatible with moderate crustal displacement focussed midway between the coast and the Drakensburg Escarpment. Some intermediate scale, positive hypsometric integral anomalies (in basins of Strahler order 4 and less) was also observed. In some cases, these anomalies correspond to differences in erosional resistance among major lithological groups. Finally, numerous local (basin-scale) positive hypsometric integral anomalies (in basins of Strahler order 3 and less) were also observed. In many cases, these latter anomalies are related to local differences in lithological resistance (e.g. the presence of dolerite sills and dikes). The analysis of the spatial variation of hypsometric integrals in basins of different size ranges is therefore a useful first-step in assessing factors that control landscape development. (c) 2007 Elsevier B.V. All rights reserved.