We investigate processes of bedrock-core bar and island development in a bedrock-influenced anastomosed reach of the Sabie River, Kruger National Park (KNP), eastern South Africa. For sites subject to alluvial stripping during an extreme flood event (~4470-5630 m3 s-1) in 2012, pre- and post-flood aerial photographs and LiDAR data, 2D morphodynamic simulations, and field observations reveal that the thickest surviving alluvial deposits tend to be located over bedrock topographic lows. At a simulated peak discharge (~4500 m3 s-1), most sediment (sand, fine gravel) is mobile but localized deposition on bedrock topographic highs is possible. At lower simulated discharges (< 1,000 m3 s−1), topographic highs are not submerged, and deposition occurs in lower elevation areas, particularly in areas disconnected from the main channels during falling stage. Field observations suggest that in addition to discharge, rainwash between floods may redistribute sediments from bedrock topographic highs to lower elevation areas, and also highlight the critical role of vegetation colonization in bar stability, and in trapping of additional sediment and organics. These findings challenge the assumptions of preferential deposition on topographic highs that underpin previous analyses of Kruger National Park river dynamics, and are synthesized in a new conceptual model that demonstrates how initial bedrock topographic lows become topographic highs (bedrock‐core bars and islands) in the latter stages of sediment accumulation. The model provides particular insight into the development of mixed bedrock‐alluvial anastomosing along the Kruger National Park rivers, but similar processes of bar/island development likely occur along numerous other bedrock‐influenced rivers across dryland southern Africa and farther afield.
Tooth, S., Heritage, G. L., Milan, D. J., Tooth, S., & Heritage, G. (2020). Topographic, hydraulic, and vegetative controls on bar and island development in mixed bedrock-alluvial, multi-channeled, dryland rivers. Water Resources Research, 56(5), https://doi.org/10.1029/2019WR026101