Abstract
Many countries still lack national digital elevation models (DEMs) and have to rely on global datasets, which can negatively influence the reliability of flood model results. Mozambique is considered the most risk prone country for floods in Southern Africa. In this study a quality and accuracy assessment of two global DEMs (Shuttle Radar Topography Mission (SRTM) and HYDRO1K) is presented for a simple static flood inundation model of lower Limpopo Basin. This is accomplished with a local fit and vertical accuracy assessment of global datasets on a local scale as well as simulations of flood extent in the floodplain carried out by filling the DEMs with water according to the 2000 flood event. The results from the vertical accuracy assessment show that global DEMs can be used on a local scale. However, flood simulations performed on original DEMs contain inadequacies and are misleading with both under- and overestimation of the flooded area, while simulation performed on locally fitted DEMs shows a better agreement with the actual event. This study clearly shows that DEMs with questionable accuracy and resolution should be used with great caution in flood inundation modelling because they could result in deceptive model predictions, and lead to devastating after-effects in risk prone areas.
Acknowledgments
Thanks to SIDA for funding the field trip to Mozambique through Minor Field Studies, and to our colleagues; Mussa Achimo, Elidio Massuanganhe, Sandra Sitoe and Fortunato Manuel Cuamba from Department of Geology, Maputo, Mozambique for the support and to Antonio Miambo from CENACARTA for his generous help. We would also like to thank Bengt Lundén and Hildred Crill who made several useful suggestions on how to improve this paper.