Flood hazard modelling in Tsholotsho district, Zimbabwe

Abstract

Among all natural hazards, floods pose the greater threat to property, safety and economic well being of human communities the world over. This study focused on understanding the nature of flooding in areas that are outside of streams. Normally, flooding has been dealing with bursting of river banks but in Tsholotsho flooding has occurred in areas that are located far away from defined drainage networks. The study established the nature of problems associated with flooding as a preamble to flood hazard modelling. Logistic regression was applied in a spatial database that had been developed and managed within a GIS to estimate the prediction power of the environmental variables (height above channel base, slope of drainage basin, distance from the nearest stream and soil type) for flood occurrence. The results indicate that height above channel base significantly (p < 0.05) predicted flood hazard for Hambeni and Sheleni while distance away from the nearest channel was significant (p < 0.05) in predicting flood hazard for Ntibu. The flood hazard map represents flood occurrence for each pixel. The developed flood hazard map will be useful in mitigating the loss of property from future flood disasters in Hambeni, Ntibu and Sheleni.