| Abstrakt: |
Flood occurrence is naturally annihilating, with enormous consequences on human lives, properties and resources. Thus, its occurrence worldwide, has serious environmental and socio-economic implications. In view of this, the incorporation of an all-encompassing flood management scheme is indispensable to mitigate the effects of the flood on human lives, resources and other livelihoods. The preliminary objective of this study was to assess the vulnerability of the Greater Accra Region of Ghana to flooding by applying the frequency ratio modeling technique to integrate nine conditioning factors derived from climatic and remote sensing datasets. The nine conditioning factors, which comprise the slope aspect, topographic roughness index (TRI), topographic positioning index (TPI), stream power index (SPI), sediment transport index (STI), profile curvature, LS factor, rainfall and drainage density were employed in outlining various flood prone zones within the study area using geographic information systems (GIS). 70% of flood inventory data (consisting of various locations of flood occurrence within the study area) were chosen randomly to produce the flood susceptibility model. The remaining 30% of the flood inventory data (validation data) were employed to validate and evaluate the efficacy of the flood vulnerability results generated. The flood vulnerability model (FVM) was discretised into five classes of very high (751.33 km2), high (333.19 km2), moderate (387.88 km2), low (479.33 km2) and very low (1590.55 km2) zones representing, respectively, 21.21%, 9.41%, 10.95%, 13.53% and 44.90% of the total study area size. This was succeeded by applying the area under the receiver operating characteristics (AUROC) curve on the output generated (flood vulnerability model) to assess its performance and accuracy. An AUROC score value of 0.788 was obtained, which shows that the model generated is accurate for the study area. |
Full text from SpringerLink