Modelling Flood Susceptibility of Calabar City, Cross River State, Nigeria

Abstract

Flooding remains a major environmental problem in many parts of the world including Nigeria, causing untold losses. This study  modelled the flood susceptibility of Calabar Metropolis using frequency ratio and geographical information system. Direct field  observation and key informant approaches were used to map 127 previous and current flood locations in the study area from which a  database of historical flood occurrence was developed. Using the geostatistical analyst tool within a geographical information  environment the flood locations were split into twopartsin the ratio of 70:30 per cent for the training and testing processes, respectively.  Eight flood conditioning factors (elevation, slope, aspect, curvature topographic position index, topographic wetness index, land cover and  normalized difference vegetation index) were extracted from SRTM-30m DEM and Landsat 8 data accordingly, and used in the spatial  analysis of floodoccurrence within the geographical information systems (GIS). The choice of these flood conditioning factors was based  on literature and nature of the study area. The frequency ratio model was then developed using the training dataset. The final product was  the flood susceptibility map of the study area. The results of susceptibility map revealed, in terms of percentage of area covered, were  as follows: Very high – 9.37, High – 19.97, Moderate –31.79, Low – 28.28 and very low – 10.59. The order of importance in the  contribution of the conditioning factors to the model in terms of percentages were elevation (21%), TPI (19%), Slope (17%), TWI (14%),  Landcover (9%), NDVI (8%), aspect (6%) and curvature (6%). Thevalidation of the model was done using the Receiver Operating  Characteristics (ROC) curve. Thevalue of the area under the curve (AUC) of the ROC was 74.81 per cent. This was interpretate as acceptable  for the prediction of flood locations in Calabar City of Cross River State in Nigeria. However, the model could be  improved upon if more historical flood locations are identifiedinthe study area and/or the employment of other prediction techniques,  including artificial intelligence and machine learning. Meanwhile, the outcome of the present study would assist thelocal town planning  authorities as well as policy makers in land use planning and policy formulations to reduce environmental threats associated with floods. It  would further assist in the effective implementation of strategies to mitigate future flood disasters and losses in the study area.