Evaluation of Flood Susceptibility in Douala Estuary Cameroon using GIS, Remote Sensing and Logistic Regression

• Main Authors: Okolie, C., Mbouendeu, C.-A.N., Silwal, A., Mukherjee, S., Arungwa, I., Tella, A., Sirorattanakul, K., Mbouendeu, S.L., Johnson, N., Ligono, L., Nkwunonwo, U., Musa, H., Eric, T.F., Onwubiko, C., Adzandeh, A., Ndongo, B., Muhire, D.
• Format: Conference or Workshop Item
• Published: 2022
• Online Access: http://scholars.utp.edu.my/id/eprint/37616/; https://www.scopus.com/inward/record.uri?eid=2-s2.0-85167614071&partnerID=40&md5=123cc3fba11dc1bb4081911e3ef5bc68
• Institution: Universiti Teknologi Petronas

In flood mapping, it is often required to combine a variety of information sources and techniques. This integration can yield an optimal, multi-perspective and comprehensive assessment of the flood susceptibility in a target area such as the city of Douala in Cameroon. Available facts indicate that Douala is highly susceptible to flood especially in the coastal and nearshore areas. In this study, we integrated eleven flood conditioning factors (elevation, slope, topographic wetness index, terrain ruggedness index, distance to water bodies, drainage density, annual rainfall distribution, land use/land cover, soil texture, normalised difference vegetation index and modified normalised difference water index) to investigate the susceptibility to flooding. The adopted methodology combines Geographic Information System (GIS) and Remote sensing techniques, and the Logistic Regression (LR) statistical model. The generated flood map was divided into five susceptibility classes of very low, low, moderate, high and very high; and the overall accuracy of the prediction was assessed. In the results, low and very low flood susceptible areas have a susceptibility index less than or equal to 0.302 and very highly susceptible areas have an index greater than 0.788. The flood susceptibility map revealed that approximately 72.1 of Douala is less susceptible to flood while about 20 of its area (mostly urbanised and populated areas) are highly or very highly susceptible. The LR model shows accuracy, precision and F1 score higher than 0.85 which indicates its theoretical reliability. This result highlights the possibility to use open-source data and tools to assess susceptibility to flood with the aim of developing resilience plans or flood emergency response/preparedness plans. This methodology which is less complex and of reduced technical and financial constraints seems reliable and replicable. The susceptibility maps developed are useful tools for decision makers, city planners and local authorities for intelligent management of flood risk. Copyright © 2022 by the International Astronautical Federation (IAF). All rights reserved.