Optimizing flood risk modelling with high-resolution remote sensing data and analytic hierarchy process

Gerald Albert Baeribameng Yiran, Clement Kwang, Lewis Blagogie

Research output: Contribution to journalArticlepeer-review

3 Citations (Scopus)

Abstract

Flooding is a global phenomenon with devastating effects on human lives, livelihoods, and properties. Flood risk management is key to reducing floods’ environmental and socio-economic impacts. The mapping and prediction accuracy of flood models have been a challenge in developing countries due to a lack of high-resolution remote sensing data. With the advent of drone technology, the challenge would be mitigated and afford more accurate spatial analysis of flood occurrence and prediction. This study focused on optimizing flood risk assessment in the Greater Accra Metropolitan Area, utilizing high-resolution imagery obtained from drones and Digital Elevation Models (DEMs) from Google Earth to understand the hydrologic processes and use them with other variables to map high flood-risk zones. An example-based classification workflow was used to classify and extract features or objects in the study area. Features were classified into bare soil (8.15%), building (30.7%), parking, pavement, and sidewalk (23.52%), road (17.74%), vegetation (18.51%), and water (1.34%). The extracted data was then reclassified as impervious (72%) and pervious surfaces (26.7%), while the remaining area was considered a water body. The DEM was used for slope, elevation, topographic wetness index (TWI), drainage density, and drainage basin. All these datasets were combined with geology using weighted overlay analysis in ArcMap. The weights of prediction variables were obtained from the pairwise comparison of the analytic hierarchy process (AHP). The flood risk map showed 13% high-risk zones in areas such as Kokomlemle, Christianborg, parts of Adabraka, and Osu. About 33% of the study area had a moderate risk of flooding, while 54% had a low risk of flooding. This approach, utilizing high-resolution imagery and AHP, surpasses previous models dependent on low-resolution satellite images and DEMs, offering a more refined understanding of flood-prone areas. Field observations validated our findings, confirming the accuracy of the identified flood-prone regions during heavy rainfall. The significance of this study lies in providing detailed, reliable flood risk information crucial for sustainable planning and emergency response, especially in an area under constant rezoning pressure for residential purposes.

Original languageEnglish
Article number111
JournalSN Social Sciences
Volume4
Issue number6
DOIs
Publication statusPublished - Jun 2024

Keywords

  • Accra
  • Flood risk management
  • Impervious surfaces
  • Rezoning
  • Sustainable planning

Fingerprint

Dive into the research topics of 'Optimizing flood risk modelling with high-resolution remote sensing data and analytic hierarchy process'. Together they form a unique fingerprint.

Cite this