Hydrological Modeling for Flood Risk Analysis in Mathare, Kamukunji and Makadara Sub-counties in Nairobi, Kenya

Abstract

The objective of this study is to utilize Geographic Information System (GIS)-based hydrological model with remote sensing to analyze the risk of floods in the residential areas of Mathare, Kamukunji and Makadara sub-counties in Kenya. Data obtained from Copernicus Open Access Hub was used to obtain Sentinel data. Synthetic Aperture Radar (SAR) data was obtained from Sentinel Mission 1. Corrected data of SAR was used to create a 10m resolution Digital Elevation Model (DEM). Multispectral Satellite Imagery was used to obtain Land-use/Land-cover (LULC) using object-based classification. 10 classes of LULC were created. Several factors affecting flood risk were identified and mapped, among them were LULC and channel flow length, which mainly affected floods in the study area. The hydrological tools found in ArcGIS Software were used to divide the area of study into 4 catchments. The Hydrological Soil group of the 4 catchments was used to further sub-divide the study area into 17 sub-catchments in order to obtain accurate Curve Number (CN) values. HEC-GeoHMS (Hydrological Engineering Center’s Geospatial Hydrological Model System) which is an extension in ArcGIS software was used to obtain channel slope and flow length that were used in calculating the time of concentration for peak discharge and runoff. U.S Soil Conservation Service Technique Release 55 (SCS TR-55) model helped in predicting rainfall-induced floods. This model predicted peak discharge and runoff in the sub-catchments. Runoff was determined using equations in the model and the peak discharge was computed by the model’s graphical method. The overall Flood Hazard Map was produced in QGIS by overlaying the factor maps while the Flood Depth Risk Map was produced using ArcGIS software by summing up the values obtained from runoff. The results indicated that rainfall-induced flood is a serious problem with flood depth of 13-19.5cm, making the whole study area to be prone to floods especially the southern part occupied by Makadara sub county. Decrease in catchment’s flow length and increase in the number of impervious areas due to growth of urbanization increased flood risk in the area. The results of this study will be useful in coming up with solutions for flood risk control which include drainage systems that will improve the infiltration capacity of runoff, appropriate infrastructure e.g. green infrastructure and early warning systems such as sensors on rivers, drainage systems etc. The models used in this study helped predict floods in the study area. The use of remote sensing techniques helped update climatic factors such as rainfall, thereby solving the shortage issue on updated information on climatic conditions in the study area and also infrastructure for accurate prediction of floods