Determining the optimal river gauge location for a flood early warning system in Uganda using HEC-RAS and AHP

Abstract

Flooding of the Manafwa River in Eastern Uganda causes significant damage in the district of Butaleja, and often occurs without advance warning. In 2012, the American Red Cross in Uganda requested MIT to develop a flood early warning system (FEWS) that can be used to expedite disaster response in the event of imminent flooding. A minimum flood warning lead time of 2 hours was requested by the American Red Cross. Although there is an existing river gauge situated upstream of Butaleja, at Busiu Bridge, it was not known whether flood warning thresholds from Busiu Bridge could provide an adequate lead time for emergency responses to reach Butaleja. The purpose of this study was therefore (1) to determine the travel time for a flood wave to reach an area of interest in Butaleja (defined by the Red Cross) from Busiu Bridge, (2) to determine the water level at Busiu Bridge that corresponds to subsequent overbank discharge in that region of Butaleja, and (3) to use the Analytic Hierarchy Process (AHP) to integrate stakeholder input with hydrologic and hydraulic modeling results in order to determine the optimal location for an additional river gauge on the Manafwa River upstream of the current gauge at Busiu Bridge. The hydraulic model HEC-RAS was used to simulate the flow of the Manafwa River through Butaleja. Based on flow data from a storm event in 2006 when precipitation was fairly widespread across the watershed, the HEC-RAS results indicated that the travel time for a flood wave at Busiu Bridge to reach Butaleja is 3.4 hours, which provides sufficient warning based upon the Red Cross criteria. It was determined that if the stage at Busiu Bridge reaches 1.90 m, over 70% of the river in the defined area of interest in Butaleja will experience bank-full stage and overbank discharge 3.4 hours later. Given a precipitation event concentrated in the Bududa highlands, the flood wave travel time can diminish to 2.5 hrs if incoming flows increase; therefore upstream gauges were considered to support the gauge at Busiu Bridge. Hydrologic analyses performed by Kaatz (2014) suggested that upstream gauges can be adequately used to forecast floods provided that the precipitation and river conditions are suitable for a slow-moving flood wave. By integrating the hydrologic and hydraulic analyses with stakeholder input, installing another river gauge upstream of Busiu Bridge at the Bridge near Bubulo Red Cross (BBRC) would provide additional accuracy and lead time for the flood early warning system. Water level monitoring efforts implemented in Butaleja would validate modeled flooding thresholds reported here.