Flash floods in Cebu: monitoring, modelling and preventing

Abstract

Flash floods are a damaging and recurring problem in Cebu city, Philippines. Very little data is known about the intensities and precipitation amounts and the resulting river discharges. This research project firstly aims to gather as much data as possible on precipitation and river discharges that could cause the floods, it focuses on a small catchment in the city called the Mahiga catchment. Data is gathered by installing three tipping buckets and two trail cameras. The cameras were able to calculate the river discharges using an innovative open-source program called OpenRiverCam. Thanks to this program a hydrograph can be
made of the river for each precipitation event. The used cameras were trail cameras of the Brand Bushnell. During this project it was concluded that, due to their unreliability, using trail cameras with OpenRiverCam is really not recommended. Security cameras with a Raspberry Pi are more suited. Due to bad luck with the weather and faulty material only three different hydrographs could be made during our time abroad (10 weeks). These hydrographs however remained useful for the second part of this research project. The second part consists of modelling the discharge of the Mahiga catchment to different
precipitation amounts using HEC-RAS. HEC-RAS is a computer program meaning Hydrologic Engineering Center’s River Analysis System. The model has been calibrated using the gathered precipitation data from the tipping buckets and the discharge results from OpenRiverCam. Graphs have been made about discharges and accumulated volumes and rating curves. The accuracy of the model is reasonable but should be improved using more discharge events. What stood out was the high infiltration rate and the fast response time of the Mahiga catchment. In section three, the results from the HEC-RAS model are used to understand the impact gabion dams make on reducing the peak flow in the Mahiga creek.
The third part summarises the effectiveness of the gabion dams in preventing flash floods. Unfortunately there is no ’real’ flash flood event captured by the tipping buckets, so three precipitation events are used based on analog measurements of a tipping bucket nearby the catchment. The gabion dams are tested on a maximum precipitation intensity of 35 mm/h, 30 mm/h and 25 mm/h with a total amount of 40 mm. Higher amounts of total precipitation
are realistic, but have a larger time duration and are not considered flash floods anymore. The volume that gabion dams can retain is too little for these large amounts of precipitation and are therefore not in the scope of this report. The results show that with at least five gabion dams, the peak flow reduces for all above mentioned precipitation intensities, but for the 35 mm/h it is getting less effective. The model also showed that the effectiveness is very dependent on the volume that can be retained by the dams. Maintenance of the gabion dams is therefore of crucial importance especially with the large amount of sediments and
debris in the creek.