University of Twente Student Theses


Rainfall-Runoff Modelling for Flash Floods in Cuong Thinh Catchment; Yen Bai Province: Vietnam

Pedzisai, Ezra (2010) Rainfall-Runoff Modelling for Flash Floods in Cuong Thinh Catchment; Yen Bai Province: Vietnam.

[img] PDF
Abstract:An assessment of the generation of runoff from severe rain events and flood propagation was carried out on a 15km2 Cuong Thinh catchment in Yen Bai province, north-western Vietnam. Cuong No stream passes through the flash flood prone Yen Bai town towards the Red River confluence. Two distributed catchment models were coupled; LISEM, event-based 1D hydro-dynamic rainfall-runoff model with SOBEK, a 2D hydraulic flood propagation model using a characteristic high intensity short duration storm. LISEM simulated two runoff scenarios, one on a whole catchment hydrograph which was compared with SOBEK normal slope scenario. Secondly, 34 sub-catchments hydrographs were simulated and incorporated into the three SOBEK complex terrain scenarios namely; dike-break on terraced slopes, terraced slopes only and normal slope in the rice fields. On the one hand, using LISEM model, it was observed that the main stream sub-catchment upslope of the rice fields contributes immensely to the catchment runoff. On the other hand, results simulated on The whole catchment scenario revealed a short lag as a fast catchment response to severe rainfall despite high interception due to dense forest and plantation cover and important surface storage by numerous ponds in the catchment. Severe tropical monsoon storms initiate a kinematic wave on the upslope which is propagated as a flood wave upon entering the rice fields as the SOBEK scenarios confirmed. A runoff coefficient of 0.44, three hour duration and peak discharge of up to 140m3/s was the main result of the LISEM whole catchment scenario. However, the SOBEK scenario predicted 20m3/s peak discharge. The SOBEK scenarios indicated that on one hand dikes play an important role in storing the runoff during the initial stages. On the other hand, they overtop and break thereby initiating a flood wave that propagates down-slope in the terraced rice fields. Breaking dikes led to doubling the flood peak which prois propagated as one wave unlike in the other two scenarios showed two subdued flood peaks separated by a few hours. During a dike-break situation the flood extent highest while the flood depth was consistent in all three scenarios. In these flood scenarios complex topography in the rice fields has an important buffering and storage function by the dikes, inter-field ridges and natural ‘bottlenecks’. Natural ‘loops’ in the rice field edges help to break the flow velocity while the bottlenecks act as the ‘hydrological valves’ regulating flow from the rice fields. While the two models predict a similar volume of flow generated in the catchment, LISEM predicted an earlier flood peak and shorter high flow duration while SOBEK predicted a longer duration of the event. The results imply a shorter warning time in both models although LISEM predicted an earlier and higher peak flow than the SOBEK scenarios. The coupling of LISEM and SOBEK was crucial to understand catchment behaviour on a complex terrain.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page