Catchment-scale spatial targeting of flood management measures to reduce flood hazard: An end-to-end modelling approach applied to the East Rapti catchment, Nepal

Abstract

Globally, practical approaches to managing flood hazards are moving away from mitigation solely at the point of the impact, and towards an integrated catchment-scale approach which considers flood source areas, flow pathways of flood waters and impacted communities. The current method for managing the fluvial flood risk in Nepal, however, generally involves localised structural interventions in affected areas using a static and reactive approach. This method does not create long term resilience to the hazards. There is therefore the need to rely less on these large-scale structural measures and focus instead on sustainable and non-structural measures for flood mitigation that allow the catchments and communities within them to be more resilient.
The three-stage, end-to-end approach developed in this thesis provides a process to help shift towards an integrated catchment management for flood hazard reduction in Nepal. The approach centres on identifying flood water source areas within the catchment and spatially targeting flood management measures in these locations. Consideration is also given to the potential impact of future, flow magnitude increasing, land cover change such as deforestation and the abandonment of terraced agriculture that is evident in many Nepali catchments.
Stage 1 adopts SCIMAP-Flood, a catchment-scale decision support framework that identifies critical source areas for flood waters. The framework uses maps flood water generating areas based on spatial rainfall patterns and land cover, the incorporation of travel times across a catchment, and modelling of hydrological connectivity. Outputs are used to create catchment-scale flood management scenarios which target flood source areas; tested flood management measures include targeted afforestation, check dams in key sub-catchments and abandoned terrace restoration.
In Stage 2 the flood management scenarios are assessed using CRUM3, a physically-based, spatially distributed, catchment-scale hydrological model. The impact of the flood management measures can be evaluated throughout the catchment using the modelled change in discharge. Stage 3 uses LISFLOOD-FP, a 2D flood inundation model, to establish the change flood inundation patterns at key flood impacted communities within the catchment from the created flood management scenarios. Stage 2 and Stage 3 utilise a coupled hydrological-hydraulic modelling approach with the results from the CRUM3 model entering the LISFLOOD-FP model as inflow hydrographs.
The approach is applied to the East Rapti catchment, a 3,084 km2 sub-catchment of the Nayarani River in southern central Nepal. The catchment contains three river flow gauges (Lothar Khola [catchment area - 169 km2], Manahari Khola [427 km2] and Rapti River [471 km2]) placed within the main sub-catchments and eight rainfall gauges. Additional data used to drive the approach was attained from global datasets and acquired during fieldwork.
This thesis has researched the potential effectiveness of the implementation of flood management interventions at the catchment-scale and evidences an alternative approach to flood management that is applicable in both Nepal and the wider Himalayan Region. Based on the integrated modelling approach, the results predict that the high flow magnitudes in the East Rapti catchment can be reduced through a catchment-scale approach.
However, even with a combined approach of large scale spatially targeted afforestation and check dam implementation (Q99.9 decrease of <=5.3%), the use of solely catchment-scale flood management approaches to combat flood hazard might not be effective at reducing the flood impact to at-risk communities. A significant outcome from the catchment-scale modelling work was that there is a far greater potential for land use change to increase, rather than reduce through mitigation, flow magnitudes in the East Rapti catchment. The model results suggest that any land within the East Rapti catchment that is altered from existing forest will contribute to increasing the flow magnitude (Q99.9 increase of up to 48.2%).