Last September, sediment-laden floodwaters in Kathmandu buried embankments in sand and gravel, then carved new channels through them—a real-world confirmation of what computer models had predicted but underestimated. This matters because most flood models treat water alone, ignoring the sand and gravel that can choke channels and redirect floodwaters in unpredictable ways. The September 2024 event in Kathmandu provides a rare opportunity to test and improve models that integrate sediment transport, filling a critical gap in flood forecasting for mountain river catchments worldwide. If this research succeeds, it will produce validated computer models that predict not just where water goes, but where sediment will accumulate and where erosion will breach defences. For cities like Kathmandu, this could mean more reliable flood warnings and better-designed embankments and drainage systems. For other mountain catchments facing similar pressures from land-use change and intensifying rainfall, the findings could reshape how engineers plan flood infrastructure—reducing the risk of defences being overwhelmed by the very material they were built to contain.
View original technical description
In most mountain river catchments, upstream changes in land-use combined with changes in climate are generating increasingly intense floods that deposit large amounts of sand and gravel and scour new channel courses. The distribution of flood waters are much harder to predict in these scenarios, and so new computer models that integrate both water and sediment transport are required. We have run a series of extreme case flood scenarios that incorporate sediment transport for Kathmandu city in Nepal that were published in August 2024. The results indicated that if modeled as floods comprising just water, then the inundation was modest and embankments held back flood waves. In contrast, once sediment was incorporated, channels became choked and spilled into neighbouring floodplains, and new channel courses incised through floodplains. These were experimental models with highly extreme scenarios. However, on the 27th and 28th of September these hypothetical scenarios became a reality when unprecedented rainfall hit the city. The model forecasts described above appear to have under-estimated the impact of sedimentation and erosion during the floods, with embanked channels being filled with sand and gravel, and water breaking through embankments to form new channels through partially developed floodplain regions. This confirmation of the impact of erosion and sedimentation in this setting provides us with a unique opportunity to provide a world leading exemplar of these processes and an opportunity to test alternative hypotheses concerning the drivers of this event. We already have high resolution digital topography in a setting with a high density of meteorological and hydrological stations that recorded the event. We now also have the record of high water mark, sediment thicknesses and erosional scour and the grainsize of the sediment. This means we will be able to test the extent to which sediment was remobilised within the channel systems, or whether additional external inputs from processes such as upstream sand mining and landsliding were a necessary condition for this event. In addition to testing detailed forecasts of sediment-rich flood water impacts, we will have the opportunity to test for controls on the accumulation of sediment throughout the catchment linked to the geometry of river channel networks. For example, where a tributary converges with a main channel, how will the backwater effects in the tributary affect sedimentation, and similarly, areas where there are low gradient channels and gorges both appear to have been impacted by localised enhancement of flood inundation by sediment accumulation.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know