Abstract. Since 2004, Halji village, home of the oldest Buddhist Monastery in north-western Nepal, has suffered from recurrent glacial lake outburst floods (GLOFs). A sudden englacial drainage of a supraglacial lake, located at a distance of 6.5 km from the village, was identified as the source of the flood. The topography of the lake basin was mapped by combining differential Global Positioning System (DGPS) measurements with a structure-from-motion (SFM) approach using terrestrial photographs. From this model the maximum filling capacity of the lake has been estimated as 1.06 × 10 6 m 3 with a maximum discharge of 77.8 m 3 s −1 , calculated using the empiric Clague-Mathews formula. A simulation of the flooded area employing a raster-based hydraulic model considering six scenarios of discharge volume and surface roughness did not result in a flooding of the village. However, both the village and the monastery are threatened by undercutting of the river bank formed by unconsolidated sediments, as it already happened in 2011. Further, the comparison of the GLOF occurrences with temperature and precipitation from the High Asia Reanalysis (HAR) data set for the period 2001-2011 suggests that the GLOF is climate-driven rather than generated by an extreme precipitation event. The calculation of geodetic mass balance and the analysis of satellite images showed a rapid thinning and retreat of Halji Glacier which will eventually lead to a decline of the lake basin. As the basin will persist for at least several years, effective mitigation measures should be considered. A further reinforcement of the gabion walls was suggested as an artificial lake drainage is not feasible given the difficult accessibility of the glacier.

Two‐dimensional (2D) hydraulic models are widely used as tools for flood hazard mapping and also to support flood risk management. Yet, only few models are capable of using high‐resolution terrain data (raster‐based Digital Terrain Model with 1 m spatial resolution) on a large scale (hundreds of square kilometres and more). Central to the model approach presented in this article is the raster‐based 2D model High Performance Computing version of FloodArea (FloodAreaHPC), which allows for multicore processing, thus being able to model large areas without having to make compromises regarding spatial details. The model has been applied for inundation modelling of rivers, dike breaks, and heavy rainfall runoff (pluvial flooding). For the latter case, the model is coupled with a hydrologic preprocessor data base which provides spatially and temporally variable runoff coefficients based on land use, soil, and slope. The case study presented in this study has an area of 144 km2 and is located close to Dortmund in Western Germany. The modelling results of two heavy rainfall scenarios, presented in analogue and digital flood hazard maps, were used in a Public Relations (PR) campaign to inform the public about pluvial flood risk and possible mitigation measures.

Abstract. The awareness of pluvial (rain-related) flood risk has grown significantly in the past few years but pluvial flooding is not handled with the same intensity throughout Europe. A variety of methods and modelling technologies are used to assess pluvial flood hazard and risk and to develop suggestions for flood mitigation measures. A brief overview of current model approaches is followed by the description of a modelling methodology that has been developed throughout the last 15 years with the focus on processing large scale areas. Experiences from several projects show that only high quality models of whole catchment areas yield results with enough accuracy to gain credibility among stakeholders, planners and the public. As a best practice example shows, the model approach also helps to plan effective decentral flood protection measures. To ensure successful flood risk management, a long-term preservation of flood risk awareness among local authorities and the public is necessary.