Sie verwenden einen veralteten Browser mit Sicherheitsschwachstellen und können die Funktionen dieser Webseite nicht nutzen.

Hier erfahren Sie, wie einfach Sie Ihren Browser aktualisieren können.

PhD-Project I4 by Georg Veh, University of Potsdam

Timescale: Oct.2015 – Sept.2018

Prof. Dr. Ariane Walz, University of Potsdam
Prof. Dr. Oliver Korup, University of Potsdam
Dr. Sigrid Rössner, Helmholtz Centre Potsdam GFZ German Research Centre for Geosciences

Background A disproportionate increase in population in mountainous regions worldwide currently coincides with climate-accelerated rates of natural hazards (Clagues et al. 2012). Glacial lake outburst floods (GLOFs) are an example for such a hazard. GLOFs are a special case of flooding that results from sudden release of meltwater ponded behind natural dams, and their connection to climate change are intensely debated (Haeberli 2013). Glacial meltwater lakes occupy the highest parts of a mountain range, so their potential energy can be transformed to hazardous and highly destructive flows (e.g. Huggel et al. 2011). The combined effects of increased human exposure (and thus vulnerability) and climate change on GLOF hazards and the corre-sponding changes to hazards and risk are subject of this proposal. Although satellite-based remote sensing is an efficient analysis and monitoring tool (e.g. Bolch et al., 2011), remaining uncertainties are substantial because of the complexity of the involved processes and the limited temporal repeat rate of data acquisitions. The same holds for dam-age potential from an increase in development as well as urbanization of mountain areas (e.g. Nussbaumer et al. 2014, accepted). Additional uncertainty pertains to GLOFs that are recon-structed from remotely sensed data and to future GLOFs projected from corresponding cli-mate simulations.

Objectives and Methods
We explicitly acknowldege these considerable uncertainties by employing a Bayesian Network (BN) approach for estimating GLOF risk at test sites. We plan to use both a data driven BN learning process as well as an expert-based network structure. The PhD-project aims at quantifying the contemporary and future risk from glacial lake outburst floods for selected sites in the Himalayas. Such risk appraisals are rare, though highly desirable in times of rapidly melting mountain glaciers and commensurately expanding meltwater lakes. Methods for this project will involve analyses of repeat satellite imagery, modelling of lake outburst and flood propagation, simulation of land-use and population scenarios, and assessment of these transient variables and their uncertainties using Bayesian networks.

Georg Veh is based at the research team “Landscape Management” of the University of Potsdam.