For a detailed description of the PhD-position:
http://www.vu.nl/nl/werken-bij-de-vu/vacatures/2014/296.asp

Contents of the research
Economic losses due to weather related natural hazards continue to rise rapidly in all regions of the world. For example, extreme flood events and storm surges all have a significant impact on society, and these impacts will increase in the future through climate change and socio economic developments. Much research to date has focused on the assessment and quantification of the natural hazard component of flood risk including the effects of climate change, which may lead to increased hazard magnitude and/or intensity in many regions. However, there is a growing understating that increasing exposure of people and economic assets has been the major cause of the long term increases in economic losses from weather and climate-related disasters. Some studies have investigated losses from natural in the past, and others assessed future increase in risk from natural hazard using simulation models or statistical extrapolation. The validation of these flood risk models is a difficult task, since extreme weather events are rare, and empirical data is on actual losses is scarce.

Tasks
This research will focus on river floods, and includes 3 activities:
(1) In the first part of the research he candidate will apply the GLOFRIS flood risk simulation model to model historic flood events in different parts of the globe. The analysis will focus on events with a temporal scale of 2-4 weeks. The model results will be validated using hydrological information such as river discharges and possibly damage information from e.g. insurers, EM-DAT and other sources

(2) the second part of the research will look into the availability of social media activity (Twitter, facebook, etc) at the time of the events that were simulated and validated in the first part. Different databases will be used, such as an existing twitter database for Indonesia. The social media data will be spatially plotted and compared with the spatial output of the flood risk/ inundation model. In this way, it is possible to see whether the spatial distribution of social media activity referring to flood hazards is within the area of modeled simulated floods, and to explore what the effect is of the social media on the behavior of people in response, or perhaps in preparation, of the flood events.

(3) Finally, an agent based decision model will be developed to simulate how people can respond just before, during and just after an event. The model will simulate the interaction between the government, households and possibly insurers and humanitarian organizations. Again, existing social media information will be integrated in this decision model to assess the potential of such information in helping people to prepare themselves against impacts.