On the potential of social media information to complement remote sensing data in natural disaster management

Kurzfassung

This master's thesis investigated the possibilities of fusing data from the social network Twitter with remote sensing data for the purposes of natural disaster management. The study was conducted on the example of Hurricane Florence in 2018, which led to significant flooding in the states of North and South Carolina. The work aims to answer the questions of how Twitter data can complement such traditional remote sensing data and how informative flood-related data is for supplementing flood maps. The results demonstrate that the proposed method of combining Twitter and remote sensing data has the potential for future use, but it requires a larger dataset of flood-related tweets. It was also found that social media data can improve urban flood maps, although with certain limitations. The proposed approaches to analyzing the informativeness of tweets include the use of both external (land cover product and elevation data from DEM) and internal data (results of tweet classification). The outcomes of this thesis may be used as the first steps in the systematic fusion of unsystematic social media data to complement remote sensing data in flood mapping.