Probabilistic Flood Extent Estimates from Social Media Flood Observations
Tom Brouwer1,2, Dirk Eilander1, Arnejan van Loenen1, Martijn J. Booij2, Kathelijne M. Wijnberg2, Jan S. Verkade1, and Jurjen Wagemaker31Deltares, Delft, Boussinesqweg 1, 2629 HV, The Netherlands 2Dept. of Water Engineering and Management, University of Twente, Enschede, Drienerlolaan 5, 7522NB, The Netherlands 3FloodTags, The Hague, Binckhorstlaan 36, 2511 BE, The Netherlands
Received: 24 Nov 2016 – Accepted: 25 Nov 2016 – Published: 25 Nov 2016
Abstract. The increasing number and severity of floods, driven by phenomena such as urbanization, deforestation, subsidence and climate change, creates a growing need for accurate and timely flood maps. This research focussed on creating flood maps using user generated content from Twitter. Twitter data has added value over traditional methods such as remote sensing and hydraulic models, since the data is available almost instantly, in contrast to remote sensing and requires less detail than hydraulic models. Deterministic flood maps created using these data showed good performance (F(2) = 0.69) for a case study in York (UK). For York the main source of uncertainty in the probabilistic flood maps was found to be the error of the locations derived from the Twitter data. Errors in the elevation data and parameters of the applied algorithm contributed less to flood extent uncertainty. Although the generated probabilistic maps tended to overestimate the actual probability of flooding, they gave a reasonable representation of flood extent uncertainty in the area. This study illustrates that inherently uncertain data from social media can be used to derive information about flooding.
Brouwer, T., Eilander, D., van Loenen, A., Booij, M. J., Wijnberg, K. M., Verkade, J. S., and Wagemaker, J.: Probabilistic Flood Extent Estimates from Social Media Flood Observations, Nat. Hazards Earth Syst. Sci. Discuss., doi:10.5194/nhess-2016-376, in review, 2016.