Improving urban flood modelling using opportunistic data
Urban floods are an increasing threat to the world due to two main developments: climate change and rapid urbanisation. Urban floods result in economic damage, services issues and problems for continuation of daily life. To cope with the growing impact of this type of disaster, hydrological models are used for flood mapping and modelling. Improvements of urban flood modelling are investigated by analysing opportunistic data sources. For this research, personal weather stations (PWS), digital images, notifications from citizens and combined sewer overflow (CSO) data are analysed. This is applied to pluvial urban floods in two case areas in the Netherlands, Amsterdam and Eindhoven. Each opportunistic data source addresses one of the conceptual urban flood model developed for this research. The data addresses forcing (PWS), validation of model state (digital images and notifications) and validation of model output (CSO). PWS precipitation data was compared with reference radar data and almost each station underestimated radar precipitation. No significant differences are found in reference and opportunistic data forcing when applied to the conceptual urban flood model. Digital images and notifications are valuable as forcing or validation data, however, methods for collection and analysis need to be considered. The study is unable to explicitly link CSO magnitude and occurrence to water inconvenience and therefore use this opportunistic data source as validation data. Opportunistic data sources do have potential to improve urban flood simulations in terms of forcing data and validation, however, data quality issues need to be concerned.