Automating GIS input for distributed urban drainage modelling

Abstract

The development of distributed urban drainage models is becoming more important as cities prepare for the challenges associated with climate change such as more intense precipitation events (McCarthy et al. 2010; Allan, 2011; Simõeset et al. 2011; Blumensaat et al. 2012; Leitãoet al. 2012). GIS-based tools were developed to generate input datasets for a 1-D distributed urban drainage model for part of Toronto's combined area, resulting in an efficient model development process compared to those utilizing manual approaches. These automatic GIS-based tools included the delineation of Wet Weather Flow (WWF) subcatchments (stormwater) and Dry Weather Flow (DWF) subcatchments (sanitary). It also included the determination of the intensity of rainfall on a more detailed scale than the coarse coverage provided by the City's rain gauges and the traditional Thiessen polygon interpolation method. Through testing the new tools designed in ModelBuilder, it was determined that 66% and 52% of DWF and WWF subcatchments respectively, were automatically delineated to a degree where they would be "Acceptable" for input into the urban drainage model, InfoWorks CS. Although the rainfall tools were able to continuously interpolate measured rainfall (on a seemingly unprecedented basis),and generate over 700 virtual rain gauges, the validity of the approach remains imperfect due to irresolvable inconsistencies between the City's gauges and those used for validation purposes.