Systematic planning of stormwater collection system improvement for current design storms in the city of Dallas

Managing urban stormwater, from quality and quality perspectives, is one of the greatest challenges in every growing metroplex. In the city of Dallas, TX, urban water management has been identified by city authorities as one of the key environmental health challenges. To address it, decision-makers need to know about the location and scope of investment within their stormwater drainage networks. This study presents a framework that identifies areas/subwatersheds that are prone to stormwater inlet overflow during design storms. To apply the framework to the case study, stormwater network data for all watersheds in the city of Dallas were collected and used to create SWMM 5.0 models to mimic hydrologic performance of the collection system during design storms. By identifying the overflowing stormwater inlets, their contributing subwatershed area, i.e., opportunity subwatersheds, were also recognized. The results showed that 6%, 10%, and 15% of the modelled area in the city are opportunity subwatersheds for the 2-, 10-, and 100-year design storms respectively. Upon classification of the overflowing results into five classes of severity, the results also showed that the opportunity subwatersheds in the very high category increase from 0.8 % to 2.8 % as the design storms increase. Finally, a series of decision tree classification models were created to predict the level of overflow severities based on subwatershed and network features. The decision trees showed that overflowing inlet depth is a significant contributing factor and, with an F-1 score >0.7 and Gini value<0.3, non-opportunity subwatersheds as well as the overflowing severity classes for the 100-year scenario can be predicted. To summarize, the presented framework assists decision-makers to further guide their stormwater investments and predict the drainage network response to different rainfall events.