Genetic algorithm-based allocation of LID practices to mitigate urban flooding

Urbanization has led to a decrease in infiltration and an increase in surface runoff which intensifies the risk, frequency, and extent of urban flood disasters. Although studies have been conducted to reduce urban flood damage by restoring the natural water cycle and thereby increasing the capacity of low impact development (LID) practices, there are few of them on land-use optimization to reduce surface runoff in urban areas. Thus, this study proposes an optimization approach that reallocates land-use parcels to reduce surface runoff using the genetic algorithm (GA) and the PCSWMM model. Incheon Gyeyang Techno-valley, one of the target districts of the 3rd New Town Project in the Seoul Metropolitan Area, South Korea, was selected as the target site. GA was embedded in the delineated catchment using the PCSWMM scripting tool to relocate land-use planning. Four LID practices, such as green roofs, permeable pavements, bio-retention, and infiltration trenches, were applied to each cell after considering the type of land-use planning. As a result, the rate of peak runoff decreased by 2.16%, 7.09%, and 7.01% under 2-, 10-, and 50-year return period rainfall, respectively. Although the updated land-use plan was not able to dramatically decrease the amount of runoff and peak flow rate, it was found that the relocation of LID practices with limited changes in the land-use plan can mitigate the peak flow rate during storm events in urban areas. Optimized land-use allocation must be considered during the planning stage because the overall capacity of low impact development practices depends on the land-use plan.

Automated summary

Urbanization leads to increased surface runoff and higher risk of urban flood disasters. This study proposes an optimization approach using genetic algorithms and the PCSWMM model to reallocate land-use parcels and reduce surface runoff.