Management of Stormwater through Green Infrastructure Systems in Coastal Urban Areas: Hydrological Modelling, Compound-Flood Response, and Water-Quality Performance

Abstract

In seven different GI scenarios, a representative 120 ha coastal urban drainage catchment, typical of Alexandria, Egypt, on the Mediterranean coast of the Nile Delta, was analyzed using SWMM 5.2 with its low impact development (LID) extension. The scenarios included four single-practice scenarios: green roof, permeable pavement, bioretention cell, and infiltration trench, each aimed at mitigating 20% of the impervious area, and three mixed portfolios with targets of 10%, 20%, and 30% impervious areas (MIX10, MIX20, MIX30). The catchment has an area-weighted imperviousness of 81.9%, and drainage is discharged from the tide gate to receive storm surge, astronomical tide, and sea-level rise (SLR) scenarios of +0.0, +0.5, and +1.0 m, together with an incrementally diminishing infiltration capacity due to rising shallow coastal groundwater levels. For a 10-year 2-hour return period design storm of 39.7 mm, baseline simulation provided 0.79 runoff coefficient, 14.09 m³/s of peak runoff, and 11,693 m³ of system flooding. Scenarios MIX30 reduced 27.9%, 35.1%, and 59.5% of runoff volume, peak runoff, and flood volume, respectively. Permeable pavement resulted in the largest flood volume reduction (56.9%) among single practices. Efficacy of GI systems decreased with an increase in storm size, as flood-volume reduction for MIX30 varied from 88.6% for a 2-year storm to 42.9% for a 50-year storm. A SLR scenario of +1.0 m increased baseline flooding by 17.1% and MIX30 flooding by 30.5%, hence decreasing the flood-volume reduction efficacy of MIX30 from 59.5% to 54.8%, including 204 m³ of water caused by limitations in infiltration associated with rising shallow groundwater level. Event loadings of total suspended solids (TSS), total nitrogen (TN), and total phosphorus (TP) were reduced with MIX30 by 44.4%, 40.5%, and 39.6%, respectively. A Monte Carlo Analysis (60 runs for each scenario) revealed that differences between BL, MIX20, and MIX30 scenarios were statistically significant for all response variables (one-way ANOVA and Tukey’s HSD, p < 0.001). From these findings, it follows that distributed GI can have a substantial effect on alleviating pluvial flooding in coastal low-lying areas; however, this effect is diminishing under both extreme rainfall events and rising sea levels and should be complemented by grey storm drainage systems and outfall management strategies within the overall coastal stormwater management program.