30
Removal Efficiency for TSS and TN 1
Our analysis indicated that green infrastructure generally succeeds in reducing both TSS and TN event 2 mean concentration. The ability of an infrastructure to reduce TSS was highly variable between and 3 within infrastructure types, removing between 58 and 86% of TSS (Figure 2). Weighted standard 4 deviations ranged from 28.5% (Detention) to 131.0% (Constructed Wetlands) (Figure 2A). Variability 5 was high and without access to the raw data and standard deviations for all sites, it was not possible to 6 determine the most effective types of green infrastructure. Detention basins had the lowest standard 7 deviation and therefore appear to be the most consistent in performance. Constructed wetlands were the 8 least consistent in their performance, evident in the very large standard deviation in TSS removal 9 efficiency. This may be because wetlands sometimes release sediments during very large storm events 10 (Carleton et al 2001). Despite the variability, we conclude that green infrastructure—with the possible 11 exception of constructed wetlands—is a viable option for reducing TSS concentration in stormwater and 12 potentially reducing the associated negative impacts on aquatic ecosystem health related to sediment and 13 particulate heavy metals. 14
15
Although green infrastructure reduced concentrations of TSS by 58 to 80%, we found that they were less 16 successful at ameliorating nitrogen pollution; no infrastructure consistently reduced the concentration of 17 TN by more than 58% (Figure 2B). Small sample sizes, ranging from 8 sites (Permeable Pavement) to 18 40 sites (Bioinfiltration) and a high degree of variability (between storm-event standard deviation: 0.95 – 19 83.05 %) hinder our ability to determine the most effective infrastructure types. Weighted standard 20 deviations ranged from 10.0 (for filtration devices) to 287.7 (for green roofs). Our findings are consistent 21 with other studies indicating that, while green infrastructure is successful at reducing concentration of TN 22 (USEPA 2000), dissolved pollution is generally more difficult to remove from stormwater than 23 particulate pollution (Vymazal 2006). 24
25
Reduction in Runoff Volume and Peak Flow 26
In addition to examining water quality measures, we also evaluated the effectiveness of green 27 infrastructures at reducing runoff volume and peak flow. Runoff quantify is particularly important 28 because it is strongly related to pollution removal. This is because reductions in runoff volume, even 29 absent any change in pollutant concentration, result in lower total pollutant loads entering stormwater 30 systems. 31
32
In our sample of papers, runoff volume and peak flow reductions were reported less frequently than TN 33 or TSS removal efficiency (Figure 2C). We calculated weighted average runoff volume or peak flow 34 reductions for permeable pavement, bioinfiltration, and green roofs, which generally reduced both peak 35 flow and runoff volume. Reductions in average peak flow ranged from 52 to 70% while 57 to 85% of 36 runoff volume was mitigated by green infrastructure (Table 2). Detention and the other categories of 37 green infrastructure did not have sufficient data for analysis. 38
39
The potential of green infrastructure to reduce peak flow by 50% or more may alleviate impacts on 40 aquatic systems such as overflows in combined sewer systems, flooding, and structural changes including 41 erosion, bank scouring and stream entrenchment in waterways. Additionally, reducing overall runoff 42 volume has the potential to reduce the risk of flooding and combined sewer overflows and increase 43 groundwater recharge. 44
