34
systems removed on average 60% (SD=36%). Other studies have found that the effectiveness of various 1 filter designs varies, and filters can serve different purposes depending upon the design (Morgan et al. 2 2005, Hipp et al. 2006, Wilson et al. 2009). For example, zeolite filters did not remove any TSS from 3 synthetic stormwater, however, Xsorb filters removed almost 95% of TSS under the same conditions 4 (Hipp et al. 2006). Similarly, the AbTech Catch Basin Insert was almost four times better at removing 5 total suspended solids than the Aquashield Catch Basin Insert under the same conditions (Morgan et al. 6 2005). It is important that practitioners take into consideration the wide range of design variability when 7 selecting filtration systems for a given site (Morgan et al. 2005. Hipp et al. 2006). 8
Scaling 9
Green infrastructure must be properly scaled for its site. For most infrastructures, engineers use data 10 regarding the size of the drainage area, proportion of impervious surface, and historical rainfall data to 11 create a design which will be effective for a given size of storm. Generally, these variables are highly 12 site-dependent, and an ideal ratio of infrastructure size to drainage area has not been defined (Heitz et al. 13 2000, Akan 2002, Mungasavalli and Viraraghavan 2006). However, for constructed wetlands, design 14 guidelines recommend a ratio of wetland area to drainage area of at least 1 to 100 (Carleton et al. 2001). 15 To examine the effect that scale might have on variation in effectiveness, we compared the removal 16 efficiency for TSS, TN, and the reduction in runoff volume and peak flow for wetlands which were small 17 for their drainage areas to those which were properly sized. Properly scaled wetlands (infrastructure to 18 drainage area ratio ≥ 0.01) were no more effective at removing TN or TSS from effluent than those with a 19 ratio below 0.01. Although scaling is expected to be important for green infrastructure design, we didn’t 20 find sizing to be a prominent source of variability in constructed wetland performance. This is consistent 21 with the results of Carleton et al. (2001) who found that detention time and hydraulic loading rates (the 22 amount of water flowing into a wetland each day) were the most important determinants of the 23 effectiveness of constructed wetlands, regardless of the scale. 24
Geographic Variation and Cold Climate Studies 25
Variability in effectiveness of green infrastructure may be related to geographical and climatic variation. 26 For example, cold temperatures are known to adversely affect pollution removal and infiltration capacity 27 (Roseen et al. 2009). Although we were not able to directly test this, it is likely that climatic and 28 geographic variation play a role in the variability we found in the effectiveness of green infrastructures. 29
30
Seasonal variability in runoff infiltration, due to lower hydraulic conductivity in lower temperatures, has 31 been demonstrated in climates with frequent winter freezes (Emerson and Traver 2008). Concentrations 32 of pollutants such as chlorine are higher in winter runoff in places where road salt is used (Semadeni-33 Davies 2006). Some, but not all, infrastructures have reduced winter performance at removing pollutants. 34 For example, winter and summer TSS removal performances were similar for filtration, bioinfiltration, 35 and retention infrastructure but winter removal efficiency declined for stone swales and hydrodynamic 36 separators (Roseen et al. 2009). Seasonal effects can also vary by pollutant. Previous studies have shown 37 that wet detention ponds show decreased removal efficiency for lead, zinc, and TSS during winter, but no 38 declines in cadmium and copper removal (Semadeni-Davies 2006). 39
40
Regional differences in climate may also impact the effectiveness of green infrastructure. Studies in our 41 dataset were primarily located in humid continental, marine west coast and dry summer subtropical 42 temperate climates. We were particularly interested in studies in humid continental climates since this is 43
