Planning considerations - Living Roofs in Integrated Urban Water Systems - page 79

signiicantly greater. In fact, sites contributing to the results in Table 3.3 that
returned the highest CN s empirically demonstrated ield-measured I a from
> 6-10 mm (the amount of rainfall that was needed before runoff was
generated); whereas the calculated I a was signiicantly less at 2-6 mm.
Nonetheless, in order for consistent application of the CN method for a given site
or watershed, the CN s for all surfaces must be derived from consistent
assumption of I a . The CN s calculated for each site presented in Table 3.3 assumed
I a = 0.2 S for consistency with most existing local stormwater design guidelines
using TR-55.
Ample evidence from these and other studies indicates that there is a
minimum threshold below which runoff is not produced from extensive living
roofs, regardless of climate conditions (Bengtsson 2005; Carpenter and Kaluva-
kolanu 2011; Fassman-Beck et al. 2013; Palla et al. 2012; Stovin et al. 2012).
More detailed inspection of the data from the 16 sites contributing to Table 3.3
suggests that regardless of living roof coniguration, meaningful runoff (more
than a couple of mm) was not typically generated from living roofs in Chicago,
New York City (two roofs), Villanova, North Carolina (three roofs) or Portland
(two roofs) except for storms larger than approximately 20-25 mm of precipita-
tion. This should be acknowledged in planning considerations. In short, where
planning requires the use of the CN method, a step function is suggested:
• Runoff volume = 0 (i.e., CN ≤ 1) where:
P ≤ S w
S w = D LR × PAW
S w ≤ 20 to 25 mm
• Runoff volume for larger rainfall events, or for events that exceed the actual
moisture storage capacity ( S w ), is determined with a maximum of CN = 85.
Despite calculated values in Table 3.3 , an arbitrary maximum value of CN = 85
is suggested based on the signiicant limitations of the CN method and availa-
ble data as described herein.
Where P = design storm depth (mm), S w = maximum water storage in the growing
media (mm) per unit area of living roof, D LR = inished growing media depth (mm),
PAW = plant available water (percent) ( Section 2.4 ). The calculations are explored
in more detail in Section 4.1 . Empirical evidence shows that at some point,
further increasing media depth ( D LR ) does not correspond to an increase in water
storage ( S w ), or the threshold below which runoff is not generated.
In applications where full runoff hydrographs (low rate versus time) are
needed for design, the process just described may introduce further deviation
from the ability to mimic living roof hydrology. The unit hydrograph represents
the base hydrologic response of a catchment to a unit input of rainfall (1 cm or
1 in), and is a fundamental element to most runoff modeling computation. In
most applications (e.g., in HEC- HMS), generalized forms of a unit hydrograph are