Planning considerations - Living Roofs in Integrated Urban Water Systems - page 84

Figure 3.8
Conceptual
representation of a living
roof in SWMM 5.1
(adapted from Rossman
2009)

Rainfall

ET

Runon

Surface layer

Overlow

Iniltration

“Soil layer” = growing media

Percolation

To roof vertical
drainage

Drainage layer

at a maximum, the growing media will only dry to the wilting point by ET
between storms. Rainfall that exceeds the ield capacity eventually discharges
through the drainage mat, and then to the roof's vertical drainage system (e.g.,
gutters and downspouts).
Additional input data needs include climate conditions. For design storm
simulations, only a precipitation depth and/or hyetograph is needed. For con-
tinuous simulations, long-term historical precipitation records or derived pre-
cipitation patterns are needed, ideally in time steps of one hour or less. ET
must also be speciied. Either may be available from local weather stations. ET
from weather station records is typically potential ET calculated from climate
conditions and one or more models such as Penman-Monteith, rather than a
direct measurement. In SWMM, a user- speciied ET rate is feasible. As dis-
cussed in Chapter 2 , living roof ET is signiicantly inluenced by available mois-
ture in the growing media, such that living roof ET may be signiicantly
different from what is predicted by potential ET models. The method of ET cal-
culation has been shown to effect hydrologic models predicting per-event
retention (DiGiovanni et al. 2013; Stovin et al. 2013), but does not appear to
have a signiicant inluence on the net effect on long-term water balances
(DiGiovanni et al. 2013; Voyde 2011).
SWMM output includes runoff hydrographs for individual storm events, and
low-rate or volume frequency spectrum curves (e.g., Figure 2.3 ), when run in
continuous simulation. Continuous simulation enables performance characteriza-
tion that accounts for a wide range of climate factors, and acknowledges varying
hydrologic response. An additional beneit of using a hydrologic model such as
SWMM is the ability to integrate living roofs into a larger site or watershed scale
model along with other SCMs.
Studies to verify the accuracy of SWMM 5.1 GI SCM simulations are currently
being pursued by many researchers. The conclusions of a few studies using earlier
versions of SWMM (Alfredo et al. 2010; Burszta- Adamiak and Mrowiec 2013;
Palla et al. 2008) are not necessarily applicable to version 5.1, as the algorithms
in the model are different. Nevertheless, SWMM has a long and successful history
with signiicant user engagement and a publicly accessible online listserv.