Water Air Soil Pollut (2007) 186:351–363

DOI 10.1007/s11270-007-9484-z

Low Impact Development Practices: A Review of Current

Research and Recommendations for Future Directions
Michael E. Dietz

Received: 1 June 2007 / Accepted: 28 July 2007 / Published online: 5 September 2007
# Springer Science + Business Media B.V. 2007

Abstract The low impact development (LID) ap- pavements continue to infiltrate even with frost in the
proach has been recommended as an alternative to ground. Although issues have been identified with
traditional stormwater design. Research on individual retention of certain pollutants, the LID approach
LID practices such as bioretention, pervious pave- has been found to result in increased retention of
ments, and grassed swales has increased in recent stormwater and pollutants on site, mimicking pre-
years. Bioretention cells have been effective in development hydrologic function. Future research
retaining large volumes of runoff and pollutants on needs have also been identified.
site, and consistently reduced concentrations of
certain pollutants such as metals. However, retention Keywords Bioretention . Green roof . Low impact
of certain pollutants such as nitrate–nitrogen and development . Pervious pavement
phosphorus has been problematic. Porous pavements
have been extremely effective in infiltrating storm-
water runoff. Concerns have been raised about 1 Introduction
groundwater contamination, but research has shown
that this is not a problem in most settings. Green roofs The effects of traditional development practices on
have been found to retain a large percentage of the hydrologic cycle have been well documented.
rainfall (63% on average) in a variety of climates. A Increases in the impervious surfaces associated with
common thread across bioretention, green roofs and urbanization have resulted in increased surface
grassed swales was found: the export of phosphorus. runoff (Hollis 1977; Jennings and Jarnagin 2002;
The issue appears to be linked to high phosphorus Waananen 1969), increased runoff velocity, de-
levels in the soil media, or possibly to fertilization of creased time of concentration (Leopold 1968), and
turf or planted areas. Solutions to this problem have decreased water quality (Makepeace et al. 1995; US
been recommended. Contrary to popular belief, EPA 1983). The earliest documentation of increased
research has shown that bioretention and pervious runoff from urban areas was in the late 1800s
(Kuichling 1889), and urban runoff continues to be
a leading cause of impairments in the nation’s
M. E. Dietz (*) waterways (US EPA 2002).
Department of Environment and Society, Low Impact Development (LID) was piloted in
Utah State University,
5215 Old Main Hill,
Maryland (Prince George’s County 1999) as a way to
Logan, UT 84322-5215, USA mitigate the negative effects of increasing urbaniza-
e-mail: michael.dietz@usu.edu tion and impervious surfaces. The preservation of the
352 Water Air Soil Pollut (2007) 186:351–363

pre-development hydrology of a site is the overall “preliminary” in this review. Also, it is not the
goal of LID. In contrast to typical stormwater design, intention of the author to endorse one product over
the LID approach advocates for more careful site another. Product names are only used to provide detail
design in the planning phases. The purpose of the site on specific research projects.
design is to preserve as much of the site in an
undisturbed condition, and where disturbance is
necessary, reduce the impact to the soils, vegetation, 2 Bioretention
and aquatic systems on the site. In contrast to
traditional stormwater treatment, which typically only Bioretention areas, or rain gardens, are depressed
mitigates peak flow rates, the use of LID will also areas in the landscape that are designed to accept
help to maintain the pre-development runoff volume. stormwater. They can be used in residential and
Cluster layouts, grass swales, rain gardens/bioreten- commercial settings, and are typically planted with
tion areas, and pervious pavements all reduce the shrubs, perennials, or trees, and covered with shred-
“effective impervious area” (Booth and Jackson 1997) ded hardwood bark mulch. The benefits of bioreten-
of a watershed, or the area that is directly connected tion areas include decreased surface runoff, increased
to the stormwater system. groundwater recharge, and pollutant treatment
Initial research on individual LID practices has through a variety of processes (Prince George’s
shown promising results. However many studies have County 1993). Several municipalities have created
occurred since an initial EPA literature review was bioretention standards. A highly detailed bioretention
published (US EPA 2000). New successes have been conservation practice standard is available through the
documented, but other unexpected outcomes have Wisconsin DNR. Site criteria, design specifications,
also arisen. In addition, questions are frequently construction guidance, and maintenance recommen-
raised in regards to the suitability of LID for all sites, dations are included in the standard (WI DNR 2006).
groundwater contamination, and winter performance Initial bioretention research focused on laboratory
of LID practices. The goal of this literature review is prototypes (Davis et al. 2001). High concentration
to present relevant research on the various LID reductions (>90%) were found for copper (Cu), lead
practices, and to synthesize the results so that the (Pb), and zinc (Zn). Nutrient concentrations were also
current status and future research needs of LID reduced: total Kjeldahl–nitrogen (TKN) retention was
investigations can be assessed. The focus of this 68%, and ammonia–nitrogen (NH3–N) retention was
review was research published in peer-reviewed 87%. The only nutrient not well retained by the
journals. However, to illustrate a point or corroborate system was nitrite + nitrate–nitrogen (NO3–N), which
a similar finding, studies published in reports or had a retention of 24%.
conference proceedings were occasionally referenced. Field investigations of bioretention have also been
If used, this research was always referred to as performed (Table 1). The first such investigation was

Table 1 Summary of bioretention pollutant retention

Location TSS NO3–N NH3–N TKN TP TN ON Cu Pb Zn Reference

Connecticut
Haddam – 67 82 26 −108 51 41 – – – Dietz and Clausen 2006
Maryland
Greenbelt – 16 – 52 65 49 – 97 >95 >95 Davis et al. 2003
Largo – 15 – 67 87 59 – 43 70 64 Davis et al. 2003
New Hampshire
Durham 96 27 – – – – – – – 99 Roseen et al. 2006
North Carolina
Greensboro −170 75 −1 −5 −240 40 – 99 81 98 Hunt et al. 2006
Chapel Hill – 13 86 45 65 40 – – – – Hunt et al. 2006
Water Air Soil Pollut (2007) 186:351–363 353

in Maryland, where synthetic runoff was applied to month period of study. If this overall retention of flow
two different bioretention areas, one in Largo, and is used to assess pollutant retention performance, the
one in Greenbelt, MD (Davis et al. 2003). Removal of system in Connecticut retained the vast majority of
Cu, Pb and Zn at the Greenbelt site was >95%; pollutants along with the flow.
however, at the Largo site removal was slightly less. The combination of phosphorus export and an
Similar variation in nutrient retention was found underdrain that is directly connected to the storm-
between the two sites (Table 1). In Connecticut, water system could cause more harm than good, if a
nutrient retention by soil media was generally lower, sensitive water body were downstream. Therefore, to
and in the case of total phosphorus (TP), more avoid this problem, the phosphorus content of the soil
actually left the system than entered it (Table 1). media used in a bioretention area should be examined,
NO3–N retention was higher in the Connecticut study and if it is very high, an alternative media should be
than in the Maryland study, especially during the used. In addition, an underdrain should be installed
second year (Dietz and Clausen 2006). High Zn, total only when the native soils have a low infiltration
petroleum hydrocarbon (TPH), and total suspended capacity. The minimum infiltration rate recommended
solids (TSS) retentions were found at another field in the Bioretention Manual is 1 in. h−1 (Winogradoff
study in New Hampshire (Roseen et al. 2006). 2002). If it is necessary to use an underdrain, it could
However NO3–N retention was low, which is consis- be drained to grade in a grassed or wooded area.
tent with other bioretention research. In North Another possible solution is to use a capped under-
Carolina, high metals retention by bioretention areas drain or controlled orifice, so that underdrain outflow
was also reported, but variable retention of nutrients can be increased if excessive ponding occurs, as
was found (Table 1). Export of TSS, TKN, NH3–N, recommended by Atchison et al. (2006). This may not
and TP were found in North Carolina (Table 1). be possible in certain situations, so as a last resort the
The export of TP noted in the Connecticut study drain could be connected to a standard stormwater
was attributed to the disturbance of the soils at the system.
beginning of the study, and did decrease over time Although NH3–N seems to be well retained by
(Dietz and Clausen 2005). However, phosphorus bioretention areas, retention of NO3–N tends to be
export from bioretention systems does not seem to low. This is due to the fact that the negatively charged
be an isolated phenomenon; similar findings have NO3–N ion does not adsorb well to soil particles. The
been noted in North Carolina (Hunt et al. 2006), and creation of NO3–N through mineralization and nitri-
in preliminary results from Ontario, Canada (Toronto fication of other forms of nitrogen in between
and Region Conservation 2006). In North Carolina, infiltration events has also been cited as a possible
the initial export of phosphorus has been attributed to mechanism for the low retention of NO3–N (Davis
high phosphorus content in the soil, or a high et al. 2001). Several researchers have performed
Phosphorus Index (Hunt et al. 2006). The TP export studies designed to increase the ability of a bioreten-
noted from the bioretention cell in Canada was tion area to treat NO3–N. An alternative design was
attributed to leaching of the mulch and organic soil proposed that involved raising the underdrain outlet
media (Toronto and Region Conservation 2006). pipe, to create a saturated zone in the bottom of the
It should be noted that the field studies in Mary- garden (Kim et al. 2003). The resulting condition
land were performed on unlined bioretention areas, would then be conducive to denitrification reactions,
whereas the Connecticut study was performed on a where NO3–N is converted to nitrogen gas (Korom
lined system. A liner is not a typical component of a 1992). In a laboratory experiment, shredded news-
rain garden; however it was used in the study in papers were found to be the most effective aid to this
Connecticut for mass balance calculations (Dietz and conversion of NO3–N in simulated bioretention
Clausen 2005). The flow mass balance for the columns, by providing a carbon source for the
Connecticut rain garden indicated that less than 1% denitrification reaction (Kim et al. 2003). This
of inflow water overflowed (Dietz and Clausen 2006). modification was tested in a field study in Connect-
In other words, this system, which was sized to icut, where increased treatment of total nitrogen (TN)
contain 2.5 cm (1 in.) of roof runoff, prevented 99% and NO3–N was found (Dietz and Clausen 2006). In
of roof runoff from leaving the site during the 24- North Carolina, significantly higher concentrations of
354 Water Air Soil Pollut (2007) 186:351–363

TKN, NH4–N, and TN were found in outflow from a tion media are likely the reason that infiltration will
similarly modified bioretention cell, as compared to a still occur despite frozen conditions.
traditionally designed cell (Hunt et al. 2006). The Bioretention design recommendations are incon-
researchers cite the conversion of organic forms of sistent. Most of the guidance has been focused on the
nitrogen to NH4–N as the suspected cause of the engineering community, for larger-scale design. A
increased NH4–N, TKN, and TN concentrations. manual from Wisconsin provides complete, easy to
Little data exist on the ability of bioretention areas follow guidance for homeowners on siting, sizing,
to reduce fecal coliform (FC) bacteria concentrations, digging, and planting a rain garden (WI DNR 2003).
a common indicator species of bacterial contamina- The sizing method used in this manual is based on the
tion. Although some grab sampling for FC bacteria WINSLAMM model, and the storage of 2.5 cm (1 in.)
was performed in the Connecticut study, inlet and of runoff from a roof (Bannerman, personal commu-
outflow concentrations were all <10 FCU 100 ml−1 nication). A similar manual for Connecticut also uses
(Dietz and Clausen 2005). Preliminary results from a the 2.5 design method (Dietz and Filchak 2006).
laboratory study indicate an average removal rate of The first bioretention design manual originated in
88% of FC bacteria in simulated bioretention columns Maryland (Prince George’s County 1993), and
(Rusciano and Obropta 2005). contained recommendations for the what media to
Increases in runoff temperature have been found use, and how to size bioretention. The more recent
as rain falls on impervious surfaces (LeBlanc et al. manual (Winogradoff 2002) contains updated recom-
1997), but there is little data on how well bioreten- mendations for the media, sizing calculations, and
tion areas attenuate temperature. In Connecticut, no ponding time. Detailed engineering specifications for
temperature difference was found between inflow the soil media have also become available, which
and underdrain outflow from a rain garden (Dietz provide guidance on the percentage/type of sand,
and Clausen 2005). The rapid infiltration rate of the percentage/type of compost, and percentage/type of
soils and northerly exposure of the roof (i.e., low topsoil. Soil pH, soluble salt content, and fertility may
influent temperature) were cited as the reasons for also be included in bioretention specifications. Due to
this lack of attenuation of the temperature of summer problems with clogging of the filter fabric recom-
runoff. Preliminary data in North Carolina have mended in the earlier manual, the 2002 Bioretention
shown decreases of 5°F to 10°F as influent storm- manual recommends the use of a pea gravel blanket
water passed through a bioretention cell (Hunt and around the underdrain pipe instead of filter fabric. The
Lord 2006). SCS curve number (SCS 1986) continues to be
A frequent concern for bioretention areas and recommended to estimate runoff for bioretention
infiltration practices in general, is that their perfor- sizing. The RECARGA model (Atchison et al.
mance in the winter months will be reduced, when 2006) provides detailed water budget modeling to
there may be frost in the soil. Despite measurable customize bioretention size, based on the desired
frost in the bioretention media in Connecticut “stay on” volume, or overall retention of precipitation
(unpublished data), the vast majority of inflow that is desired. The RECARGA model also uses the
(99%) was either infiltrated or evapotranspired over SCS curve number to estimate runoff from pervious
the course of a 2-year period (Dietz and Clausen surfaces, although a bioretention area could easily be
2006). A similar finding has been reported for designed with this model for a totally impervious
infiltration practices, including bioretention, at the watershed, such as a parking lot or roof, and the curve
University of New Hampshire (Roseen, personal number would not be part of the calculation.
communication). Preliminary results from Norway The Natural Resources Conservation Service rec-
also support the previous findings that bioretention ommends that the curve number approach not be used
functions well through the winter months: no for rains less than 1.3 cm (SCS 1986). Furthermore,
seasonal differences in retention time or lag time the overall accuracy of the curve number approach for
were found (Muthanna et al. 2006). Rapid thawing estimating runoff volumes has been brought into
of soil media has been found to occur when runoff question. Large discrepancies have been noted be-
enters bioretention areas. The organic material, tween runoff predicted by the curve number method
macropore structure, and porous nature of bioreten- and actual runoff for small storms (Pitt 1999).
Water Air Soil Pollut (2007) 186:351–363 355

Table 2 Summary of green

roof precipitation retention Location Precipitation Media Roof Reference
retention (%) thickness (cm) slope (%)

Augustenborg, Sweden 63.0 3.0 2.6 Bengtsson et al. 2005

Oregon, USA 69.0 12.7 – Hutchinson et al. 2003
Michigan, USA 38.6 2.0 2.0 Monterusso et al. 2004
Michigan, USA 58.1 1.0 2.0 Monterusso et al. 2004
North Carolina, USA 62.0 7.6 – Moran et al. 2004
North Carolina, USA 63.0 10.2 3.0 Moran et al. 2004
Michigan, USA 69.8 2.5 2.0 VanWoert et al. 2005
Michigan, USA 70.7 4.0 2.0 VanWoert et al. 2005
Michigan, USA 65.9 4.0 6.5 VanWoert et al. 2005
Michigan, USA 68.1 6.0 6.5 VanWoert et al. 2005
Average 62.8

WinSLAMM (Source Loading and Management roofs consisted of a thick soil layer with plants, grass,
Model) has been recommended as an alternative and/or trees, and extra structural support was required.
design tool for bioretention areas, and LID in general These “intensive” green roofs are being replaced by
(Pitt 2004). Runoff depths were well predicted by “extensive” green roofs, which have a much thinner,
WinSLAMM for a variety of watersheds, and for a lighter media (thus fewer structural requirements),
wide range of precipitation events (Pitt 1999). and different plants (Davis and McCuen 2005). A
Recently, a model called the Western Washington variety of research projects on the energy benefits of
Hydrology Model (WWHM) was constructed (http:// green roofs have been performed, however only the
www.aquaterra.com/software.html). WWHM is built research related to stormwater will be highlighted
on the Hydrologic Simulation Program Fortran here.
(HSPF) platform, but it was customized with a Retention of precipitation on a green roof is a
simpler interface, and included local soils and combination of storage in the media and evapotrans-
precipitation data for western Washington. piration by plants. Research on green roofs in a
Research on bioretention has produced positive variety of locations has consistently shown between
results, and provided insight into the mechanisms of 60% and 70% retention of precipitation, with an
pollutant retention. Despite certain problems with average retention of about 63% (Table 2). The study
phosphorus export and low TN retention, bioretention green roofs have used media with different thick-
areas have proven to significantly reduce stormflow nesses, and one researcher has specifically investigat-
volumes and concentrations of many pollutants. ed the effects of media thickness and slope on
Longer term, field based research is still necessary precipitation retention (VanWoert et al. 2005). Al-
to provide data on how these systems perform over though increased media depths and lower slopes
time, and under varying seasonal conditions. The use resulted in slightly higher (statistically significant)
of specific media and/or design variations to reduce retention, the gain in retention was not large. In
certain target pollutants is a research area that should general, for the studies examined, the thickness of the
be further explored. Although metals and nutrient media, ranging from 2 cm to over 12 cm did not result
retention in bioretention systems have been studied in in any noticeable gain in precipitation retention
detail, research on bacteria retention and water (Fig. 1). This suggests that to minimize installation
temperature attenuation are other possible research costs and structural requirements, a thinner media
areas. may be acceptable for the purposes of stormwater
retention. However, thinner media depths (5 cm) have
been found to result in winter frost injury of perennial
3 Green Roofs plants than thicker media (10 or 15 cm) in Ontario,
Canada (Boivin et al. 2001). Sedum spp. are typical
Vegetated roof systems, or green roofs, have been in plants used in green roofs, due to their drought
use in Europe for many years. Historically, green tolerance. Although Boivin et al. (2001) did not
356 Water Air Soil Pollut (2007) 186:351–363

trations over 20 mg l−1 in roof runoff were also

reported for one of the green roof systems in
Michigan. TP and NO3–N concentrations in the
precipitation were not reported for these two studies,
thus a percent retention could not be calculated.
However, in Wisconsin, TP concentrations in roof
runoff from residential and commercial areas were
0.15 and 0.20 mg l−1, respectively (Bannerman et al.
1993), which is much lower than the concentrations
reported for the runoff from the green roof studies
cited above. Preliminary results from North Carolina
are showing a similar effect: TP concentrations and
mass export in runoff from a green roof were
significantly higher than in precipitation, although
Fig. 1 Green roof media thickness (cm) vs precipitation
retention (%) no significant differences in concentration were
reported between green roof runoff and runoff from
a traditional roof (Moran et al. 2004). TN concen-
examine Sedum spp., other researchers have found trations were also significantly higher in runoff from
that Sedum spp. planted in 10-cm thick media were the green roof runoff than in precipitation. No
found to be suitable for use in Michigan, with fertilization was reported in the study in North
minimal winter damage and irrigation requirements Carolina, but the authors speculate that the export of
(Monterusso et al. 2005). TP and TN was due to leaching of these nutrients
The roofs with the lowest reported retention were from the media.
from a preliminary study (Monterusso et al. 2004), Only one study examined Cu concentrations in
and may have been the result of a low sample number runoff from green roofs. In Oregon, preliminary
(n=4), and two rainfall events that were very close to results showed that the acute water quality criteria of
each other. It should also be noted that certain green 9 μg l−1 was exceeded three times, although the
roofs had extra water storage built into the structure. concentrations were not far above the criteria.
For example, an extra water holding capacity of 2 and Although the Cu concentration in the precipitation
4 l m−2 was reported for the green roofs studied by was not reported, the authors speculate that the Cu
VanWoert et al. (2005), and Moran et al. (2004), may have come from treated wood used on the roof,
respectively. Although average retention values are or from the soil media itself.
presented in Table 1, wide-ranging retention percent- Research has shown that by using a green roof,
ages for individual storms were reported by several 60–70% reductions in stormwater runoff volume from
researchers, depending on the intensity of the storm, a roof are to be expected. This finding may have
and the time of year. In addition to overall retention of particular relevance in cities where space for storm-
flow, reductions in peak flow rate and increases in lag water treatment is costly and limited, and reductions
time were also reported for several green roofs in stormwater flows will provide great relief to
(Hutchinson et al. 2003; Moran et al. 2004; VanWoert overburdened combined sewer overflow (CSO) sys-
et al. 2005). tems. These benefits may outweigh the high initial
Very little data exist on water quality measure- cost of green roofs. This has been demonstrated in an
ments from green roofs. In Oregon, TP concentrations analysis for the city of Toronto, Canada, where after
ranging from 0.2 to 1 mg l−1 were reported in installation costs were considered, large cost savings
preliminary measurements of runoff from a green were hypothesized if green roofs were placed on all
roof (Hutchinson et al. 2003). Mean TP concentra- feasible buildings in the city, due to the reduced
tions in green roof runoff ranging from around 0.5 to stormwater load (Banting et al. 2005).
more than 4 mg l−1 were reported in Michigan, Despite the reductions in stormwater volume, the
although the authors reported that the plots were export of TP and TN need to be carefully considered
fertilized (Monterusso et al. 2004). NO3–N concen- when installing green roofs. In many cases, the large
Water Air Soil Pollut (2007) 186:351–363 357

reductions in stormwater volumes will offset any More recently, several types of permeable pave-
increase in concentration. However, plants should be ments have been monitored in the field in Renton,
selected that do not require fertilization, as this would Washington (Booth and Leavitt 1999; Brattebo and
increase the TP available for leaching. In addition, Booth 2003). Two concrete products, Turfstone®
more research needs to be performed on the effects of (turf infill) and UNI Eco-Stone® (gravel infill) were
green roof media on the export of pollutants. studied. Over the entire 6-year period of study,
Research is especially lacking on the treatment and/ negligible surface runoff was noted from both of
or export of metals from green roofs. these products, indicating that virtually all of the
precipitation infiltrated (Booth and Leavitt 1999;
Brattebo and Booth 2003). Water quality was also
improved: copper and zinc concentrations in infil-
4 Permeable Pavements trate water sampled below all four pavement types
were significantly lower (p=0.01) than in concen-
A variety of alternatives to traditional asphalt and trations in runoff from an adjacent asphalt lot
concrete paving have become available. Although (Brattebo and Booth 2003). Interestingly, the
these products share the same goal (to infiltrate authors report that from 1996 to 2002, average
stormwater), there are several design variations. The zinc concentrations in the permeable pavement
focus of this paper is research results; a thorough infiltrate and the asphalt runoff significantly in-
review of the various products and specifications is creased, yet two of the permeable systems showed
available in a recent book (Ferguson 2005) for those simultaneous decreases in copper concentrations
interested in more background information or specific (Brattebo and Booth 2003).
details. UNI Eco-Stone® was also monitored in Connect-
icut, where runoff depth from the paver surface was
4.1 Concrete Blocks or Grids 40% of precipitation depth for the 22-month study
(Gilbert and Clausen 2006). Although this is substan-
These products are precast concrete that can be in the tially more runoff than reported by Brattebo and
shape of a grid or a block with open voids to allow for Booth (2003), it was still 72% less than the runoff
infiltration. They are typically laid down by hand over depth from a nearby asphalt driveway. In addition,
a specially prepared base, although mechanical concentrations of all pollutants measured (TSS, NO3–
installation methods have become available, which N, NH 3 –N, TKN, TP, Cu, Pb and Zn) were
reduces the installation cost. The void spaces are significantly lower in runoff from the UNI Eco-
typically filled with crushed stone or pea gravel, or Stone® driveways than in runoff from the asphalt
some products can be filled with topsoil and planted driveways (Gilbert and Clausen 2006). As would be
with turf. expected, due to lower concentrations and runoff
Laboratory monitoring of concrete grid pavers volumes, mass export for all of the pollutants was also
began as early as 1981 (Day et al. 1981). Three lower for the Eco-Stone® driveways than the asphalt
products were monitored: Monoslab®, Grasscrete®, driveways (Gilbert and Clausen 2006). In another
and Turfstone®. Runoff coefficients were calculated study, Unilock® pavers were installed in one section
for a range of simulated precipitation events, up to a of a large parking lot near Toronto, Canada. Prelim-
20-year storm. The highest average runoff coeffi- inary results indicate that no surface runoff occurred
cient for the concrete grid pavers was 0.005, from the Unilock® parking lot for 9 storm events,
whereas the control concrete slab had an average with a maximum intensity of 31 mm h−1 (Toronto and
runoff coefficient of 0.78 (Day et al. 1981). Region Conservation 2006).
Percolate was also sampled for quality. High reten- Monitoring data from warmer climate zones are
tion (>81%) for all metals (Cu, Pb, Zn) was noted, showing similar results. Preliminary data from North
yet phosphorus and nitrogen fractions were incon- Carolina for three different types of concrete paver
sistently retained. In some cases, ortho-phosphate blocks indicate reductions in runoff volume, and peak
(ortho PO4–P) and NO3–N were exported from the exfiltrate flow rate, as compared to an asphalt lot
systems (Day et al. 1981). (Collins et al. 2006).
358 Water Air Soil Pollut (2007) 186:351–363

4.2 Plastic Grids remove water from the surface of a road which
reduces the risk of hydroplaning (Fitts 2002). Due to
Several types of plastic grid structures have become some structural issues, modifications have been made
available in recent years. Design and installation to the mix specs. When stormwater infiltration is
techniques may vary slightly from the concrete blocks desired, the major design difference is that the OGFC
or grids, with the largest difference being the volume material is typically put down over a coarse aggregate
of fill material in the pavement structure. In contrast storage layer that is designed to rapidly infiltrate and
to concrete blocks which are mostly impervious, the store water.
plastic grid structure is mostly pervious. The large Research on pervious asphalt began with some
spaces are designed to be filled either with topsoil and EPA funded projects in the early 1970s (Ferguson
planted with turf, or filled with a small diameter, 2005). Research in Europe began in the 1990s. In
sharp crushed stone. Installation specifications vary France, a street section was repaved with pervious
according to the manufacturer, but in general, the base asphalt, and a 61-cm thick crushed stone reservoir
preparation is critical to encourage rapid infiltration was included below the pervious asphalt layers
into the subgrade. Despite their growing popularity, (Legret and Colandini 1999). The authors reported
there are not very many monitoring studies to that on average, 96.7% of the storm water volume
document the benefits of the plastic grid structures. infiltrated in the soil below the reservoir structure
Two plastic grid structures, Grasspave® and (Legret and Colandini 1999). In Sweden, in a pervious
Gravelpave® were monitored in Renton, Washington asphalt road section with swales, between 30% and
(Booth and Leavitt 1999; Brattebo and Booth 2003). 40% of precipitation ran off the site (Stenmark 1995).
The only difference between the two installations was The swales were a confounding factor in the study in
the infill material; topsoil and turf were used in the Sweden; the individual infiltration capacity of the
Grasspave®, and gravel in the Gravelpave® structure. pervious asphalt compared to the swales was not
As with the other products monitored at this site, determined.
virtually no surface runoff was reported from either
of these two products (Brattebo and Booth 2003). 4.4 Pervious Concrete
The largest amount of runoff reported was from the
Grasspave® section, for a long-duration storm, where Pervious concrete is a variation on the typical
121 mm of rain fell, and 4 mm of surface runoff was concrete mixture. Fine sands are typically omitted
observed (Brattebo and Booth 2003). Copper and from the mix, and the slurry is tamped or rolled in
zinc concentrations in infiltrate water below all four placed, rather than the traditional floating. This type
pavement types were significantly lower (p=0.01) of concrete is much less forgiving than traditional
than asphalt runoff concentrations (Brattebo and concrete, and a proper installation requires experi-
Booth 2003). In Georgia, runoff from a Grassy enced installers. It has been installed in many
Paver™ plastic grid parking lot filled with sand and locations throughout the country, however very little
planted with grass was 93% less than runoff from an monitoring on pervious concrete installations has
adjacent asphalt lot (Dreelin et al. 2006). been performed.
A pervious concrete parking lot section with a
4.3 Pervious Asphalt swale in Florida had a runoff coefficient of 0.20,
which was lower than the coefficients for an asphalt
Pervious, or permeable asphalt, is a variation on the lot with a swale, and cement lot with a swale, which
typical hot mix asphalt (HMA) that is commonly used were 0.35 and 0.33 respectively (Rushton 2001). It
as a road surface. The mix, which omits the fine should be noted that the asphalt lot also contains a
portion of the aggregate typically included in HMA, small “garden” area, which the author felt was
was developed to be installed as a wearing course responsible for the fairly low runoff coefficient from
over a standard asphalt layer. The mix was termed the asphalt lot (Rushton 2001). Pollutant export load
open graded friction course (OGFC), and it has been from the pervious lot with a grass swale was reduced
used around the country since the 1970s because of its for TSS, NO3–N, NH3–N, and TN by 91%, 66%,
ability to dampen road noise and tire spray, and 85%, and 42%, respectively, as compared to the
Water Air Soil Pollut (2007) 186:351–363 359

asphalt lot with no swale (Rushton 2001). Metal load in France (Legret and Colandini 1999). The mainte-
reductions (Cu, Fe, Pb, Mn, Zn) were all greater than nance recommendation for UNI-Ecostone® pavers
75%. However, TP loads were only reduced by 3%, is the removal and replacement of the infill material.
despite the large decrease in runoff volume, and some The time interval for the replacement depends upon
of the systems with grassed swales exported more TP the local conditions, and the loading of fine particles
than came in (Rushton 2001). This phenomenon is on to the pavement surface. Research on pervious
consistent with the TP export noted earlier from some pavement sites in North Carolina, Maryland, Virginia,
bioretention systems (Dietz and Clausen 2006; Hunt and Delaware (Bean et al. 2007) has shown that
et al. 2006; Toronto and Region Conservation 2006). although the infiltration capacity of concrete grid,
A large pervious concrete plaza was installed at concrete block, and pervious concrete pavements may
Villanova University. Although there were some decrease if fine particles are loaded on to the surface,
problems with the installation of the material and they can still infiltrate large quantities of water
some sections had to be reinstalled (Traver et al. (comparable to grassed sandy loam), and the infiltra-
2005), the problems have been corrected, and the site tion rate can be improved with replacement of the
has shown promising results. The site takes runoff infill material. In place of sand, the authors recom-
from adjacent standard concrete areas, several roof- mend the use of crushed aggregate as an infill
tops, and grassed areas. To date, the site has material to help encourage high infiltration rates
successfully captured and infiltrated runoff from all (Bean et al. 2007). A laboratory experiment on
storms 5 cm or less in size (Kwiatkowski et al. 2007). pervious concrete found that even when the surface
Water quality measurements were also taken at the of the material was clogged with fine sand, the vast
site; chloride concentrations were found to be highest majority of simulated rainfall (up to a 100-year
during winter months, as deicers were applied to event for the Columbia, SC region) was infiltrated
pedestrian areas. In addition, concentrations of copper (Haselbach et al. 2006).
in roof runoff were fairly high (Kwiatkowski et al.
2007). However, neither copper nor chloride concen- 4.5.2 Winter Performance
trations in groundwater below the pervious concrete
were high enough to be of concern. The authors Another concern with pervious pavements, just as
concluded that with proper siting, an infiltration BMP with bioretention, is the ability of the system to
such as the pervious concrete would not adversely perform in the winter. Numerous studies on pervious
impact the groundwater (Kwiatkowski et al. 2007). pavements in cold climates (e.g., Connecticut, Wash-
ington, New Hampshire, and Ontario Canada) have
4.5 Other Concerns been performed or are ongoing. Research findings
support the claims of manufacturers that with a proper
4.5.1 Clogging of Surfaces base and proper installation, the system will continue
to infiltrate through the winter, and the surface can be
A frequent concern with porous pavements is the plowed, although some care should be exercised with
clogging of the surface over time. Rather than sanding (to avoid clogging of the pores) and salting
particles becoming lodged in the internal structure of (to avoid potential groundwater contamination).
the pavement, clogging of pervious asphalt pavements
seems to be confined to the surface 2 cm of the 4.5.3 Soils
pavement (Baladès et al. 1995). The specifications for
these types of products (e.g., pervious asphalt, In addition to concerns about winter performance,
pervious concrete) state that the pavement surface fine grained soils with slow infiltration rates have
should be cleaned out with vacuum suction on a been cited as a reason why a pervious pavement or
specified maintenance interval, so that the infiltration bioretention cannot be used. However, research has
rate can be maintained. A more intensive vacuuming, shown that with appropriate design, pervious pave-
high pressure washing, and suction removal of the ments can be used in clay soils. A previously cited
remaining sludge was found to greatly improve the example in Georgia (Dreelin et al. 2006) was installed
infiltration rate of a partially clogged pervious asphalt over well-drained soils with clayey subgrade that
360 Water Air Soil Pollut (2007) 186:351–363

could contain as much as 35–60% clay. An under- It should be noted that LID advocates a distributed
drain system was installed in the subgrade, below a approach to treatment practices, rather than an “end of
10-in. thick layer of open graded gravel. Runoff from pipe” approach. If this strategy is adhered to, the
the underdrain was only observed one time, during a stormwater will have less of a chance to accumulate
1.85 cm precipitation event (Dreelin et al. 2006). Just large masses of pollutants. Therefore, the likelihood
as with bioretention (Winogradoff 2002), in areas of having high concentrations of pollutants will be
where native soils may not have high infiltration rates, reduced if the distributed approach is used, and
a thicker reservoir of coarse aggregate can be installed concentrations of pollutants will largely be driven by
beneath the pavement structure and underdrain. This atmospheric deposition rates. Collecting and treating
provides a greater storage capacity, and a longer time stormwater from high traffic areas or areas with high
for water to exfiltrate to the native soils before potential pollutant loads, while infiltrating “cleaner”
underdrain flow would begin. runoff from buildings and low traffic areas, may
provide a good margin of safety where groundwater
4.5.4 Groundwater Contamination contamination is a concern.

Due to the fact that stormwater runoff is known to

contain a wide variety of pollutants (Makepeace 5 Large-scale LID Studies
et al. 1995), concerns of groundwater contamination
have been raised where infiltration practices such as Many large-scale housing projects utilizing LID
pervious paving or bioretention have been recom- techniques have been installed around the country.
mended. The results from a multiyear research However, to date only one such project has had
project sponsored by the US EPA on this topic have monitoring of stormwater quantity and quality to
been summarized (Pitt et al. 1999). For residential investigate the cumulative impact of LID systems.
and light commercial applications, the pollutants of The Jordan Cove Urban Watershed Project in Water-
concern are typically nutrients, petroleum residue ford, Connecticut was designed to monitor the effects
from automobile traffic, heavy metals, pathogens, of a traditional (17 lots) and LID (12 lots) subdivi-
and possibly pesticides. Due to the fact that these sion. Grassed swales, bioretention areas, pervious
pollutants are usually found in fairly low concen- pavements and a cluster layout were among the
trations in stormwater, and are well retained by soils, practices utilized in the LID subdivision. Large
the contamination potential is low or moderate (Pitt increases in runoff and pollutant export were found
et al. 1999). Two exceptions to this general finding as the traditional subdivision was developed, however
exist: pathogens may be present in high concen- no change in runoff depth or nitrogen or phosphorus
trations, and may not be well attenuated in the soil. export was found as the LID subdivision was
Fecal coliform bacteria was well retained by bio- developed (Dietz and Clausen 2007). Stormwater
retention columns in the previously cited preliminary runoff lag times were also significantly greater in
laboratory study (Rusciano and Obropta 2005), the LID subdivision, indicating that the LID practices
however field research on bacteria and virus removal were effective in increasing the time of concentration
in bioretention or pervious paving systems is lack- in the watershed, and maintaining the hydrologic
ing. Also, chloride may be present in stormwater, function of the undeveloped site (Hood et al. 2007).
and concentrations may be high during winter
months (Pitt et al. 1999). Chloride is also very
mobile in soil, and can easily travel to shallow 6 Conclusions
groundwater. Research is showing that concentra-
tions of chloride have been increasing in local LID is a relatively new suite of practices and is
waterways in New England (Kaushal et al. 2005), constantly evolving. To date, the research on LID
and if current trends continue, chloride levels in practices has not been as extensive as the research on
streams will reach dangerous levels, threatening agricultural or traditional urban stormwater best
aquatic life. The ability of LID systems to treat management practices. The LID practices investigat-
bacteria and chloride needs to be investigated further. ed in this review have shown great promise in
Water Air Soil Pollut (2007) 186:351–363 361

mitigating the impacts of development on down- Although the individual practices often have
stream water bodies. However, as this paper shows, detailed specifications, it seems that engineers to do
some strong conclusions can be drawn from the not have a consistent design tool that can credit the
research to date when LID practices are used in runoff reductions that LID components can provide,
developed areas. The research cited in this paper has that is also based on research results. The standard
shown generally that LID practices are effective at curve number calculation (SCS 1986), while fairly
preserving the natural hydrologic function of a site, easy to apply, does not have the flexibility to give
and retaining pollutants. Also, some frequently credit for the variety of LID components available,
voiced concerns about the function of pervious and its accuracy has been brought into question.
pavements and bioretention areas have been shown Engineers are using models like RECARGA, Win-
to be inaccurate: pervious pavements and bioreten- SLAMM, and P8 to design LID practices, although
tion have been found to work effectively in cold they may use another model such as SWMM for
climates, with frost in the ground. Proper base design hydraulic routing on a site. The Western Washington
and installation are critical to this function. In Hydrologic Model is accurate, easy to use, and
addition, substantial infiltration in tight soils beneath provides credits for LID practices. The widespread
pervious pavers has also been found. Again, proper adoption of an accurate model to give proper credit to
design and installation are critical components of LID components is critical for widespread adoption of
LID systems in any application, not just cold LID techniques.
climates or tight soils. Future research needs have been identified. Longer
There are certain conditions where it may not be term studies for all of the practices are justified, as
appropriate to use an LID practice that relies on very few studies exist on how these systems perform
infiltration. Areas with high contaminant loading for long periods of time. In addition, investigations on
such as recycling centers or gas stations, or the effect of different media mixtures for bioretention
brownfield areas with high soil contamination, may and green roofs to minimize the risk of phosphorus
not be appropriate for infiltration, due to increased export are needed. Also, further research on the
risks of contaminating the groundwater. Conditions ability of LID systems to retain and destroy bacteria
such as steep slopes, shallow (<3 ft) depth to and viruses is needed. Despite limitations in certain
bedrock or seasonal high water table are also places situations, it seems clear that LID is a viable storm-
where traditional pavement and stormwater manage- water treatment option that has broad applicability.
ment practices may be more appropriate. However,
rarely is an entire site composed of such limiting
conditions, and LID practices can be used wherever
possible to reduce to cumulative impact on down- References
stream water bodies.
Atchison, D., Potter, K. W., & Severson, L. (2006). Design
A common thread across green roofs, bioreten-
Guidelines for Stormwater Bioretention Facilities. Uni-
tion, and grass swales has been noted: the export of versity of Wisconsin Water Resources Institute, WIS-
phosphorus. This issue appears to be linked to high WRI-06-01.
phosphorus levels in the media (for bioretention and Baladès, J.-D., Legret, M., & Madiec, H. (1995). Permeable
pavements: Pollution management tools. Water Science
green roofs), or possibly to fertilization of turf or
and Technology, 32(1), 49–56.
planted areas. This can be a concern in areas where Bannerman, R. T., Owens, D. W., Dodds, R. B., & Hornewer,
underdrains or roof leaders are tied into a storm- N. J. (1993). Sources of pollutants in Wisconsin storm-
water system; in these cases the excess loading of water. Water Science and Technology, 28(3–5), 241–259.
Banting, D., Doshi, H., Li, J., Missios, P., Au, A., Currie, B. A.,
TP to surface waters may worsen an existing
& Verrati, M. (2005). Report on the environmental
problem. Care should be taken to ensure that in benefits and costs of green roof technology for the city
cases where a drain is directly connected and is of Toronto. City of Toronto and Ontario Centres of
likely to be utilized, the media does not contain high Excellence–Earth and Environmental Technologies.
Bean, E. Z., Hunt, W. F., & Bidelspach, D. A. (2007). Field
levels of phosphorus. Proper education of mainte-
survey of permeable pavement surface infiltration rates.
nance personnel and homeowners can also address Journal of Irrigation and Drainage Engineering, 133(3),
the issue of excessive fertilization. 249–255.
362 Water Air Soil Pollut (2007) 186:351–363

Bengtsson, L., Grahn, L., & Olsson, J. (2005). Hydrological Gilbert, J. K., & Clausen, J. C. (2006). Stormwater runoff quality
function of a thin extensive green roof in southern and quantity from asphalt, paver, and crushed stone drive-
Sweden. Nordic Hydrology, 36(3), 259–268. ways in Connecticut. Water Research, 40, 826–832.
Boivin, M., Lamy, M., Gosselin, A., & Dansereau, B. (2001). Haselbach, L. M., Valavala, S., & Montes, F. (2006).
Effect of artificial substrate depth on freezing injury of six Permeability predictions for sand-clogged Portland cement
herbaceous perennials grown in a green roof system. pervious concrete pavement systems. Journal of Environ-
Horticulture Technology, 11(3), 409–412. mental Management, 81, 42–49.
Booth, D. B., & Jackson, R. (1997). Urbanization of aquatic Hollis, G. E. (1977). Water yield changes after the urbanization
systems: Degradation thresholds, stormwater detection and of the Canon’s Brook catchment, Harlow, England.
the limits of mitigation. Journal of the American Water Hydrological Sciences Bulletin, 22, 61–75.
Resources Association, 33(5), 1077–1089. Hood, M., Clausen, J. C., & Warner, G. S. (2007). Comparison
Booth, D. B., & Leavitt, J. (1999). Field evaluation of of stormwater lag times for low impact and traditional
permeable pavement systems for improved stormwater residential development. Journal of the American Water
management. Journal of the American Planning Associa- Resources Association, 43(4), 1036–1046.
tion, 65(3), 314–325. Hunt, W. F., Jarrett, A. R., Smith, J. T., & Sharkey, L. J. (2006).
Brattebo, B. O., & Booth, D. B. (2003). Long-term stormwater Evaluating bioretention hydrology and nutrient removal at
quantity and quality performance of permeable pavement three field sites in North Carolina. Journal of Irrigation
systems. Water Research, 37, 4369–4376. and Drainage Engineering, 132(6), 600–608.
Collins, K. A., Hunt, W. F., & Hathaway, J. M. (2006, Hunt, W. F., & Lord, W. G. (2006). Bioretention performance,
November). Evaluation of various types of permeable design, construction, and maintenance. North Carolina
pavements with respect to water quality improvement and State University Cooperative Extension.
flood control. Paper presented at the 8th International Hutchinson, D., Abrams, P., Retzlaff, R., & Liptan, T. (2003),
Conference on Concrete Block Paving, San Francisco, Stormwater monitoring two ecoroofs in Portland, Oregon,
CA. USA. Paper presented at the First Annual Greening
Davis, A. P., & McCuen, R. H. (2005). Stormwater manage- Rooftops for Sustainable Communities Conference,
ment for smart growth. New York, NY: Springer. Awards and Trade Show, Chicago, IL.
Davis, A. P., Shokouhian, M., Sharma, H., & Minami, C. Jennings, D. B., & Jarnagin, S. T. (2002). Changes in
(2001). Laboratory study of biological retention for urban anthropogenic impervious surfaces, precipitation and daily
stormwater management. Water Environment Research, 73 streamflow discharge: A historical perspective in a mid-
(1), 5–14. Atlantic subwatershed. Landscape Ecology, 17, 471–489.
Davis, A. P., Shokouhian, M., Sharma, H., Minami, C., & Kaushal, S. S., Groffman, P. M., Likens, G. E., Belt, K. T.,
Winogradoff, D. (2003). Water quality improvement Stack, W. P., Kelly, V. R., et al. (2005). Increased
through bioretention: Lead, copper and zinc removal. salinization of fresh water in the northeastern United
Water Environment Research, 75(1), 73–82. States. Proceedings of the National Academy of Sciences,
Day, G. E., Smith, D. R., & Bowers, J. (1981). Runoff and 102, 13517–13520.
pollution abatement characteristics of concrete grid pave- Kim, H., Seagren, E. A., & Davis, A. P. (2003). Engineered
ments. Virginia Water Resources Research Center, Virginia bioretention for removal of nitrate from stormwater runoff.
Polytechnic Institute and State University, Project A-090- Water Environment Research, 75(4), 355–367.
VA VPI-VWRRC-BULL 135 4C. Korom, S. F. (1992). Natural denitrification in the saturated zone:
Dietz, M. E., & Clausen, J. C. (2005). A field evaluation of rain A review. Water Resources Research, 28(6), 1657–1668.
garden flow and pollutant treatment. Water, Air and Soil Kuichling, E. (1889). The relation between the rainfall and the
Pollution, 167(1–4), 123–138. discharge of sewers in populous districts. Transactions of
Dietz, M. E., & Clausen, J. C. (2006). Saturation to improve the American Society of Civil Engineers, 20, 1–60.
pollutant retention in a rain garden. Environmental Science Kwiatkowski, M., Welker, A. L., Traver, R. G., Vanacore, M.,
& Technology, 40(4), 1335–1340. & Ladd, T. (2007). Evaluation of an infiltration best
Dietz, M. E., & Clausen, J. C. (2007), Stormwater runoff and management practice (BMP) utilizing pervious concrete.
export changes with development in a traditional and low Journal of the American Water Resources Association, (in
impact subdivision. Journal of Environmental Manage- press).
ment, (in press). DOI 10.1016/j.jenvman.2007.03.026. LeBlanc, R. T., Brown, R. D., & FitzGibbon, J. E. (1997).
Dietz, M. E., & Filchak, K. F. (2006). Rain gardens: A design Modeling the effects of land use change on the water
guide for homeowners in Connecticut. University of temperature in unregulated urban streams. Journal of
Connecticut Cooperative Extension System. Environmental Management, 49, 445–469.
Dreelin, E. A., Fowler, L., & Carroll, C. R. (2006). A test of Legret, M., & Colandini, V. (1999). Effects of a porous
porous pavement effectiveness on clay soils during natural pavement with reservoir structure on runoff water: Water
storm events. Water Research, 40, 799–805. quality and fate of heavy metals. Water Science and
Ferguson, B. K. (2005). Porous pavements. Boca Raton, FL: Technology, 39(2), 111–117.
CRC Press, Taylor & Francis Group. Leopold, L. B. (1968). Hydrology for urban land planning—a
Fitts, G. (2002). The new and improved open graded friction guidebook on the hydrologic effects of urban land use.
course mixes. Asphalt, 2002, 16–18, Fall. Geological Survey Circular, 554.
Water Air Soil Pollut (2007) 186:351–363 363

Makepeace, D. K., Smith, D. W., & Stanley, S. J. (1995). Urban stormwater. Paper presented at the North American
stormwater quality: Summary of contaminant data. Critical Surface Water Quality Conference and Exposition,
Reviews in Environmental Science and Technology, 25(2), Orlando, FL, July 18–25.
93–139. Rushton, B. (2001). Low-impact parking lot design reduces
Monterusso, M. A., Rowe, D. B., & Rugh, C. L. (2005). runoff and pollutant loads. Journal of Water Resources
Establishment and persistence of Sedum spp. and native Planning and Management, 127(3), 172–179.
taxa for green roof applications. Horticultural Science, 40 SCS (1986). Urban hydrology for small watersheds. U.S.
(2), 391–396. Department of Agriculture Technical Release No. 55
Monterusso, M. A., Rowe, D. B., Russell, D. K., & Rugh, C. L. (revised), 210-VI-TR-55.
(2004). Runoff water quantity and quality from green roof Stenmark, C. (1995). An alternative road construction for
systems. Acta Horticulturae, 639, 369–376. stormwater management in cold climates. Water Science
Moran, A., Hunt, W., & Jennings, G. (2004). Greenroof and Technology, 32(1), 79–84.
research of stormwater runoff quantity and quality in Toronto and Region Conservation. (2006). Performance eval-
North Carolina. NC State University, A&T State Univer- uation of permeable pavement and a bioretention swale.
sity, Cooperative Extension, ISSN 1062-9149. Seneca College, King City, Ontario. Toronto and Region
Muthanna, T. M., Thorolfsson, S. T., & Viklander, M. (2006, Conservation Authority, Interim Report #2.
Winter hydrology in a cold climate rain garden. Paper Traver, R. G., Welker, A. L., Horst, M., Vanacore, M., Braga,
presented at the 2006 American Society of Agricultural A., & Kob, L. (2005). Lessons in porous concrete.
and Biological Engineers Annual International Meeting, Stormwater, July/August, 30–45.
paper No. 062309, Portland, OR. US EPA (1983). Results of the nationwide urban runoff
Pitt, R. (1999). Small storm hydrology and why it is important program. United States Environmental Protection Agency,
for the design of stormwater control practices. In W. James NTIS PB84-185552.
(Ed.), Advances in modeling the management of storm- US EPA (2000). Low impact development (LID), a literature
water impacts, volume 7. Guelph, Ontario: Computational review. United States Environmental Protection Agency,
Hydraulics International and Lewis Publishers/CRC Press. EPA-841-B-00-005.
Pitt, R. (2004, September). WinSLAMM and low impact US EPA (2002). National water quality inventory, 2000 report.
development. Paper presented at the Putting the LID on United States Environmental Protection Agency, EPA-
Stormwater Management, College Park, MD. 841-R-02-001.
Pitt, R., Clark, S., & Field, R. (1999). Groundwater contami- VanWoert, N. D., Rowe, D. B., Andresen, J. A., Rugh, C.
nation potential from stormwater infiltration practices. L., Fernandez, R. T., & Xiao, L. (2005). Green roof
Urban Water, 1, 217–236. stormwater retention: Effects of roof surface, slope, and
Prince George’s County. (1993). Design manual for use of media depth. Journal of Environmental Quality, 34,
bioretention in stormwater management. Prince George’s 1036–1044.
County, MD Department of Environmental Resources, Waananen, A.O. (1969). Effects of watershed changes on
Watershed Protection Branch, MD Department of Envi- streamflow. In W. L. Moore, & C. W. Morgan (Eds.),
ronmental Protection, Landover, MD. Urban effects on water yield (pp. 169–182). Austin and
Prince George’s County. (1999). Low-impact development design London: University of Texas Press.
strategies: An integrated design approach. Prince George’s Winogradoff, D. (2002). The bioretention manual. Prince
County, MD Department of Environmental Resources. George’s County, MD Department of Environmental
Roseen, R. M., Ballestero, T. P., Houle, J. J., Avelleneda, P., Resources, Watershed Protection Branch, MD Department
Wildey, R., & Briggs, J. (2006). Storm water low-impact of Environmental Protection, Landover, MD.
development, conventional structural, and manufactured WI DNR. (2003). Rain gardens: A how-to manual for home-
treatment strategies for parking lot runoff. Transportation owners. Wisconsin Department of Natural Resources,
Research Record: Journal of the Transportation Research DNR Publication PUB-WT-776 2003.
Board, 1984, 135–147. WI DNR. (2006). Bioretention for infiltration. Conservation
Rusciano, G. M., & Obropta, C. C. (2005). Efficiency of Practice Standard (1004). Wisconsin Department of
bioretention systems to reduce fecal coliform counts in Natural Resources.