Portland’s Ecoroof Avian Monitoring Project
2012-14
Final report
Western Scrub-Jay on the Hamilton West ecoroof in Fall
2015
Casey Cunningham, Bureau of Environmental Services
Joe Liebezeit, Audubon Society of Portland
Portland’s Ecoroof Avian Monitoring Report
Table of Contents
Background ..................................................................................... 2
Methodology.................................................................................... 2
Site Descriptions ............................................................................. 4
Results and Discussion ................................................................... 8
Conclusions………………………………………………………..……14
Preliminary Management Recommendations……………………… 15
Considerations for Future Studies ................................................. 15
Acknowledgements ...................................................................... .16
References ………………………………………………………….…..16
Background
Ecoroofs are an alternative to conventional roofing practices. Ecoroofs provide
multiple benefits beyond those gained from a conventional rooftop. They manage
stormwater, save money by extending the life of the waterproof roofing
membrane, cool and clean the air, save energy, and provide habitat. Portland is
taking a closer look at the habitat benefits of ecoroofs for macroinvertebrates and
birds. Research in Switzerland and England, where there’s a longer history with
greenroofs, has shown that they can function as valuable habitat for rare or
threatened birds, plants, insects and other wildlife, and can be designed to
increase that value (Baumann, 2006; Brenneisen, 2006; Gedge, 2005). Little
research in this area has been performed in North America although this is
changing (see Eakin, 2012).
This study compares Portland ecoroofs with conventional roofs and ground-level
greenspaces to find patterns of presence and use by birds. Birds can act as an
indicator of the broader biodiversity value of a site, and are conspicuous and
cost-effective to monitor. Data was collected during peaks of spring and fall
migration periods because this is a critical time in many species life-cycles, and
the time when birds are most abundant in our area. Sampling began in spring
2012 and continued through fall 2014, for a total of six seasons. Future ecoroof
designs for Portland which maximize habitat value may be monitored and
compared to the baseline data collected in this study.
Methodology
Three study areas with three treatments each (nine sites) were monitored per
season, all in highly urbanized areas near or in downtown Portland, Oregon.
Treatments included 1) three greenroofs (two extensive ecoroofs, one intensive
roof garden); 2) three conventional roofs to serve as controls, and 3) three
ground-level landscaped areas (grassy parks or parking lots with tree cover and
2
shrubs). Within each study area (i.e. cluster of 3 treatment sites) the sample
areas for each treatment were approximately the same size, and the
conventional roofs and ground sites were chosen to be as close to the greenroof
as possible. Ecoroofs and control roofs were selected to be of similar heights, as
best as could be arranged. All roofs were flat and on commercial buildings.
Study areas are unique in their proximities to the Willamette River and natural
areas or areas with higher tree canopy.
Monitoring occurred from 7-9 AM or 8-10 AM, on nine mornings in spring during
April and May, and 12 mornings in fall between August and October. Each
individual site was sampled three times in spring for a total of six hours of
monitoring per season; and four times in fall for a total of eight hours per season.
On each date, avian monitoring occurred simultaneously at one ecoroof, one
nearby conventional roof, and one nearby ground-level landscaped area. BES
and Audubon staff modified an existing bird survey protocol developed by Huff et
al (2000) to suit the logistics and constraints of roof access. Birds heard and
seen on or flying directly above the sites were recorded, along with observed
behaviors. Monitors ranked their identification confidence for each bird as high,
medium or low. Fifteen trained Audubon volunteers, two Audubon staff and one
City staff conducted the monitoring shifts. Each site had one or two monitors who
sat still, recording bird activity from the location with the most expansive view of
the study area.
Only birds that landed within the sites are included in the data analysis. However,
birds in flight that were actually using habitat in a given treatment were included
(e.g. a hummingbird hovering and collecting nectar from a flower). For the
species richness estimate, only high-confidence identifications were included.
Low-confidence or unidentified birds were included in the avian abundance data.
This method generally followed recommendations described in FernandezCanero and Gonzalez-Rodondo (2010).
NCSS 8 (Hintze 2012) was used to test for significant differences in bird
abundance among years and treatments using a 2-way ANOVA. For the
comparison among years, only data for treatments with data for all three years of
the study per site were included. No abundance differences were detected
among year for each treatment. Years were then pooled together and a one-way
ANOVA was used to test for significant differences among treatments using all
abundance data. Year and site were treated as fixed effects. Residual distribution
tests and the modified Levine equal variance test (Zar 1999) were used to check
for violations in ANOVA test assumptions.
3
Site Descriptions
Downtown Portland showing all monitoring sites
Some sites changed during the course of the study due to changes in building
management and accessibility, as noted below.
Central Wine Warehouse Sites
An area of approximately 20,000 square feet was monitored for each of the
following sites:
 Central Wine Warehouse ecoroof - monitored 2012-2014:
Ecoroof Constructed: 2008
Number of Stories: 2
Distance from Willamette River: 3 blocks
Design: Mix of all native sedum, grasses and forbs in extensive greenroof
soil blend with red cinder drainage channels. Some volunteer non-native
plants are present. Soil depth averages about 5”. Not irrigated or
accessible.
4



Tazo control roof - monitored 2012:
Number of Stories: 2
Distance from Willamette River: 1 block, with the interstate and railroad
running between.
Roofing Type: Asphalt membrane with a light-colored granular coating.
American Medical Response control roof - monitored 2013-2014:
Number of Stories: 2
Distance from Willamette River: 1 block, with the interstate and railroad
running between and the Burnside Bridge adjacent to the north.
Roofing Type: Asphalt membrane with a light-colored granular coating.
Tazo parking lot - monitored 2012-2014:
Distance from Willamette River: 2 blocks, with the interstate and railroad
running between.
Vegetation: Site is 65% paved with narrow landscape strips with mediumsize non-native trees providing about 40% canopy.
Central Wine ecoroof in Spring
AMR control roof
Tazo ground site
Hamilton West ecoroof
5
West Apartments Sites
An area of approximately 4,000 square feet was monitored for each of the
following:
 Hamilton West Apartments ecoroof - monitored 2012-2014:
Ecoroof Constructed: 1999
Number of Stories: 10
Distance from Willamette River: 14 blocks
Design: Dominated by non-native sedum and volunteer grasses, with
some forbs. Includes a geotextile drainage membrane beneath two
different soil blends between 3 and 5” deep. An adjacent patio for tenants
is separated from the ecoroof by a fence. Irrigation status is unknown.
 12th Avenue Terrace Apartments control roof - monitored 2012-2014:
Number of Stories: 6
Distance from Willamette River: 14 blocks
Roofing Type: Covered in 1/2” depth pea gravel.
 Portland State University (PSU) park block (SW 12th Ave. and Market
St.) - monitored 2012-2014:
Distance from Willamette River: 13 blocks
Vegetation: Open lawn with large non-native street trees around two
sides providing about 50% canopy.
12th Avenue Terrace control roof
Park block at PSU
Louisa roof garden
Crystal Ballroom control roof
6
Louisa Apartments Sites
An area of approximately 8,700 square feet was monitored for each of the
following sites:
 Louisa Apartments roof garden – monitored 2012-2013:
Ecoroof Constructed: 2005
Number of Stories: 2
Distance from Willamette River: 14 blocks
Design: Soil depth undulates between 6 and 18”. Planted with various
non-native ornamental grasses, shrubs and small trees in pots. Paths and
seating for tenants are integrated throughout the roof garden. Irrigated.
 Crystal Ballroom control roof – monitored 2012-2013:
Number of Stories: 4
Distance from Willamette River: 15 blocks.
Roofing Type: Asphalt membrane with a light-colored granular coating.
 North Park Block (NW Couch St. and Park Ave.) - monitored 20122013:
Distance from Willamette River: 9 blocks.
Vegetation: Lawn with large non-native street trees. 80% tree canopy
(deciduous).
North Park Blocks
Gray’s Landing ecoroof
Southwest Charter School control roof
Caruthers Park
7
Gray’s Landing Apartments Sites
An area of approximately 14,700 square feet was monitored for each of the
following sites:
 Gray’s Landing Apartments ecoroof – monitored 2014:
Ecoroof Constructed: 2012
Number of Stories: 2
Distance from Willamette River: 2 blocks
Design: Soil depth 4”. Planted with various non-native sedum species.
Irrigated. Not accessible.
 Southwest Charter School control roof – monitored 2014:
Number of Stories: 4
Distance from Willamette River: 2 blocks.
Roofing Type: EPDM (black).
 Caruthers Park (SW Moody Ave. and Gaines St.) – monitored 2014:
Distance from Willamette River: 1 block.
Vegetation: Native and ornamental shrubs and trees with undulating
topography. Includes bioswales, paths and benches. Large lawn adjacent
to sampling area.
Results and Discussion
This data should be considered preliminary and ideally, it would have included
more sampling days and measured additional key variables to help compensate
for potential confounding factors inherent to this study. Each site has a unique
context, aspect, age, planting plan, soil mix, distance to natural areas, and
elevation, among other variables. Despite these constraints, broad patterns in
avian richness and abundance were assessed.
Because of tree canopy and other visual barriers, birds in ground-level
greenspaces were more difficult to detect than at the roof top sites. As a result,
ground-level greenspaces likely had a lower detection probability and number of
birds compared to the rooftops was probably underestimated.
Cumulatively, for both abundance and species richness, greater numbers were
found landing on ground-level greenspaces than on greenroofs, and greater
numbers on greenroofs than on conventional roofs. However, a significant
difference in bird abundance between years or treatments was not detected
(F=0.61, df=2, P=0.58; F=1.81, df=13, P=0.21) likely because of high withintreatment variation in abundance estimates (Fig. 2). Nevertheless, almost twice
the number of individual birds and species landed on greenroofs than on
conventional roofs. This suggests that greenroofs can serve as extensions of
urban habitats for aerial species. The lack of ground-level predators and lower
number of human disturbances on most roofs could benefit species using
greenroofs as well (Fernandez-Canero and Gonzalez-Rodondo 2010), although
there may be other dangers, such as increased bird strike potential because of
exposure to glass and potentially increased numbers of predatory bird species.
8
Figure 1. Total avian abundance and richness at three treatments collected in Portland, OR, 2012-14
Results were highly variable from day to day and from season to season (Fig. 2).
Abundance on the Louisa greenroof ranged from zero landing birds in a two-hour
monitoring period to 27 in another. The Central Wine Warehouse ecoroof varied
between five and 45 landing birds; and five to 53 birds landed on the Hamilton
ecoroof on different sampling dates. This variability highlights the need for a
larger data set to more confidently determine patterns in bird abundance and
richness across sites and treatments. During both seasons the Louisa roof
garden had noticeably fewer detections and less activity than other greenroofs,
which may be due to its accessibility to tenants and proximity to a busy street.
Despite the lower detections, it had comparable diversity to the other ecoroofs
and was the only roof where breeding was observed during the study.
The Central Wine ecoroof and its corresponding conventional roof and groundlevel greenspace were the largest sites sampled, and cumulatively the ecoroof
showed slightly more bird activity than the other ecoroofs. This was not the case
for its corresponding conventional roof, suggesting the larger the greenroof, the
higher the bird activity, while larger conventional roofs do not appear to support
increased bird activity. Also, the Central Wine ecoroof is the only study site with
predominantly native plants, which may contribute to its higher observed avian
activity. The Tazo parking lot had much lower abundance than the other ground
sites, perhaps related to the mostly impervious hardscape beneath the tree
canopy.
9
Abundance by Land Type - Spring 2013
60
Ecoroof
Control Roof
Ground
50
40
Abundance
30
Mean
20
10
44
4- -13
5
4 -1
4- - 6-13
2
4- 5-13
2
4- 6-1 3
27 3
5- -1 3
2
5- -13
3
5- -13
413
44
4- -13
5
4 -1
4- - 6-13
2
4- 6-13
27 3
5- -1 3
2
5- -13
3
5- -13
413
44
4- -13
5
4 -1
4- - 6-13
2
4- 5-13
2
4- 6-1 3
27 3
5- -1 3
2
5- -13
3
5- -13
413
0
Figure 2. Abundance by land type at all sites combined during the spring 2013 season
The Gray’s Landing sites are closest to the river and a large vegetated area
(Ross Island), yet increased abundance or species richness was not detected at
these sites. However they were only monitored during the final season so the
results reflect a small sample size.
Table 1. Number of detections of observed foraging and
breeding behaviors of birds using each treatment.
Foraging was the most frequently observed behavior at all sites. Greenroofs
provide sites to forage unlike on conventional roofs where foraging sites are
fewer (Table 1). On conventional roofs, foraging was relegated to areas of pea
gravel ballast or in cracks beneath objects on the roof. Foraging on the ground
sites consisted largely of bird flocks utilizing the tree canopy, with some ground
and understory use. Generalist omnivores were most frequently observed on
both roof types (Fig. 3) while in the ground-level greenspaces insectivorous
10
species were most abundant, gleaning in tree canopies. Birds observed on the
greenroofs were primarily granivores (pigeons, finch, sparrows) and omnivores
(crows, jays, starlings) observed feeding on the soil surface among the low
plants. Swallows, an aerial insectivore, were seen feeding over all sites.
Hummingbirds, a nectarivore, were seen perched or passing over all site-types.
Species richness was equal or greater at ground sites across all foraging guilds
(Fig. 4) compared to greenroof and conventional roofs. It should be noted that
these dietary classifications include bird species that focus on a particular food
type during the migration season. Many of the species that were classified
separately do overlap in diet to some degree, eating a combination of plant and
animal matter.
Figure 3. Number of individuals observed by foraging guild at three treatments, 2012-14
Multiple species used parapets and other objects on both roof types for perching,
sometimes in flocks. This was the primary avian use of conventional roofs. More
singing, visual displays and nest material collecting were recorded on the
greenroofs than on the ground sites (Table 1). This could be a result of better
visibility in the more open sites, and an indication that greenroofs provide suitable
habitat.
11
Figure 4. Number of species observed by foraging guild at three treatments, 2012-14. Only species
observed more than five times on a site-type were included
Figure 5. Total detections of native vs. non-native birds in
each treatment, 2012-14.
Almost half of the birds that landed on ecoroofs were non-native but at a slightly
lower ratio than those landing on conventional roofs, while ground sites had
considerably fewer non-native species (Fig. 5). Three species made up all nonnative species recorded: Rock Pigeon, European Starling and House Sparrow.
12
Figure 6. Number of individuals by species detected on the three treatments, 2012-14. Only species
observed more than five times on a particular site-type were included.
A few neotropical migrants were observed. The most common were Violet-Green
Swallows and Yellow-Rumped Warblers, which were recorded on or near all sites
at some point during the study. The ground sites had a greater diversity of
migrants, including Wilson’s, Yellow, and Nashville Warblers, House Wren,
Western Wood-Pewee and Western Tanager. Savannah Sparrows were
observed on the Central Wine Ecoroof, a migratory grassland species, a habitat
which ecoroofs superficially resemble.
Several Portland Special Status Species (City of Portland, 2011) were observed
on ground-level sites: Bushtit, Downy Woodpecker and House Wren, as well as
Yellow, Nashville and Wilson’s Warblers. Special Status Species are considered
rare, declining or of special interest because of their associations with important
habitat attributes or conditions that support functioning ecosystems.
Common among all land types was the introduced Rock Pigeon. While this
species is not native, it is likely a major prey species for urban raptors including
Red-Tailed and Cooper’s Hawks, and Peregrine Falcons. Regular on most roofs
were Western Scrub-Jay, American Crow, European Starling, House Sparrow
and House Finch.
White-Crowned Sparrows were observed nesting in a shrub on the Louisa
greenroof during both years, and a House Finch nested in an awning
13
overhanging the same greenroof. This greenroof has taller and more structurally
complex vegetation, as well as deeper soils than others in the study, and this
additional cover could result in more nesting attempts. Combined with the lack of
ground-dwelling mammalian predators, greenroofs may make suitable nesting
sites for urban birds.
Birds flying high over sites were recorded but not included in the analysis
because it was assumed they were not interacting with the sites. Species
observed in this category included Osprey, Great Blue Heron, Vaux’ Swift, Bald
Eagle, Turkey Vulture, Red-Tailed Hawk, California Gull, Glaucous-winged Gull,
Canada Goose, Cackling Goose, Double-Crested Cormorant and Mallard. Some
additional species have been observed utilizing ecoroofs by City staff, but were
not recorded during these surveys. These additional species include Mourning
Dove, Killdeer, Barn Swallow, Rufous Hummingbird, Red-Tailed Hawk, Peregrine
Falcon, Cooper’s Hawk, Canada Goose, Band-Tailed Pigeon and Pileated
Woodpecker.
Conclusions







Higher avian abundance and species richness was detected on
greenroofs than on conventional roofs, but fewer than on ground-level
landscaped sites.
Greenroofs appear to function as an extension of urban habitats such as
ground-level parks. A diversity of native bird species, including several
species of concern, were recorded at the ground-level sites and could
therefore access and benefit from ecoroofs if they were designed for that
purpose. The absence of ground-level predators may make them
particularly beneficial to migratory aerial species, particularly if vegetative
cover were provided.
Greenroofs appear to support much more foraging activity than nearby
conventional roofs, and are comparable to nearby ground sites.
Ground sites had more insectivorous bird species utilizing the greater
plant diversity, size and associated cover, while greenroofs tended toward
more generalists species. An exception to this is the Savannah Sparrow, a
species associated with open areas with low-vegetation, which was
observed on greenroofs.
The proportion of non-native birds using ecoroofs was higher than at
ground sites.
Greater complexity of plant size and structure (on ground sites and on the
Louisa roof garden) appears positively linked to avian richness,
suggesting ecoroofs could be designed for increased habitat value.
Similar findings have been documented on other ecoroofs (Cantor 2008).
Two native species were observed nesting on or just over the Louisa roof
garden during the study: White-crowned Sparrow and House Finch. This
suggests that higher vegetation complexity and diversity would increase
the likelihood of nesting attempts on greenroofs. However, greenroofs that
14
do not offer complex vegetative structure could immediately benefit birds
that require simple nesting substrates (e.g. some ground-nesting birds;
Baumann 2006).
Preliminary Management Recommendations
To increase habitat value for birds on ecoroofs:
 Maximize vegetation and soil complexity and diversity on greenroofs when
possible.
 Consider the connectivity of greenroofs in the urban landscape,
particularly along natural features that may act as corridors (e.g. along
waterways) where they can act as extensions of existing urban habitats.
 Consider larger roofs as those will offer a larger habitat patch.
 Because more bird activity was observed on ecoroofs, abiding by more
stringent bird strike prevention measures for buildings with ecoroofs
compared to those with conventional roofs is recommended (see resource
for bird friendly building design, pg. 16:
http://audubonportland.org/files/hazards/bfbdd).
 Collaboration and good communication between green roof designers and
natural resource managers.
Considerations for Future Studies





Associated flyovers (under 50 feet) were recorded but not included in
these results as they were common at all sites and difficult to determine a
degree of association. Recording only species landing on the treatments
gave a clearer picture of bird use.
Monitoring at and just after dawn when birds are most active would likely
produce a larger sample size. Timing of these surveys was based on
arrival times of building staff, and the earlier the hour, the more difficult the
access.
Obtaining rooftop access is always an issue and occasionally a site is
missed. Getting keys from the building owner or manager is ideal.
Building management staff turnover was frequent and regular
communication and reminders are recommended.
The more sampling days the better. With a small number of sampling
days, the number of variables inherent to this type of study, and any
atypical events, makes the broader statistical patterns less clear.
Disturbances are frequent on ground sites, including construction, dogs,
maintenance or removal of vegetation, and organized events in parks.
These clearly impact the results but were included as inherent to urban
landscaped spaces. Observer presence on roofs similarly deters some
birds which would otherwise have landed. To avoid this, monitor from an
adjacent roof if possible.
15


Similarly, resident accessibility on a condominium greenroof may play a
large role in bird presence and use. The Louisa roof garden had extremely
low (zero) numbers of birds on some days. A future study could separate
data into extensive (minimally accessible) ecoroofs, and intensive roof
gardens in order to tease apart these effects more clearly.
Each season brings different birds and though this study focuses on
migration, when a high number of species are active in the area, it would
be worthwhile to monitor in summer and particularly winter, when many of
our ground-foraging species are typically most abundant.
Acknowledgements
Thank you to the following for volunteering time and resources to this study: Linda Leavens,
Dena Turner, Litzy Venturi, Will Risser, Jan Risser, Marlene Huntsinger, Margo DeBeir,
Heather Durham, Phyllis Wolfe, Wendy Shoemaker, Wendy Lee, Peggy Hackenbruck,
Jenny Jones, Cassie Deitz, Justin Bauer, Dave McCabe, Corey Ostin, Demetra Ariston,
Laura Blades, and Barbara Buhman. Audubon Society of Portland: Candace Larson, Mary
Coolidge and Bob Sallinger. City of Portland Bureau of Environmental Services: Paul
Ketcham, Linda Dobson and Dave Helzer.
References
Baumann, N. 2006. Ground-Nesting Birds on Green Roofs in Switzerland: Preliminary
Observations. Urban Habitats 4, 37-50.
Brenneisen, S. 2006. Space for Urban Wildlife: Designing Green Roofs as Habitats in
Switzerland. Urban Habitats 4, 27-36.
Cantor. S.L. 2008. Green Roofs in Sustainable Landscape Design. W.W. Norton and Co., New
York.
City of Portland (COP). 2011. Terrestrial Ecology Enhancement Strategy.
https://www.portlandoregon.gov/bes/article/354989
Eakin, C.J. 2012. Assessing Wildlife Habitat Contributions of Green Roofs in Urban Landscapes
in Michigan and Illinois, USA: Measuring Avian Community Response to Green Roof
Factors. Master’s Thesis, Michigan State University.
Fernandez-Canero. R. and P. Gonzalez-Rodondo 2010. Green Roofs as a Habitat for Birds: A
Review. Journal of Animal and Veterinary Advances 9(15): 2041-2052.
Gedge, D.; Kadas G. 2005. Green Roofs and Biodiversity. Biologist 52, 161-169.
Hintze, J. 2012. NCSS 8. NCSS, LLC. Kaysville, Utah, USA. www.ncss.com.
Huff, M.; Bettinger, K.; Ferguson, H.; Brown, M.; Altman, B. 2000. A Habitat-Based Point-Count
Protocol for Terrestrial Birds, Emphasizing Washington and Oregon. United States
Department of Agriculture (USDA).
Zar, J.H. 1999. Biostatistical Analysis, 4th ed. Englewood Cliffs, New Jersey: Prentice Hall.
16