Content uploaded by Max R Piana
Author content
All content in this area was uploaded by Max R Piana on Jun 11, 2015
Content may be subject to copyright.
Cities and the Environment (CATE)
Volume 7
|
Issue 2 Article 1
8-20-2014
Green Roofs Over Time: A Spatially Explicit
Method for Studying Green Roof Vegetative
Dynamics and Performance
Max R. Piana
KieranTimberlake, mpiana@kierantimberlake.com
Stephanie C. Carlisle
KieranTimberlake, scarlisle@kierantimberlake.com
is Article is brought to you for free and open access by the Biology at Digital Commons at Loyola Marymount University and Loyola Law School. It
has been accepted for inclusion in Cities and the Environment (CATE) by an authorized administrator of Digital Commons at Loyola Marymount
University and Loyola Law School. For more information, please contact digitalcommons@lmu.edu.
Recommended Citation
Piana, Max R. and Carlisle, Stephanie C. (2014) "Green Roofs Over Time: A Spatially Explicit Method for Studying Green Roof
Vegetative Dynamics and Performance," Cities and the Environment (CATE): Vol. 7: Iss. 2, Article 1.
Available at: hp://digitalcommons.lmu.edu/cate/vol7/iss2/1
Green Roofs Over Time: A Spatially Explicit Method for Studying Green
Roof Vegetative Dynamics and Performance
In the past decade, conventional green roof research methodology has emphasized performance measures that
assume a static state condition of vegetative composition based on design intent and establishment
conditions. Such research has predominantly been limited to short-term observations for low diversity,
rigorously maintained systems. ese conditions, however, are not the reality of many installed green roofs,
and as a result knowledge of how these living systems change over time is limited. Given this perspective, this
paper presents an ecologically grounded and spatially explicit methodology aimed at assessing the long-term
performance and dynamics of green roof vegetation. e method allows for observations of plant composition
and performance based on both statistical and graphical analysis of plant cover and diversity measures.
Application of this methodology is presented through a multi-year case study of a single, six year-old, intensive
green roof in Ithaca, New York. Applicable to any green roof, this method promotes an understanding of green
roofs as adaptive, ecological systems, a perspective that will aid in beer predicting green roof performance
over time, and inform the design, construction, and maintenance of resilient, high-performance roofscapes.
Keywords
green roofs, dynamic systems, vegetation composition, urban ecology, green infrastructure, ecosystem
services, built environments
Acknowledgements
e authors would like to thank Cornell University, particularly Art Fives, for assisting in this research and
allowing access to their facilities. Special thanks to Steve Kieran and James Timberlake for their support of this
research, as well as Billie Faircloth, Taylor Medlin, Jacob Mans, Andrea Calabrea, and Mark Ashton for their
contributions in both the eld and review of this paper.
is article is available in Cities and the Environment (CATE): hp://digitalcommons.lmu.edu/cate/vol7/iss2/1
INTRODUCTION
Green roofs, also known as vegetated roofs, eco-roofs, or living roofs, use plants to improve
building performance. Through the design and construction of engineered landscape systems,
green roofs transform otherwise unused roof surface into a living piece of green infrastructure
designed to positively affect local climate, hydrology, building energy consumption, human
comfort and well-being (Theodosiou 2003; Lazzarin et al. 2005; Villarreal and Bengtsson 2005).
Green roofs can also provide direct ecological benefits by creating important habitats for plants
and animals and increasing biodiversity and resilience of urban plant communities (Brenneisen
2003, 2006; Baumann 2006). As green roofs have come to be understood as an established
element of urban sustainability, there is an increased desire to quantify and predict the
performance attributes of green roofs over time.
While green roofs are designed holistically as engineered landscape systems, research has
determined that a significant proportion of the thermal and hydrological benefits of a green roof
assembly can be attributed to the biological function and physical properties of green roof
vegetation (Snodgrass and Snodgrass 2006; Getter and Rowe 2006; Oberndorfer et al. 2007).
While a great deal of green roof research focuses on the horticultural attributes of individual
species, in recent years, research activity has begun to employ ecological theory to ask complex
questions about the behavior and performance of green roofs, exploring spatial and temporal
relationships between vegetation and roof environments. The study of vegetation assemblage and
dynamics is by definition concerned with the “the change in species composition or in the three-
dimensional architecture of the plant cover of a specified place through time” (Picket et al.
2013). Studies of floristic relationships to resource distribution and micro-heterogeneity in roof
systems have begun to describe green roofs as complex ecosystems (Brenneisen 2003; Kohler
2006, 2010; Martin 2007; Dunnett et al. 2008; Nagasse and Dunnett 2010, 2011; Olly et al. 2011;
Nardini et al. 2012). Additionally, researchers have begun to explore the benefits of utilizing
mixed species and diverse plantings (Lundholm et al. 2010; Nagase and Dunnett 2010; Butler
and Orians 2011) to increase thermal, hydrologic, and ecological performance of green roof
assemblies (ASTM E 2400 2006; Lundholm et al. 2009, 2010).
However, an ecologically-driven approach to green roof research is only just emerging
and does not define larger green roof research trends and conventional methodology. Instead,
most green roof vegetative studies have treated plant communities as static assemblages in which
success is primarily measured through survival rates of initial vegetation, measured in average
percent cover for highly controlled and species-limited scenarios (Dvorak and Volder 2010). The
majority of green roof vegetation studies are conducted under extremely regulated and
manipulated roof conditions, over a short period of time, thereby offering little insight into the
long-term dynamics, fluctuations, and changes associated with the expected mechanical lifespans
of green roofs, which may range 40-100 years (Koasero and Ries 2007). As green roof design in
the United States has moved towards an increase in species-rich, meadow-style intensive green
roofs, the results of such studies are difficult to extrapolate and apply to novel green roof
assemblies and site-specific designs with maintenance expectations specific to the practices of
individual building owners.
1
Piana and Carlisle: A Spatially Explicit Method for Studying Green Roof Dynamics
Published by Digital Commons at Loyola Marymount University and Loyola Law School, 2014
Figure 1 The Yale Sculpture Gallery, New Haven, CT., at time of planting (2006) and in 2012. Long-term
ecological study is needed to improve the understanding of how these building systems change over time.
Efforts to accurately predict the performance of green roofs are challenged by a lack of
long-term data on green roofs, as well as a lack of research conducted on real buildings with
diverse plant communities (Cook-Patton and Bauerle 2012) under representative conditions.
While long-term vegetation dynamic studies are limited in general, there are in fact only four
commonly cited studies of this kind that are specific to green roofs (Kohler 2006, 2010; Dunnett
2008; Rowe et al. 2012). While significant in their contributions, two of these studies are devoted
to roofs with minimal diversity (Kohler 2006; Rowe et al. 2012), and they again introduce
scenarios of significant human control through the act of targeted weeding, thereby altering
floristic associations and changes over time. Even in long-term studies focused on plant
dynamics, plots may be highly maintained, excluding non-planted species through regular
weeding (Dunnett 2008; Rowe 2012).
Such industry and methodological trends reflect an assumption that the success of a green
roof is tied to maintaining a static expression of initial designed conditions with a highly
controlled growth scenario focused on “filling out” original plantings (Dunnett and Kingsburry
2004; Oberndorfer et al. 2007). This model of stasis and stability fits with a desire to achieve an
idealized condition—a state in which performance is understood and quantifiable—through a
combination of careful detailing, specification, and ongoing, rigorous maintenance activities
(Beck 2013). Indeed, many aspects of how green roofs function as dynamic novel ecosystems—
whether they are strictly maintained to preserve original planting patterns or allowed to mature
and develop—are not sufficiently understood by landscape architects, plant biologists, architects,
or engineers. While ecological theory is increasingly being applied to the study of green roof
vegetation, a gap remains in our understanding and the methods applied to address these pressing
questions.
2
Cities and the Environment (CATE), Vol. 7 [2014], Iss. 2, Art. 1
http://digitalcommons.lmu.edu/cate/vol7/iss2/1
With this perspective, this paper presents a rapid assessment methodology aimed at
collecting spatially explicit green roof vegetation data that, when interpreted through an
ecological lens, provides insight into the dynamics of green roof vegetation, and therefore roof
performance, over time. For the purpose of this paper, the application of the methodology is
demonstrated on the Flora Rose House at Cornell University, a mature intensive green roof,
subjected to minimal human disturbance since its time of planting in 2006. Field data is utilized
to provide two forms of information: 1) characterization of green roof vegetation, including
vegetation performance metrics, and 2) plant dynamics and phytosociological relationships.
Given that the data is spatially explicit, census results may be assessed visually, in
graphic form, as well as quantitatively, thereby introducing a novel means of data
communication and potentially increasing the legibility of complex spatial-temporal data. This
survey and analysis methodology is transferable across projects, allowing for comparison of
green roof system dynamics between multiples roofs varying in design intent, site context, or
geographic location. The dialogue that emerges from such inquiry encourages new questions and
alternative perspectives on how we may design, maintain, monitor, and evaluate green roofs.
Ultimately, such ecologically grounded research directed at increasing our fundamental
understanding of green roof plant dynamics has the potential to improve both initial green roof
design and the overall performance and resiliency of these systems over time.
Figure 2 Flora Rose House, Cornell University West Campus, Ithaca, New York. The green roof was planted in
2006 and surveyed in August of 2012 and 2013. The planting pallet of the roof is representative of a warm-season
grass meadow, with a mix of drought-tolerant perennial grasses and herbaceous forbs. The planting scheme is
nontraditional, deviating from the more common pallet of low biomass succulents, the minimal use of short-lived
ornamental perennials, as well as the choice of grasses with tall average height and significant periods of fallow.
3
Piana and Carlisle: A Spatially Explicit Method for Studying Green Roof Dynamics
Published by Digital Commons at Loyola Marymount University and Loyola Law School, 2014
FIELD METHODOLOGY AND DATA COLLECTION
The field methodology utilized to assess vegetation dynamics is based on the Relevé Method
(Table 1) (Poore 1955), the most accepted method for conducting vegetative surveys in the
United States and Europe (e.g. Talbot and Talbot 1994; Walker et al. 1994; Klinka et al. 1996).
Within this study methodology, the census emphasizes spatial mapping of vascular plants,
including herbaceous dicots, woody dicots, and grasses present on the green roof, and
quantitative measurement of species presence, percent cover, sociability, and vitality. Both the
mapping of plant species and the measurement of species presence and percent cover are
completed through the survey of 2 m
2
quadrants across the entirety of the roof.
In the field, the location and identity of each plant is recorded by field note diagrams of
the species footprints for each quadrant, which are later scanned and digitally transcribed, coding
each species by color. Quantitative analysis is recorded and analyzed according to cover classes
designated by the Braun-Blanquet cover/abundance scale (Braun-Blanquet 1932; Shimwell
1971; Mueller-Dombois et al. 1974), allowing the survey to accurately document growth and
coverage by a mixture of plant types, including easily identifiable individuals and clonal species
such as ground covers, emergent forbs, and grasses. Percent cover (Table 1) is recorded by
calculating the relative area occupied by the vertical projection of all aerial parts of plants as a
percentage of the surface area of the sample plot at time of survey. Species names are identified
and recorded for each plant species making up at least 5% of the cover in any quadrant. Dead
plant material from previous growing seasons is not included in the survey. It is important to
note that on the intensive roofs of this study, vegetation is often defined by complex multi-strata
layers, featuring ground covers and diverse understory composition—a characteristic of such
roofs that may be correlated to roof performance and resilience. Photographic analysis may be an
option for survey of less vertically complex green roof systems, such as sedum-dominated roofs,
but is limited in its ability to accurately characterize diverse roofs or roofs that have been
minimally maintained.
Figure 3 Field methods for assessing a mature intensive green roof. Over time, and certainly in instances of minimal
maintenance, green roofs may become ecologically more complex. The methodology presented here accommodates
a high level of plant diversity and multi-strata composition. Such community structure may increase a green roof’s
ecological function, performance and resilience to disturbance events over time.
4
Cities and the Environment (CATE), Vol. 7 [2014], Iss. 2, Art. 1
http://digitalcommons.lmu.edu/cate/vol7/iss2/1
Table 1 Braun-Blanquet cover and abundance scale (Braun-Blanquet 1932; Poore 1955).
Cover
Class
Percent
Cover
Description
Baseline
Mean
+ < 5%
Only a few (approximately 2-20) individuals of the species, and those
individuals collectively cover less than 5% of the sample plot area
0.1
1 < 5%
Numerous individuals of the species, but those individuals collectively
cover less than 5% of the sample plot area
2.5
2 5% - 25% Species cover is between 5% and 25% of the sample plot area. 15.0
3 25% - 50% Species cover is between 25% and 50% of the sample plot area. 37.5
4 50% - 75% Species cover is between 50% and 75% of the sample plot area. 62.5
5 75% - 100%
Species cover is between 75% and 100% of the sample plot area. 87.5
The field methods defined above may be repeated at regular intervals to allow for
comparison across time. Comparisons to initial establishment conditions are ideally achieved
from census data collected at year one or time of plant establishment, allowing for a robust
investigation of plant establishment and system dynamics. If an initial conditions survey cannot
be completed at time of planting, initial plant data may be approximated from the original
planting design, specifications, construction documents and nursery receipts. If available, design
and construction documentation provide detailed and spatially explicit information regarding
demarcation of planting zones, detailed stem, tray or pot counts, and plant spacing for each
species and zone of planting. Because there is no data on true plant establishment in this
approach, percent vegetative cover is not based on the percent area of roof coverage for each
species but the number of individual stems planted for each species relative to the total number
of plants.
DATA ANALYSIS
Census data from a single site survey may be analyzed to characterize and assess
phytosociological relationships and plant community composition and structure. This
information may be aggregated and simplified to generate basic vegetative performance metrics
that relate to green roof services. Additionally, data may be spatially defined and interpreted in a
number of ways, including: across the entire roof; at the plot scale; by designated zones
(planting zones, microclimate conditions, etc.); and by plant community groupings (species,
family, functional groups). In each of these forms of analysis, census results are analyzed
quantitatively and may be diagramed visually to provide snapshots of roof composition.
Ultimately, this analysis may be extended to consider data from multiple years, allowing for
observations of vegetation dynamics, including fluctuations and larger successional changes, and
their relationship with environmental stressors and disturbance events such as climate extremes
or maintenance regimes.
5
Piana and Carlisle: A Spatially Explicit Method for Studying Green Roof Dynamics
Published by Digital Commons at Loyola Marymount University and Loyola Law School, 2014
Quantitative Analysis
Vegetative characterization and performance metrics are established through analysis of percent
area of vegetative cover, species presence and richness, and species diversity. Basic percent
vegetative cover, the presence or absence of vegetation, is the most commonly used green roof
performance metric in green roof literature and remains an important indicator of basic green
roof performance and function. In addition to vegetative cover across the roof, spatially explicit
census data also provides information on the percent cover for each identified species, indicating
the cumulative area of roof represented by each plant species per survey, as well as over time.
Figure 4 Changes in species presence and percent cover for species on Flora Rose House roof (limited to species
representing over 2% cover). As the roof did not undergo any supplemental planting, species present in 2006 (in
red) represent planted species, while all others represent emergent vegetation.
Diversity, which identifies variation among species at multiple levels, is identified in this
methodology through species richness and species diversity. From a community ecology
perspective, it has been well established that there is a positive correlation between ecosystem
function and species diversity (Naeem and Tjossem 1999; Hooper et al. 2005) and species
richness (Naeem et al. 1994; Tilman and Downing 1996; Aarssen 1997; Freitas 1999; Spehn et
al. 2000). Species richness refers to the simple count of the number of distinct species present in
a given plot or across the entire roof (Cook-Patton and Bauerle 2012). Species diversity refers to
both species richness and evenness and was calculated using the Shannon-Weiner Diversity
0
5
10
15
20
25
S. scoparium
E. spectabilis
A. virginicus
T. ohiensis
C. rotundiflolia
S. heterolepis
H. americana
C. pennslyvania
A. cernuum
P. digitalis
P. subulata
A. canadensis
A. divaricatus
C. appalachica
M.didyma
A. astromontana
O. fruticosa
G. lindheimeri
J. horizontalis
A. uva
-
ursi
S. tenatum
I. cristata
T. cordifolia
T. pratense
C. varia
L.corniculatus
M. officinalis
A. dumosus
S. viridis
E. annuus
C. bibersteinii
F. rubra
S. canadensis
A. pilosus
M. lupulina
A. artemisiifolia
C. canadensis
% Cover
2006 Initial Planting
2012 Survey
2013 Survey
6
Cities and the Environment (CATE), Vol. 7 [2014], Iss. 2, Art. 1
http://digitalcommons.lmu.edu/cate/vol7/iss2/1
Index (Shannon 1948) and converted to true diversity (TD=e
Shannon Index
) (Hill 1973). While
precedent is limited, the Shannon Index has previously been used by green roof researchers
(Kadas 2002; Brenneisen 2003; Coffman 2007) as an indicator of relative biodiversity. The
strength of the Shannon Index is its utilization of a straight measure of diversity that allows for a
summary and comparison of biodiversity over time or across multiple roofs and may be utilized
to tease out relationships between diversity and services provided by green roofs such as heat
flux or stormwater retention (Bass 2009).
Table 2 Summary of vegetation performance measures across the entire roof for 2006-2012. Percent cover of
vegetation cannot be determined from construction records for initial planting years (2006). This data may be
spatially analyzed if considered at the plot
Year Species Richness % Cover Plant Families True Diversity
2006 30 NA 21 21.22
2012 74 75% 31 21.57
2013 84 87% 34 26.83
In addition to these diversity metrics, more general categorization of represented plant
families and plant types can be identified to describe and assess the green roof system. Plant
species exhibit different resource use patterns, adaptations to the external environment, and life
history strategies. Increased and complimentary functional diversity may improve green roof
performance (Naeem et al. 1994; Spehn et al. 2000). Plant types, as defined by the USDA, can be
considered a coarse proxy for functional diversity within a plant community (Lavorel and
Garnier 2002) and are a categorization method previously used in green roof studies (Lundholm
et al. 2010). It is important to note that additional methods of defining functional trait diversity,
including phylogenetics, exist and may be integrated into future iterations of this methodology
(Cook-Patton and Bauerle 2012).
Phytosociological analysis is achieved through species mapping and the Braun-Blanquet
scores collected in each plot, which translate to a sociability score for each species, a method
similar to those utilized in other plant dynamics studies (e.g. Hansen and Stahl 1993). Sociability,
as defined in this methodology, is a plot-by-plot measure of a species’ tendency to exclude other
species by forming large groups or patches, or alternatively, to grow individually or in small
aggregate clusters and integrated with neighboring species. In this study, sociability is calculated
by dividing the sum of the percent coverage value for each species in each plot by the average
coverage of that species across the entire roof. The higher the sociability score, the more likely
that species is to occur in homogenous groups or patches, while a lower score represents greater
comingling with other species.
Data Visualization
Spatially explicit data is uncommon in green roof research and presents a unique opportunity to
pair quantitative analysis with graphic-based data interpretation of vegetative performance
metrics and species distribution. Here we demonstrate examples of filters through which one
may interpret data and develop plant distribution maps, which depict overall roof composition by
species, as well as plot-based maps of vegetation performance metrics, including coverage,
7
Piana and Carlisle: A Spatially Explicit Method for Studying Green Roof Dynamics
Published by Digital Commons at Loyola Marymount University and Loyola Law School, 2014
species richness, and biodiversity. The combination of quantitative and graphic representation of
survey results provides a means for researchers to observe and communicate plant community
relationships over time and to focus on specific zones or growth patterns that may not be visible
from numerical outputs alone. As this methodology entails a full survey of the entirety of the
green roof studied, it is able to bypass many of the difficulties of geographic interpolation that
challenge traditional landscape studies on very large tracts of land, such as the tension between
data extent and sampling intensity.
Figure 5 Sequential species maps allow for visualization of plant communities over time. Within these high-
resolution species maps, each color represents a uniquely identified species. Color groupings represent family types
while white space represents areas of bare ground.
To create plant distribution maps, field notes on the footprint of each plant are recorded,
compiled, and drawn spatially. Field drawings (See Figure 2) are mapped through Adobe
Illustrator software. Individual planting masses are grouped by species and represented by a
unique colored layer; they may be imported to any spatial analysis software, such as GIS,
allowing for an analysis of floristic associations (e.g. species-specific, functional groups,
families, emergent). Species are further identified through color ramps that distinguish plant type
and allow for identification of each individual plant found on the roof.
It is common for green roof research and industry reports to discuss green roof
performance based on plant establishment, coverage, and in some instance, biodiversity. These
performance metrics can be assessed quantitatively and displayed graphically across the entire
roof or at the plot level. Plot-level analysis of performance metrics may be calculated for each
individual plot and visually mapped with each respective score communicated through a color
gradient.
8
Cities and the Environment (CATE), Vol. 7 [2014], Iss. 2, Art. 1
http://digitalcommons.lmu.edu/cate/vol7/iss2/1
Figure 6 Plot-level analysis performance matrix maps from 2012 roof census. Variability in vegetative cover and
diversity measures reveals dynamics of resilience and interaction with biophysical conditions. These maps
summarize a single survey and allow for comparison across survey years or to original planting design.
Plant assemblages and phytosociological relationships may be further analyzed visually
through various filters of inter and intra-species relationships and community groupings. The
composition and structure of plant communities is reflective of the system’s ecosystem health
and function; as such, the grouping of plants into defined communities provides another method
of insight into green roof vegetation dynamics and system performance (Cook-Patton and
Bauerle 2012). At the most basic level, one may examine dominant plant species—those species
that are the greatest contributors to total system biomass. Dominant plant species, which are
often few in number (Grime 1998) and represent less than 25% of total species richness
(Schwartz et al. 2000), are indicators of ecosystem health and function, including system
productivity. Another strategy may be to categorize vegetation by functional traits or
phylogenetic similarity—classifications that are again indicators of ecosystem productivity and
inter- and intra-species niche dynamics, such as facilitation and inhibition, within the plant
community.
These strategies of grouping plant communities may be extended to consider both species
that were part of the initial planting design and ruderal, or emergent, species. Through such
comparisons, we may confront our assumptions on the importance of maintaining a static state of
species composition on green roofs, and, perhaps, begin to consider not only what change has
occurred, but how it may affect the function, performance, and resilience of these built systems.
With limited long-term study of green roof maintenance regimes, the role of ruderals is often
overlooked or simply left to anecdotal evidence and “rules of thumb.” Few studies have a means
of systematically exploring the role of emergent species in plant succession and species
fluctuations over time, despite the potential value of such knowledge to inform the design,
maintenance, and evaluation of green roof systems.
9
Piana and Carlisle: A Spatially Explicit Method for Studying Green Roof Dynamics
Published by Digital Commons at Loyola Marymount University and Loyola Law School, 2014
Figure 7 Mapping of green roof plant community assemblages allows for assessment of various inter and intra-
species relationships that provide information regarding green roof ecosystem function, health, resilience, and
change. The above maps consider spatial distribution of plant species included in the original design (left) and
emergent vegetation (right) in 2012. Plant populations have fluctuated and shifted over time.
MAPPING AND RELATING SITE CONTEXT
Plant assemblage and dynamics, whether in natural or designed contexts, relate to the
environmental conditions of site, both abiotic and biotic. While a green roof is often composed
of a consistent assembly, the distribution of resources on a green roof is rarely homogenous,
varying in such factors as solar and wind access, thermal loading from interior or rooftop
mechanical equipment, drainage, and original species presence. Even those conditions that are
originally similar may change over time as soil characteristics evolve in response to plant life
cycles or as solar and wind access shifts in response to surrounding construction or changes in
neighboring trees and mechanical equipment. With these environmental changes, roof vegetation
and other living aspects of a roof assembly can be expected to respond and adapt in turn.
Site characteristics, such as solar access and moisture regimes (see above) may be
modeled or measured, creating spatially explicit data that provides a nuanced analysis and
explanation of the heterogeneity across the site to which vegetation responds over time. For
example, one may examine solar access on green roofs, determining hours of direct and indirect
solar radiation during growing season through the use of 3-dimensional context models (Carlisle
and Piana 2014). When examined at sufficient resolution, such data may be related to species
and performance metrics at the plot level, including percent cover, species richness, and
biodiversity. By better understanding the relationship between site conditions and green roof
dynamics, we may be better able to describe and predict the fluctuations and changes of these
living building systems over time.
10
Cities and the Environment (CATE), Vol. 7 [2014], Iss. 2, Art. 1
http://digitalcommons.lmu.edu/cate/vol7/iss2/1
Figure 8 Other spatially explicit building data, derived from models or field assessments, can be integrated with the
methodology to improve understanding of the impact of building and site context on green roof vegetation
dynamics. Examples include, analysis of solar access (left) and drainage conditions (center), which reveal a
correlation between areas of high shade, lowest roof elevations (areas of drainage), and both vegetative coverage
and species diversity. Additionally, real-time monitoring of surface and root zone temperature and moisture across
the roof (right) allows for analysis of the effects of plant coverage, community composition, and biodiversity on
plant health, biomass, and thermal performance. The thermal map (right) represents a snapshot of an animation
illustrating plot-scale soil temperature readings over time.
Often, vegetative dynamics are affected by a combination of infrastructural and
microclimatic elements. For example, when seen together, analysis of drainage systems and
overshadowing from surrounding buildings, help identify areas where roof vegetation may be
more exposed to stressors and disturbance events. An understanding of fine-scale landscape
factors, such as moisture gradients and wind exposure may inform management strategies or
supplemental planting strategies. In connecting vegetative assessments to complex climate or
building performance assessments, there is also an opportunity to introduce emerging
technologies, such as temperature and moisture
sensors, which can provide spatially comparable
data over seasons or years that may provide additional insight into the relationship between
floristic associations and system performance over time. Ultimately, developing a holistic
understanding of these constructed systems will allow for better green roof design and
management over time.
FUTURE CONSIDERATIONS AND CONCLUSIONS
An appreciation of green roofs as both adaptive ecosystems and viable pieces of green
infrastructure calls for an increased understanding of the variables that drive plant community
dynamics and their impact on long-term performance. The survey and analysis methodology
presented in this paper asks for increased rigor and nuance in our understanding of the
performance impacts of changes in roof vegetation and conditions over time. The study
methodology allows for an examination of the functional value of both intentionally planted,
nursery-grown meadow species and spontaneous urban meadow species and seeks to better
11
Piana and Carlisle: A Spatially Explicit Method for Studying Green Roof Dynamics
Published by Digital Commons at Loyola Marymount University and Loyola Law School, 2014
understand the dynamic relationships within these communities. A shift to a performance-based
and community ecology perspective allows for exploration of the possible benefits of
establishing a diverse pallet of spontaneous climate-adapted plants over time. The dialogue that
emerges from such inquiry may encourage new questions and alternative perspectives on how
we may design, maintain, monitor, and evaluate green roofs. As our focus and time scale of
consideration shifts—beyond assuring successful establishment and maintenance of initial design
intent through short-term warranty periods and toward the long-term resilience of a green roof
system and of life-cycle building performance—may the rapid adaptation and tolerance to
unpredictable changes exhibited by ruderal species be determined beneficial to these systems?
Or do the presence and establishment of such species represent a trend towards a less
ecologically rich, single-species dominant system, in which system function and resilience are
diminished?
In addition to fostering such discussion, this methodology also calls attention to a broader
need for increased legibility and communication of ecological research. In discussing the
discursive power of drawings, the architectural historian Robin Evans remarked, “To translate is
to convey” (Evans 1997). The abstraction of architectural drawings and diagrams allow for
concepts to take the place of pictorial representation—emphasizing how a systems works more
than how it appears. A map is one such form of analytical drawing that is particularly useful in
that it seeks to represent what is present—to convey an idea and to ask questions without the
assumptions and presumptions of a model and its expectation of prediction. Methodologically, as
environmental scientists and as designers, we can find agency in the act of mapping as a process
of exploration (Corner 1999), in which critical analysis of visual patterns and relationships may
be as useful as quantitative measures. For those seeking to describe and communicate patterns in
dynamic systems, reliance on fixed images remains as much of a challenge as the limitations of
capturing a snapshot of a system at a specific moment in time.
Given the potential for replicated experiments across green roofs (Felson and Pickett
2005), a collective effort should be made to observe and communicate the long-term dynamics of
green roofs subject to true building conditions and maintenance regimes. The methodology
presented here should be viewed as a building block onto which additional site data and
performance metrics may be added. Future research questions might include: What is the
dynamic relationship between vegetation and the biophysical conditions of the roof and site?
How do changes in plant composition impact the thermal or hydrologic performance of a
building? How may shifting and adaptive plant communities provide ecosystem services?
Questions such as these identify the need to increase the capacity of research to relate green roof
performance to vegetative dynamics over time and to begin to reconsider effective design and
maintenance regimes. It is through this more nuanced understanding of vegetative dynamics, and
their relationship to performance, that we may begin to develop a more holistic and relevant
approach to green roofs, from time of planting onward.
12
Cities and the Environment (CATE), Vol. 7 [2014], Iss. 2, Art. 1
http://digitalcommons.lmu.edu/cate/vol7/iss2/1
LITERATURE CITED
Aarssen, L. W. (1997). High productivity in grassland ecosystems: effected by species diversity
or productive species?. Oikos, 80, 183-184.
ASTM E 2400 (2006). Standard Guide for Selection, Installation, and Maintenance of Plants for
Green Roof Systems. West Conshohochen, PA: ASTM International.
Bass, B. (2009). Biodiversity Research on Green Roofs: Developing a Research Protocol [pdf].
http://greenroofs.org/resources/Biodiversity_Research_on_Green_Roofs_Protocol_2009.pdf
(Accessed 06/01/2013)
Baumann, N. (2006). Ground-nesting birds on green roofs in Switzerland: Preliminary
observations. Urban Habitats, 4(1), 37-50.
Beck, T. (2013). Principles of Ecological Landscape Design. Island Press.
Braun-Blanquet, J. (1932). Plant sociology: The Study of Plant Communities. New York:
McGraw Hill.
Brenneisen, S. (2006). Space for urban wildlife: Designing green roofs as habitats in
Switzerland. Urban Habitats, 4(1), 27-36.
Brenneisen, S. (2003). The benefits of biodiversity from green roofs: key design consequences.
In Proceedings of the First North American Green Roofs Conference ‘03: Greening
Rooftops for Sustainable Communities, Chicago. Toronto, ON: The Cardinal Group.
Butler, C. and Orians, C. M. (2011). Sedum cools soil and can improve neighboring plant
performance during water deficit on a green roof. Ecological Engineering, 37(11), 1796-
1803.
Carlisle, S.C. and Piana, M.R. (2014). Growing Resilience: Long-term plant dynamics and green
roof performance. In Proceedings of the 11th North American Green Roof Conference
‘11: Greening Rooftops for Sustainable Communities, San Francisco, CA. Toronto, ON:
The Cardinal Group.
Coffman, RR. (2007). Comparing wildlife habitat and biodiversity across green roof type. In
Proceedings of the Fifth Annual Greening Rooftops for Sustainable Communities
Conference, 2007: Awards and Trade Show, Minneapolis, MN. Toronto, ON: The
Cardinal Group.
Cook-Patton, S. C. and Bauerle, T. L., (2012). Potential benefits of plant diversity on vegetated
roofs: A literature review. Journal of Environmental Management, 106, 85-92.
13
Piana and Carlisle: A Spatially Explicit Method for Studying Green Roof Dynamics
Published by Digital Commons at Loyola Marymount University and Loyola Law School, 2014
Corner, J. (1999). The Agency of Mapping: Speculation, Critique and Invention. In Cosgrove, D.
(Ed.), Mapping (213-252). London: Reaktion Books.
http://www.msaudcolumbia.org/summer/wp-
content/uploads/readings2010/Corner_AgencyOfMapping.PDF (Accessed 04/28/2014)
Dunnett, N., and Kingsbury, N. (2004). Planting Green Roofs and Living Walls. Portland, OR:
Timber Press.
Dunnett, N., Nagase, A., and Hallam, A. (2008). The dynamics of planted and colonizing species
on a green roof over six growing seasons 2001–2006: Influence of substrate depth. Urban
Ecosystems, 11(4), 373-384.
Dvorak, B., and Volder, A. (2010). Green roof vegetation for North American ecoregions: A
literature review. Landscape and Urban Planning, 96(4), 197-213.
Evans, R. (1997). Translations from Drawing to Building and Other Essays. London:
Architectural Association.
Felson, A. J., and Pickett, S. T. (2005). Designed experiments: new approaches to studying urban
ecosystems. Frontiers in Ecology and the Environment, 3(10), 549-556.
Freitas, H. (1999). Biological diversity and functioning of ecosystems. In Pugnaire, F.I. and
Valladares, F. (Eds.), Handbook of Functional Plant Ecology (719-733). New York: CRC
Press.
Getter, K. L. and Rowe, D. B. (2006). The role of extensive green roofs in sustainable
development. HortScience, 41(5), 1276-1285.
Grime, J. P. (1998). Benefits of plant diversity to ecosystems: immediate, filter and founder
effects. Journal of Ecology, 86(6), 902-910.
Hansen, R. and Stahl, F. (1993). Perennials and Their Garden Habitats. Cambridge: Cambridge
University Press.
Hill, M. O. (1973). Diversity and evenness: a unifying notation and its consequences. Ecology.
54, pp. 427–432.
Hooper, D. U., Chapin Iii, F. S., Ewel, J. J., Hector, A., Inchausti, P., Lavorel, S., Lawton, J. H.,
Lodge, D. M., Loreau, M., Naeem, S., Schmid, B., Setälä, H., Symstad, J., Vandermeer,
J., and Wardle, D. A. (2005). Effects of biodiversity on ecosystem functioning: a
consensus of current knowledge. Ecological Monographs, 75(1), 3-35.
Kadas, G. (2002). Study of Invertebrates on Green Roofs: How Roof Design Can Maximise
Biodiversity in an Urban Environment. (Masters Thesis). Department of Geography,
University College, London.
14
Cities and the Environment (CATE), Vol. 7 [2014], Iss. 2, Art. 1
http://digitalcommons.lmu.edu/cate/vol7/iss2/1
Klinka, K., Chen, H. Y. H., Wang, Q. and de Montigny, L. (1996). Forest canopies and their
influence on understory vegetation in early-seral stands on West Vancouver Island.
Northwest Science, 70, 193-200.
Köhler, M. (2006). Long-term vegetation research on two extensive green roofs in Berlin. Urban
Habitats, 4(1), 3-25.
Köhler, M. and Poll, P. H., (2010). Long-term performance of selected old Berlin greenroofs in
comparison to younger extensive greenroofs in Berlin. Ecological Engineering, 36(5),
pp.722-729.
Kosareo, L. and Ries, R. (2007). Comparative environmental life cycle assessment of green
roofs. Building and Environment, 42(7), 2606-2613.
Lavorel, S. and Garnier, E. (2002). Predicting changes in community composition and ecosystem
functioning from plant traits: revisiting the Holy Grail. Functional Ecology, 16(5), 545-
556.
Lazzarin, R.M., Castellotti, F. and Busato, F. (2005). Experimental measurements and numerical
modeling of a green roof. Energy and Buildings, 37, 1260-1267.
Lundholm, J.T., MacIvor, J.S., and Ranalli, M.A. (2009). Benefits of green roofs on Canada’s
east coast. In Proceedings of the 7th North American Green Roof Conference ‘09:
Greening Rooftops for Sustainable Communities, Atlanta, GA. Toronto, ON: The
Cardinal Group.
Lundholm J, MacIvor JS, MacDougall Z, and Ranalli M. (2010). Plant Species and Functional
Group Combinations Affect Green Roof Ecosystem Functions. PLoS ONE 5(3): e9677.
doi:10.1371/journal.pone.0009677
Martin, M. (2007). Native Plant Performance on a Seattle Green Roof. (Master’s Thesis).
University of Washington, Seattle, WA.
Mueller-Dombois, D and Ellenberg, H. (1974). Aims And Methods of Vegetation Ecology. New
York: Wiley Press.
Naeem, S., Thompson, L. J., Lawler, S. P., and Lawton, J. H. (1994). Declining biodiversity can
alter the performance of ecosystems. Nature, 368, 21.
Naeem, S. and Tjossem, S.F. (1999). Plant neighborhood diversity and production. Ecoscience,
6(3), 355-365.
Nagase, A., and Dunnett, N. (2010). Drought tolerance in different vegetation types for extensive
green roofs: effects of watering and diversity. Landscape and Urban Planning, 97(4),
318-327.
15
Piana and Carlisle: A Spatially Explicit Method for Studying Green Roof Dynamics
Published by Digital Commons at Loyola Marymount University and Loyola Law School, 2014
Nagase, A., and Dunnett, N. (2011). The relationship between percentage of organic matter in
substrate and plant growth in extensive green roofs. Landscape and Urban
Planning, 103(2), 230-236.
Nardini, A., Andri, S., and Crasso, M. (2012). Influence of substrate depth and vegetation type
on temperature and water runoff mitigation by extensive green roofs: shrubs versus
herbaceous plants. Urban Ecosystems, 15(3), 697-708.
Oberndorfer, E., Lundholm, J., Bass, B., Coffman, R., Doshi, H., Dunnett, N., Gaffin, S., Kohler,
M. Liu, K., and Rowe, B. (2007). Green Roofs as Urban Ecosystems: Ecological
Structures, Functions, and Serivces. BioScience, 57(10): 823 – 833.
Olly, L. M., Bates, A. J., Sadler, J. P., and Mackay, R. (2011). An initial experimental
assessment of the influence of substrate depth on floral assemblage for extensive green
roofs. Urban Forestry and Urban Greening, 10(4), 311-316.
Pickett, S. T., and Cadenasso, M. L. (2013). Vegetation dynamics. Vegetation Ecology, 172-198.
Wiley-Blackwell, Hoboken, New Jersey.
Poore, M. E. D. (1955). The use of phytosociological methods in ecological investigations: I.
The Braun-Blanquet system. Journal of Ecology, 43(1), 226-244.
Rowe, DB, K Getter, and A Durhman. (2012). Effect of green roof media depth on Crassulacean
plant succession over seven years. Landscape and Urban Planning 104(3-4):310-319.
Schwartz, M. W., Brigham, C. A., Hoeksema, J. D., Lyons, K. G., Mills, M. H., and Van
Mantgem, P. J. (2000). Linking biodiversity to ecosystem function: implications for
conservation ecology. Oecologia, 122(3), 297-305.
Shannon, C. E. (1948) A mathematical theory of communication. The Bell System Technical
Journal, 27, 379-423 and 623-656.
Shimwell, D. W. (1971). The description and classification of vegetation. London: Sidgwick and
Jackson.
Snodgrass, EC and L. (2006). Green Roof Plants. London: Timber Press.
Spehn, E.M., Joshi, J., Schmid, B., Diemer, M., and Körner, C. (2000). Above‐ground resource
use increases with plant species richness in experimental grassland
ecosystems. Functional Ecology, 14(3), 326-337.
Talbot, S. S., and Talbot, S. L. (1994). Numerical classification of the coastal vegetation of Attu
Island, Aleutian Islands, Alaska. Journal of Vegetation Science, 5, pp.867–876.
Theodosiou, T. G. (2003). Summer period analysis of the performance of a planted roof as a
passive cooling technique. Energy and Buildings, 35(9), 909-917.
16
Cities and the Environment (CATE), Vol. 7 [2014], Iss. 2, Art. 1
http://digitalcommons.lmu.edu/cate/vol7/iss2/1
Tilman, D., and Downing, J. A. (1996). Biodiversity and stability in grasslands. Ecosystem
Management: Selected Readings, 367, 363-365.
Villarreal, E. L., and Bengtsson, L. (2005). Response of a Sedum green-roof to individual rain
events. Ecological Engineering, 25(1), 1-7.
Walker, M. D., Walker, D. A. and Auerbach, N. A. (1994). Plant communities of a tussock
tundra landscape in the Brooks Range Foothills, Alaska. International Journal of
Vegetable Science, 5, 843–66.
17
Piana and Carlisle: A Spatially Explicit Method for Studying Green Roof Dynamics
Published by Digital Commons at Loyola Marymount University and Loyola Law School, 2014
