&EPA
United States
Environmental Protection
Agency
Evaluation of Green Roof Plants
and Materials for Semi-Arid Climates
:
* «••.;"
.
Office of Research and Development
-------
EPA/600/R-12-592
September 2012
Evaluation of Green Roof Plants and
Materials for Semi-Arid Climates
By
James E. Klett
Jennifer M. Bousselot
Ronda D. Koski
Colorado State University
Fort Collins, CO 80523-1173
Thomas P. O'Connor
Urban Watershed Management Branch
Water Supply and Water Resources Division
National Risk Management Research Laboratory
Edison, NJ 08837
Cooperative Agreement No. X3-83350101
Submitted to
National Risk Management Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
October 10, 2012
-------
Notice
The U.S. Environmental Protection Agency, through its Office of Research and Development, funded and managed,
or partially funded and collaborated in, the research described herein. It has been subjected to the Agency's peer
and administrative review and has been approved for publication. Any opinions expressed in this report are those of
the authors and do not necessarily reflect the views of the Agency, therefore, no official endorsement should be
inferred. Any mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
-------
Abstract
While green roof systems have proven to be highly effective in the evaporative cooling of buildings, reduction of
roof top temperatures, protection of roof membranes from solar radiation degradation, reducing stormwater
runoff, as well as beautification of the urban rooftop landscapes throughout Europe and in several regions in North
America, green roof systems have not been evaluated in the high elevation, semi-arid regions in the United States.
Because of the risk of plant failure from incorrectly selected species, the paucity of information on green roofs in
this region, and the large potential for environmental benefits, studies were conducted on various performance
parameters on the green roof of the building that houses the EPA Region 8 Headquarters in Denver, Colorado.
Green roofs are vegetated rooftops. Green roofs provide several benefits to urban environments, including
reduction of stormwater run-off volumes and intensity, filtration of stormwater discharge, reduction of the urban
heat island effect, temperature moderation within the building underlying the green roof, and beautification of
urban rooftop landscapes. In order to provide these benefits, the green roofs must receive sufficient amounts of
water and nutrients to keep the plants alive. In the semi-arid, high elevation environment of the Front Range of
Colorado, green roof plants have not been scientifically tested for long term survivability and adaptability. The low
annual precipitation, short periods of snow cover, low average relative humidity, high solar radiation (due to high
elevation above sea level, approximately 1.6 km), high wind velocities, and predominantly sunny days all add up to
challenging growing conditions for many species of plants.
Due to the porous and well-drained nature of the typical growing media used in extensive (shallow) green roof
systems, plant species considered for use in such systems need to be evaluated for their response water
requirements and survivability and growth habits over multiple years. Thus, relative rate of dry down of the
moisture content of the media for plant species considered for use in such systems is an important characteristic to
assess. In semi-arid regions, such knowledge will help to determine the need for irrigation and the frequency of
irrigation events for these species.
-------
Table of Contents
Notice ii
Abstract iii
Contents iv
List of Figures vi
List of Tables viii
Acronyms and Abbreviations ix
Acknowledgements x
Executive Summary 1
Chapter 1. Introduction 1-1
Chapter 2. Conclusions and Recommendations 2-1
Conclusions 2-1
Single Species Study 2-1
Mixed Species Study 2-1
Zeolite Amendment Study 2-2
Volumetric Moisture Content Analysis 2-2
Moisture Deficit Study 2-3
Implications 2-3
Recommendations for Further Study 2-4
Chapter 3. Materials and Methods 3-1
Environmental Conditions of Study Sites 3-1
Plant Cover by Digital Images 3-2
Plant Cover by Converted Two-dimensionsal 3-3
Biomass 3-3
Growing Media Volumetric Moisture Content 3-4
Analytical Methods 3-5
Chapter 4. Single Species Study 4-1
Results 4-2
Comparison between Trials 4-6
Biomass Accumulation 4-8
Conclusions 4-9
Chapter 5. Mixed Species Study 5-1
Introduction 5-1
Results 5-1
Conclusions 5-5
iv
-------
Chapter 6. Zeolite Amendment Study 6-1
Introduction 6-1
Results 6-2
Conclusions 6-4
Chapter 7. Analysis of Volumetric Moisture Content 7-1
Calculated Evapotranspiration Rates 7-1
Relative Volumetric Moisture Content Analysis 7-3
Comparison of Volumetric Moisture Content and Plant Cover 7-15
Conclusions 7-19
Chapters. Moisture Deficit Study 8-1
Introduction 8-1
Greenhouse Trials 8-2
Outdoor Trial 8-6
Comparison between Trials 8-8
Recommendations 8-9
Chapter 9. References 9-1
-------
List of Figures
Figure 3-1: Layout of plants in individual modules for studies 1, 2 and 3 3-2
Figure 3-2: Initial layout of blocks for studies 1, 2 and 3 3-3
Figure 3-3: Examples of planting and volumetric moisture content sampling locations 3-5
Figure 4-1: Example of 2009 recovery as series of the same four D. cooper/ plants on Days 413, 455, 497 and
538 4-4
Figure 4-2: Plant cover determined by DIA analysis for the five experimental species over period of study 4-4
Figure 4-3: Plant cover determined by C2D analysis for the five experimental species over period of study 4-5
Figure 4-4: Plant cover determined by DIA analysis for the five experimental species for eight measurements.... 4-6
Figure 4-5: Plant cover determined by C2D analysis for the five experimental species eight measurements 4-6
Figure 4-6: Three examples of regrowth patterns 4-7
Figure 4-7: Correlation analysis between DIA and C2D plant cover methods 4-7
Figure 5-1: Plant cover for herbaceous plants of mixed species study in existing green roof growing media 5-3
Figure 5-2: Plant cover for succulent plants of mixed species study in existing green roof growing media 5-4
Figure 5-3: Plant cover for herbaceous plants of mixed species study in 50% zeolite amended green roof
growing media 5-4
Figure 5-4: Plant cover for succulent plants of mixed species study in 50% zeolite amended green roof growing
media 5-5
Figure 6-1: Plant cover as determined by DIA over eight dates during two growing seasons 6-2
Figure 6-2: Example block on July 1, 2008 showing four media 6-3
Figure 7-1: Irrigation and rainfall totals for September 7-3
Figure 7-2: Average volumetric moisture content for each species for three dates in 2008 7-4
Figure 7-3: Average volumetric moisture content for each species for three dates in 2009 7-5
Figure 7-4: Box and whisker plot of effect of species on volumetric moisture content for study 1 7-6
Figure 7-5: Box and whisker plot of effect of block on volumetric moisture content for study 1 7-6
Figure 7-6: Box and whisker plot of effect of block on volumetric moisture content for study 1 7-7
Figure 7-7: Box and whisker plot of effect of block and species on volumetric moisture content for study 1 7-7
Figure 7-8: Box and whisker plot of effect of species and irrigation type on volumetric moisture content for
study 1 7-8
Figure 7-9: Box and whisker plot of effect of zeolite on volumetric moisture content for study 3 7-9
Figure 7-10: Box and whisker plot of effect of bock on volumetric moisture content for study 3 7-10
Figure 7-11: Box and whisker plot of effect of type of irrigation on volumetric moisture content for study 3 7-11
Figure 7-12: Box and whisker plot of effect of block and amendment on volumetric moisture content for
study 2 7-11
Figure 7-13: Box and whisker plot of effect of amendment and irrigation type on volumetric moisture content
for study 2 7-12
Figure 7-14: Additional volumetric moisture content measurements of the control in study 2 7-13
vi
-------
Figure 7-15: Box and whisker plot of effect of block and irrigation type on volumetric moisture content for
study 2 7-14
Figure 7-16: Box and whisker plot of effect of amendment and block on volumetric moisture content for
study 2 7-15
Figure 7-17: Box and whisker plot of effect of plant, irrigation type and zeolite amendment on volumetric
moisture content 7-16
Figure 7-18: Box and whisker plot of effect of plant, irrigation type and zeolite amendment on plant area 7-17
Figure 7-19: Box and whisker plot of effect of block, irrigation type and zeolite amendment on volumetric
moisture content 7-17
Figure 7-20: Box and whisker plot of effect of block, irrigation type and zeolite amendment on volumetric
moisture content 7-18
Figure 8-1: Mean volumetric moisture content measurements of growing media for herbaceous plants in
greenhouse trials 8-3
Figure 8-2: Mean volumetric moisture content measurements of growing media for succulent plants in
greenhouse trials 8-4
Figure 8-3: Example block showing the change in plant appearance a) the day after the trial began and b)
day 12 8-5
Figure 8-4: Mean volumetric moisture content measurements of growing media for herbaceous plants in
outdoor trials 8-6
Figure 8-5: Mean volumetric moisture content measurements of growing media for succulent plants in
outdoor trials 8-7
Figure 8-6: Photo examples of different containers used in green house (a) and outdoor (b) trials 8-9
VII
-------
List of Tables
Table 1: Plant Responses to Media Amendments, Type of Irrigation and Environmental Conditions on Region 8
Green Roof - 2 -
Table 3-1: Weather Monitoring Equipment on the EPA Region 8 Green Roof 3-1
Table 3-2: Weather Monitoring Equipment at the Fort Collins, CO Research Location 3-1
Table 4-1: Plant Species Evaluated in the Single Species Study 4-2
Table 4-2: Days of Study and Corresponding Calendar Dates 4-2
Table 4-3: Weather Data from Region 8 Green Roof 4-3
Table 4-4: Biomass Accumulation by Species 4-8
Table 4-5: Final Correlations between Plant Cover Analysis Methods and Dry Biomass 4-9
Table 5-1: Plant Species in the Mixed Species Study 5-1
Table 5-2: Overwintering Results for the Mixed Species Study Evaluated on Day 413 (May 13, 2009) 5-2
Table 5-3: Change in Percent Plant Cover for the Mixed Species Study from September 19, 2008 to May 13,
2009 5-2
Table 5-4: Comparison of peak plant cover in 2009 in the Mixed Species Study 5-3
Table 6-1: Chemical and Physical Characteristics of the Four Growing Media 6-1
Table 6-2: Overwinter Survival for Each Sedum Taxa for Controls and Zeolite Amendments as Determined on
May 13, 2009 6-3
Table 7-1: Calculated Evapotranspiration Rates and Irrigation and Rainfall Totals 7-2
Table 7-2: Factorial Analysis of Volumetric Moisture Content for Study 1 7-5
Table 7-3: Factorial Analysis of Volumetric Moisture Content for Study 3 7-9
Table 7-4: Factorial Analysis of Volumetric Moisture Content for Study 2 7-10
Table 7-5: Factorial Analysis of Additional Volumetric Moisture Content for Study 2 7-13
Table 7-6: Factorial Analysis of Volumetric Moisture Content and Plant Cover for Study 2 7-16
Table 8-1: Physical Characteristics of the Growing Medium Used in all Three Trials 8-1
Table 8-2: Species Evaluated in Greenhouse and Outdoor Trials 8-2
Table 8-3: Mean Relative Water Use, Days to Top Growth Dieback and Percent Revival after Watering in
Greenhouse Trials 8-5
Table 8-4: Mean Relative Water Use, Days to 50% Volumetric Moisture Content Loss and Rate of Loss in
Outdoor Trials 8-8
Table 8-5: Mean Relative Water Use, Days to Top Growth Dieback and Percent Revival after Watering for
Outdoors Trial 8-8
Table 8-6: Environmental Conditions Daily Means Derived from Measurements in the Greenhouse and Outdoor
Trials 8-9
VIII
-------
Acronyms and Abbreviations
ASTM American Society for Testing and Materials
CEC cation exchange capacity
CAM crassulacean acid metabolism
CSU Colorado State University
C2D converted two-dimensional
DIA digital image analysis
EPA Environmental Protection Agency
ET evapotranspiration
ET0 evapotranspiration reference rate
ETC crop evapotranspiration
FAO Food and Agriculture Organization of the United Nations
FLL Forschungsgesellschaft Landschaftsentwicklung Landschaftsbau
Kc crop coefficient
NA not applicable
RARE Regional Applied Research Effort
SE standard error
UHI urban heat island
VMC volumetric moisture content
IX
-------
Acknowledgements
This project would not have been possible without the Environmental Protection Agency's (EPA) Regional Applied
Research Effort (RARE) grant. Thanks to the EPA, as the funding agency and primary support: Patti Lynn Tyler, Joni
Tetter, Greg Davis, Gerard Bulanowski, Patty Provencher, Bill Monson, Kathy Dolan, Craig Greenwell and especially
Thomas Slabe of EPA Region 8 who collected and organized weather data from Region 8 headquarters roof. Ms.
Carolyn Esposito of EPA Office of Research and Development reviewed the quality assurance project plan. Thanks to
many at CSU: Dr. Tony Koski, Dr. Ken Barbarick, David Staats, John Ray, James zumBrunnen, Dr. Bill Bauerle and Leah
Meyer. Thanks to those who donated advice and supplies: Gene Pielin, Panayoti Kelaidis, Mark Blanchard, Mark
Fusco, Rich Andrews, Jim Shaw, Jeff Ottersberg, Bill Adams, Harlan Hamernik, Brian Core and Dr. John White. Dr.
Robert Berghage of the State University of Pennsylvania and Dr. Matthew Palmer of Columbia University performed
reviews of this report. Primary data collection and analysis was performed by CSU and primary statistical analysis
and interpretation of Chapter 7 was performed by EPA.
-------
Executive Summary
This report is the result of a three part Environmental Protection Agency (EPA) Regional Applied Research Effort
(RARE) project that involved evaluating green roof stormwater management, mitigation of the Urban Heat Island
(UHI) effect and biological evaluation in a high elevation semi-arid area of the United States. The overall research
strategy was to facilitate the development of green roof systems within the urban communities encompassed within
the boundaries of EPA's Region 8.
Green roofs (roofs covered with vegetation) have been proven to provide several benefits to urban environments,
including reduction of volume and intensity of stormwater run-off, filtration of stormwater discharge, reduction of
the UHI Effect, temperature moderation within the building underlying the green roof, and beautification of urban
rooftop landscapes. Green roofs are based on a design that uses mineral aggregates such as expanded clays,
expanded shales, expanded slates or pumice in the growing medium. These materials provide good drainage and
stability but they tend to have low nutrient- and water-holding capacities.
There are two main types of green roofs: extensive and intensive. Extensive green roofs are characterized by
shallow growing medium, usually less than 15 cm (6.0 in.) deep, while intensive green roofs are characterized by
deep growing medium, from 15 cm (6.0 in.) up to 1 m (3 ft), but can even be deeper. The shallower extensive green
roofs are typically the only option for existing structure but are also utilized more often on newer buildings due to
increased cost for structural support and growing media for intensive roofs. However, the shallow depth and
comparatively quick drainage of the media in extensive roofs traditionally has not supported a large diversity of
plant species due to root zone limitations. Intensive green roofs are more like rooftop gardens or raised beds
because the deeper rooting depths support a wider variety of plants. Although intensive green roofs can be
aesthetically similar to at-grade gardens, the weight bearing capacity of most buildings limits their use. Therefore,
most intensive green roofs are installed on newly constructed buildings. Many environmental factors affect the
moisture content of the growing medium such as surface temperature, ambient air temperature, intensity and
duration of solar radiation, plants, relative humidity, and rate of air movement (wind), as well as growing medium
depth and composition.
Due to the porous and well-drained nature of the typical growing medium used in extensive green roof systems, the
success or failure of an extensive green roof is primarily dependent on a plant species' ability to grow in the media.
These challenges are intensified for extensive green roofs on rooftops of buildings in areas characterized by high
elevation and semi-arid climate. Success of an extensive green roof is primarily dependent on plant species ability
to survive the low moisture content of the growing medium. Plants adaptable to dry, porous soils are primarily used
in extensive green roof applications. Although Sedum species have dominated the plant palette for extensive green
roofs, there is growing interest in expanding the plant list for extensive green roof systems.
This report is the culmination of a series of three studies conducted in 2008 and 2009 to determine some
performance characteristics of several plants species grown in extensive green roof growing media. Most of the
studies were conducted on an extensive green roof located on the roof of the building housing the offices of the
EPA's Region 8 Headquarter in Denver, Colorado with a portion at the Fort Colllins Colorado State University (CSU)
- 1 -
-------
campus. Plant taxa monitored included Allium cernuum (nodding onion), Antennaria parvifolia (small-leaf
pussytoes), Artemisia frigida (fringed sage), Bouteloua gracilis (blue grama), Buchloe dactyloides (buffalograss),
Carexflacca (heath sedge), Delosperma cooper/ (hardy ice plant, trailing ice plant or pink carpet), Delosperma
nubigenum (yellow ice plant), Eriogonum umbellatum aureum (Kannah Creek® buckwheat) 'Psdowns', Penstemon
pinifolius (pineleaf penstemon), Opuntia fragilis (brittle prickly pear), Sedum acre (goldmoss stonecrop), Sedum
album (white stonecrop), Sedum lanceolatum (lanceleaf stonecrop), Sedum spurium (two-lined ortworow
stonecrop) 'Dragons Blood', Sedum spurium (two-lined stonecrop) 'John Creech', Sempervivum (hens and chicks)
'Royal Ruby' and Thymus pseudolanuginosus (woolly thyme). All plants in the study are considered perennials
(USDA, 2012). Table 1 is a summary of the plants responses to environmental conditions, differing media and
irrigation regimes across the three studies performed on the Region 8 green roof. Some plants are listed multiple
times.
Table 1 Plant Responses to Media Amendments, Type of Irrigation and Environmental Conditions on Region 8 Green Roof
Criteria
Plants benefiting from overhead rotary irrigation
Plants not benefiting from overhead rotary irrigation
Plants benefiting from 50% zeolite amendment
Plants not benefiting from 50% zeolite amendment
Plants benefiting from 33% or 66% zeolite
amendment
Plants not benefiting from 33% or 66% zeolite
amendment
Plants subject to over wintering stress
Succulents
Delosperma cooperi
Opuntia fragilis
Sedum acre1
Sedum album
Opuntia fragilis
Sedum lanceolatum
Sempervivum 'Royal Ruby'
Delosperma cooperi
Sedum spurium 'Dragons Blood'
and 'John Creech'
Sedum acre
Sedum album
Sedum acre
Sedum album
Herbaceous
Antennaria parvifolia
Bouteloua gracilis
Allium cernuum
Bouteloua gracilis
Antennaria parvifolia
Eriogonum umbellatum aureum
Not Applicable
Not Applicable
Antennaria parvifolia
Buchloe dactyloides
Eriogonum umbellatum aureum
1 Potential overwintering stress on these species due to desiccation may confound this result as plants may leaf out
in spring and die later.
In the Single Species Study, plant area covered (plant cover) was determined using two methods: digital image
analysis data (DIA) and two-dimensional data (C2D). The first is based on pixel analysis of digital images of the
plants, and the latter is based on manually collected measurements of the plants; comparisons were then made
between the two methods. For each of six plant species in the study, digital images and manual two-dimensional
measurements were taken on four dates (at six week intervals) in 2008 and on four dates (at six week intervals) in
2009. Using SigmaScan Pro 5.0 image analysis software, DIA was performed on these images. Additionally,
comparisons were made between DIA data and final biomass, and C2D and final biomass. Eight individual plants
were planted, each with a 93 cm2 (1.0 ft2) square of growing space and 10 cm depth. Plant cover increased for all six
species during the 2008 growing season. Due to the low overwintering rate (12.5%) of E. umbellatum aureum, this
species was removed from analysis in 2009. In the spring of 2009, four of the five remaining species exhibited
decreased plant cover due to winter dieback; the one exception was the cactus O. fragilis though this increase may
have resulted in part from damaged pads being replanted early on in the study. In terms of plant cover, both
quantification methods (C2D and DIA) revealed that B. gracilis and D. cooperi outperformed A. parvifolia, O. fragilis,
and S. lanceolatum. Thus, five of the six species evaluated in this study appear to be appropriate for use in extensive
green roof applications in high-elevation semi-arid areas with little snow cover, if irrigated in the growing season.
In the Mixed Species Study, five modules were filled with the existing green roof growing medium used for EPA
Region 8 green roof and the other five were planted with a 50% by volume zeolite amendment with the existing
-2-
-------
growing medium. These ten modules were planted with one plant of each of the following eight species (A.
cernuum, A. parvifolia, B. gracilis, D. cooper/, E. umbellatum aureum, O.fragilis, S. lanceolatum, and Sempervivum
'Royal Ruby'). One module of each growing medium type was placed in each of the five blocks. Similar plant cover
data as in the first study were collected. At the end of the study, the two species that had the highest plant cover
were B. gracilis and D. cooper/; plant cover for all other species was much lower. Similar to the Single Species Study,
E. umbellatum aureum and A. parvifolia had the lowest overwintering rates, though E. umbellatum aureum had
greater plant cover than A. parvifolia and increased in plant cover through 2009, thereby benefiting from the mixed
stand planting, while A. parvifolia declined, especially in the 50% zeolite amendment. As in the Single Species Study,
O. fragilis increased in plant cover over; this increase was not as dramatic, though it confirms increase was not due
to extra plantings.
In the Zeolite Amendment Study, blends of a typical green roof growing medium (GreenGrid®) and a zeolite product
(ZeoPro™ H-Plus) were examined on the EPA green roof. The four growing media evaluated included a GreenGrid®
control with no ZeoPro™ H-Plus, 67% GreenGrid® with 33% ZeoPro™ H-Plus, 34% GreenGrid® with 66% ZeoPro™ H-
Plus, and no GreenGrid® with 100% ZeoPro™ H-Plus. Plants used in this study included S. acre, S. album, and S.
spurium cultivars 'Dragons Blood' and 'John Creech', all of which were already in use on the EPA green roof. The four
growing media mixes were evaluated based on plant taxa growth performance. During the initial year, the additions
of zeolite to the typical extensive green roof growing medium improved plant cover for all four plants. However, S.
acre and S. album had poor overwintering success and died out late in the spring of 2009. Conversely, the two
cultivars of S. spurium, which were native to the area, exhibited an increase in plant cover during the second year. S.
spurium cultivars 'Dragons Blood' and 'John Creech' had the highest plant cover in the with 33% and 66% ZeoPro™
H-Plus mixtures. The 100% ZeoPro™ H-Plus negatively impacted plant cover for all these species, but serves as a
proof of concept of potential alternative green roof media.
Volumetric moisture content (VMC) data were collected from the modules described in the above studies over
variety of dates. The overhead rotary irrigation system installed in June, 2009 delivered a more consistent amount
of water throughout the green roof as measured by instantaneous VMC measurements. Less irrigation was applied
in 2009 with the spray irrigation, than in 2008 with the drip irrigation system. Overall, the overhead rotary irrigation
increased biomass and plant cover. Many individual plants benefitted from the switch from drip irrigation to
overhead rotary irrigation, however, several Sedum species, i.e., S. acre, S. album, S. lanceolatum actually declined.
S. acre and S. album had low overwintering success to begin with, while S. lanceolatum had lower plant cover at the
end of the Mixed Plant Study in the control medium (without zeolite amendment) but significantly increased in plant
cover with the 50% zeolite amendment. As the percentage of zeolite in the growing media increased, VMC also
increased, despite the fact that laboratory results showed decreasing water holding capacity as zeolite percentage
increased. The effects of shading, i.e., increased VMC in one section of the roof, were statistically observed in the
Single Species Study but not in the Mixed Species Study due to different plant uptake rates of water under varying
conditions.
In the fourth study, Moisture Deficit Study (or dry down study), eight succulent species and seven herbaceous
species were dried down at different rates over a period of five months. The Moisture Deficit Study was conducted
on Fort Collins campus of Colorado State University; this was a two part study with one part taking place inside a
greenhouse, and the second part taking place out of doors. In this study, fifteen plant taxa were evaluated for
response to gradual and long-term drying of the porous extensive green roof growing medium. Taxa evaluated
were A cernuum, A. parvifolia, A.frigida, B. gracilis, B. dactyloides, C.flacca (heath sedge), D. cooper/, D.
nubigenum, P. pinifolius, S. acre, S. album, S. lanceolatum, S. spurium 'John Creech', Sempervivum 'Royal Ruby', and
T. pseudolanuginosus. Despite differences in dry down, the succulent species, as a group, maintained viable foliage
for over five times longer than the herbaceous species. The revival rates of the succulent species were nearly double
those of the herbaceous species. These results indicate that succulent species are more likely to be longer-lived
during periods of drought and are more likely to resume growth soon after water is made available. Based on these
-3-
-------
results, irrigation frequency is recommended for succulent species at a maximum of 28 day intervals and
herbaceous species at maximum of 14 day intervals in the semi-arid, high elevation environment of the Front Range
of Colorado during non freezing conditions; however, this minimal irrigation recommendation assumes the media
be brought to field capacity at the beginning and end of intervals, either through rainfall, irrigation or combination
thereof.
Due to diverse effects observed in this study due to changes in irrigation regime, varying results for species between
the monoculture and mixed stand plantings, and interaction effects with zeolite amendments, future studies should
look at root growth in addition to top growth of plants, especially for herbaceous species. The low overwintering
success or eventual die-off of several species in the study, i.e., S. acre, S. album, A. parvifolia, B. dactyloides and E.
umbellatum aureum and over all winter dieback of most of the observed species may be an indication desiccation of
roots due to limited snow cover and winter precipitation. Plants that did survive the winter may be competitively
better at obtaining water resources. An additional limited irrigation regime to prevent plant desiccation during
winter months may improve some plants survival in extensive green roofs in arid regions.
-4-
-------
Chapter 1 Introduction
Green roofs are planted for many reasons, including stormwater management, reducing the urban heat island (UHI)
effect, and beautification of urban roof top landscapes. In order to maximize these benefits, the plants of the green
roofs have to remain alive. In the semi-arid, high elevation environment of the Front Range of Colorado, green roofs
have not been scientifically tested for long term survivability and adaptability. The low annual precipitation, short
periods of snow cover, low average relative humidity, high solar radiation (due to elevation above sea level,
approximately 1.6 km), high wind velocities, and predominantly sunny days all add up to challenging growing
conditions for many species of plants. Therefore, plants from other environments characterized by more ideal
growing conditions, i.e. high moisture, high humidity and more cloud cover, may not survive when planted in an
extensive green roof.
The success of plants on an extensive green roof is primarily dependent on the particular species' ability to survive
the often low moisture content of the growing media. Due to the well-drained nature of the soil-less growing media,
plants capable of surviving dry, porous soils are primarily used in extensive green roof applications. Cuttings of
Sedums are often used on extensive green roofs because of their relative tolerance to moisture deficit conditions
and the fact that many are evergreen groundcovers. Although Sedums have dominated the plant palette for
extensive green roofs, there is increased interest in expanding the plant list for extensive green roof systems.
Researching additional plant species not already in use on extensive green roofs will expand the plant palette.
Diversifying the plant palette of green roofs, especially with native species, will potentially open additional habitat
choices for macroinvertebrates and bird species in urban areas.
The modern extensive green roof is based on a design that uses mineral aggregates such as expanded clays, shales,
and slates, or pumice in the growing media. These materials provide good drainage and stability but tend to have
low nutrient- and water-holding capacities. Amendments with compost can increase water holding capacity and
nutrients, but can also leach nutrients. To date, region-specific research on extensive green roof growing media
mixes has been limited. For example, zeolite was tested as a 10% blend to pumice and compost by weight to develop
local green roof materials for New Zealand, reducing cost for transport (Passman and Simcock, 2008). Therefore
additional research on growing media mixes appropriate for use on green roofs is necessary. Similar to the need for
diversifying plant species on a green roof, additional growing media amendments and mixes will benefit green roof
systems as well, especially as the plant palette increases.
Colorado has several native plant species that grow in shallow, rocky, well-drained soils which may be good
candidates for extensive green roof plants (Getter and Rowe, 2006). Several plant species native to Colorado along
with one species of plant native to an area of South Africa with growing conditions similar to those that occur in
1-1
-------
Colorado, Delosperma cooper/ (hardy ice plant, trailing ice plant or pink carpet), were tested alongside other Sedums
already on the EPA Region 8 green roof. In order to effectively select suitable plants, species need to be evaluated in
terms of response to the climatic conditions and ability to adapt to the extensive green roof growing media.
Most commercially available extensive green roof growing media is predominantly made up of expanded slate, shale
or clay. These lightweight, well-drained materials (but not so light that they blow away) do not break down like
organic materials. However, these materials typically drain quickly (due to macro-pore space) and do not hold
nutrients very well due to low cation exchange capacity (CEC). Compost has been commonly added to increase
nutrient and water retention content but this is an organic amendment. The German Forschungsgesellschaft
Landschaftsentwicklung Landschaftsbau (FLL) guidelines limit organic content within green roof media to reduce
potential for shrinkage of the vegetation support and to limit impact to long term success of the roof (Philippi,
2011). The FLL recommends 8% (dry weight) or less organic content based on the loss on ignition method; however,
recommendations for higher organic matter content have topped 15% in drier climates (Miller, 2011). As an
extensive green roof matures though, organic content will drop to about 4% due to oxidation and biodegradation
(Miller, 2011). The FLL (2008) guidelines also recommend a distribution of aggregate sizes for vegetation substrates.
Inorganic materials that have all of the benefits of expanded slates, shales or clays, plus have more micro-pore space
and higher CEC would be ideal amendments to existing extensive green roof growing media. A potential amendment
or alternative medium would be a zeolite. Zeolites are hydrated aluminosilicates of alkaline and alkaline-earth
metals (Virta, 2001). There are about 40 naturally occurring zeolites and hundreds of synthetic zeolites; both
naturally occurring and synthetic zeolites have widespread commercial applications because of the unique sorption,
ion-exchange, molecular sieve and catalytic properties zeolites offer (Virta, 2001). Usage of naturally occurring
zeolites has been increasing in horticulture, typically as a soil conditioners or growth media, though this is not
currently one of the top four domestic usages of these minerals (Virta, 2009). Research with turfgrass demonstrates
higher moisture contents in substrates that contain clinoptilolite (a natural zeolite) than in sand alone (Miller, 2000;
Murphy et al., 2005).
Data collected through this project were used to complete the following four studies:
1. The Single Species Study determined suitability of six plant species for extensive green roof use in the semi-
arid, high elevation Front Range of Colorado.
2. The Zeolite Amendment Study determined suitability of zeolite as growing media amendment for
supporting plant growth in an extensive green roof system.
3. The Mixed Species Study evaluated mixed stands of trial plant species when grown in extensive green roof
growing media and media with 50% zeolite amendment.
4. The Moisture Deficit Study determined the impact of moisture deficit on 15 plant species through controlled
dry downs, i.e. periods without irrigation.
The research for studies 1, 2 and 3 took place on the EPA Region 8 Headquarters green roof in downtown Denver,
CO. The research for Study 4 was performed on the Colorado State University (CSU) campus in Fort Collins and
included eight succulent and seven herbaceous species. The plant studies 1 and 3 are presented in Chapter 4 and 5,
respectively, followed by studies 2 and 4 in Chapter 6 and Chapter 8, respectively. Chapter 7 discusses the
measurements of volumetric moisture content (VMC) for studies 1, 2 and 3.
Overwintering success in these experiments is of vital importance as Front Range Colorado winters are typically
characterized by warm sunny days (frequently up to 15^C [SO^F] or above) and freezing nights with high winds
occurring often and unpredictable precipitation and snow cover duration. These environmental conditions are
difficult for plants due to moisture limitations. Plants still require moisture during the winter to prevent winter
1-2
-------
desiccation and maintain adequate root metabolism. Drought resistance in the shallow, well-drained media of
extensive green roofs is a significant factor of plant survivability. In addition, different plants use water at different
rates; therefore plant water use determines the appropriateness of plants in green roof applications.
Error bars have been used in the figures. However, standard error was used in figures for plant cover analysis rather
than standard of deviation. This is because error bars with standard of deviation would obscure data and general
trends. Standard error was calculated by dividing the standard deviation by the square root of the number of test
performed. Standard error still shows relative variance of the individual species to the various trials without
obscuring data points or general trends. Standard deviation was used in the figures for VMC analysis.
1-3
-------
Chapter 2 Conclusions and Recommendations
Conclusions
The following conclusions are based on evaluations over two consecutive growing seasons on an extensive green
roof in a semi-arid, high elevation location with irrigation.
Single Species Study
In general, all six species studied increased in plant cover during 2008 for both digital image analysis data (DIA) and
two-dimensional data (C2D) data sets. However, four of the five species showed temporary declines in plant cover
after winter dormancy with the initiation of the second year of monitoring in May, 2009, the exception being
Opuntiafragilis (brittle prickly pear). This reduction in plant cover is likely due to dieback due to overwintering
stress; the sixth species, Eriogonum umbellatum aureum (Kannah Creek® buckwheat) 'Psdowns', had severe die-off
and was not monitored in 2009. A similar phenomenon can be observed in the growth index graphs for species
evaluated in a Michigan study, specifically, Agastachefoeniculum, Aster laevis, and Coreopsis lanceolate/
(Monterusso et al., 2005). Analysis by DIA and C2D were comparable though the early bloom of S. lanceolatum and
subsequent senescing in 2009 led to higher readings by C2D than with DIA.
On the final date of plant cover comparisons (Day 538 [9/15/2009]), the two species with the highest plant cover
were Bouteloua gracilis (blue grama) and D. cooper/, with the remaining three species closely grouped in plant cover
(Antennaria parvifolia (small-leaf pussytoes), O. fragilis and Sedum lanceolatum (lanceleaf stonecrop)). While 6.
gracilis and D. cooper/were more successful than A. parvifolia, O. fragilis and S. lanceolatum these latter species
survived and resulted in a net increase in plant cover so these species should still be considered for use on extensive
green roofs.
Mixed Species Study
Similar to the Single Species Study, the two species that had the highest plant cover were D. cooper/ (3950 cm2) and
B. gracilis (1220 cm2) while plant cover for all other species was much lower (> 1000 cm2). The results for both these
species were also higher in the 50% zeolite amendment, approximately 5 and 25%, respectively. Unlike the Single
Species Study, B. gracilis showed poor overwintering in both regular media (65% loss) and 50% zeolite amendment
(75% loss), possibly due to competition from the mixed stand of species. However, E. umbellatum aureum was much
more successful than in the single species study, indicating that this species benefitted from the mixed stand
2-1
-------
planting. The 50% zeolite amendment also appeared to detrimentally affect A parvifolia, which had 60%
overwintering loss and 63% reduction in plant cover, and Allium cernuum (nodding onion), which had 25%
overwintering loss and 43% reduction in plant cover. Both of these species also had less than 100 cm2 cover. At end
of the 2009 growing season, surviving plants in the zeolite had greater peak plant cover with increases of 26% for
the herbaceous plants and over 36% for the succulents, with the exception of A cernuum, which actually declined.
Zeolite Amendment Study
The addition of zeolite to the growing media used in an extensive green roof system potentially improved
establishment year growth for Sedum acre (goldmoss stonecrop) at 33 and 66% mixtures and Sedum album (white
stonecrop) at all zeolite mixtures, but hindered overwintering success, particularly at 66% and 100% zeolite for both
species. The ultimate die-off of almost all Sedums at all levels of zeolite mixture and in the controls was most likely
an indication of desiccation of the roots during winter.
Of the two cultivars of S. spurium, both 'John Creech' and 'Dragons Blood' had minor benefits at the 33% and 66%
addition of zeolite in the first year (2008) but had greater benefits in the second year (2009). This coincided with a
switch in form drip irrigation to overhead rotary irrigation. For 'John Creech' this was at all levels of zeolite, while
'Dragoons Blood' was limited to improvements in the 33% and 66% mixtures.
The addition of zeolite to extensive green roof growing media may be beneficial for some but not all species. The
100% zeolite negatively impacted over winter success of all species studied; however, this is a proof of concept that
other media may be pursued for green roof applications. Competition through variation in water use by the plants in
this study may have impacted the survivability of S. acre and S. album.
Volumetric Moisture Content Analysis
VMC data suggest that the overhead rotary irrigation system was more efficient than the drip irrigation at supplying
uniform distribution of water. This is due to the quick, vertical draining properties of the media which does not allow
for lateral water movement (except along drainage layer which may be beyond root system) and could nullify the
benefits of drip irrigation if plants are not directly under the emitter.
The potential effects of shading a portion of the green roof by the upper floors of Region 8 headquarters building
were observed in the Single Species Study, as volumetric moisture content (VMC) of the east side of the roof was
statistically higher than other sections of the roof during the study. This effect was not observed during the Mixed
Species Study. In the Single Species Study, plants would be expected to use water at the same rate, while there was
greater variation in plant water uptake rates in the Mixed Species Study, due in part to varying conditions across the
roof and interactions between plants.
In general, VMC increased with increasing zeolite content of the growing media despite the fact that laboratory
results showed decreasing water holding capacity as zeolite percentage increased. Statistical analysis indicated that
lower observed VMC in 33% and 66% ZeoPro™ H-Plus mixtures correlated with increased plant cover in 2009 when
the overhead rotary irrigation system was in use.
Less irrigation was applied in 2009 with the overhead rotary irrigation, than in 2008 with the drip irrigation system.
Year to year for the months July through September, there was 10% more rainfall in 2009, i.e., 97 mm compared to
88.1mm, but there was 32% less irrigation required, i.e., 200 mm compared to 270 mm.
2-2
-------
O.fragilis , a cactus, consistently had higher VMCthan other species and plant cover consistently increased after
two seasons. It would appear that irrigation and rainfall rates were more than this species required.
Moisture Deficit Study
While there was no clear division between succulent and herbaceous species in dry down curves, there were
differences among species within plant types. Additionally, relative water use during the 18 day dry down was
inconsistent within plant type. However, the general trend was that the growing medium planted with succulent
species retained more moisture for a longer period of time than did the growing medium planted with the
herbaceous species.
Dieback and revival rates differed by plant type as well. The succulent plant species had viable foliage for over five
times longer than the herbaceous plants in the greenhouse. After dieback, the revival rates of the succulent plants
were nearly double the herbaceous. Therefore, not only are the succulents more resistant to drought stress at the
onset of extended periods of insufficient moisture in planting medium, but they have a better chance of recovering
after a drought once water is again made available.
Irrigation frequency recommendations for extensive green roof culture in the high elevation environment of the
Front Range of Colorado varies by plant type; succulent species should be irrigated at least every 28 days while
herbaceous species should be irrigated at least every 14 day intervals (more frequently for species with high water
use requirements). Irrigation frequency will need to increase if the duration of an irrigation event supplies an
insufficient amount of water to satisfy a species water use requirements, i.e., irrigation, rainfall event or
combination thereof is below field capacity of the media.
Implications
After the initial year of this study, it became obvious that the drip irrigation system was not suitable for supplying
the water needs of the Sedum plants growing in an extensive green roof in a semi-arid, high elevation environment
due to the well-drained nature of the growing medium. Sedums are shallow rooted plants that can be started on
roofs by planting individual plugs or by spreading cuttings out over the roof. Only a small cone of moisture
developed around the emitters of the drip irrigation system which were spaced approximately 30 cm (1 ft) apart.
The overhead rotary irrigation system provided more uniform coverage of water over the area than did the drip
irrigation. The implication is that overhead rotary irrigation overall increased plant cover, partly by allowing Sedums
to spread across roof because more uniform moisture was available.
For the Zeolite Amendment Study, complications from overwintering i.e., cold temperatures, lack of rainfall and
absence of irrigation, affected the species S. acre and S. album, and to a lesser extent the S. spurium cultivars. While
many factors could influence overwintering in the amended green roof growing media, it is clear that having a
portion of zeolite in the growing media improved plant cover for some of the species but did not improve plant
cover in for all species monitored. Winter irrigation may be required for some plants if there is lack or rainfall or
snow cover to prevent desiccation of roots.
For the Moisture Deficit Study (dry down study), rates of growing media dry down over the initial 18 day period
were variable by species for both the herbaceous and succulent groups. However, the days to dieback and revival
after rewatering show clear differences in the two groups of plants with succulents taking over five times longer to
dieback and almost twice as successful at recovering compared to the herbaceous species.
2-3
-------
Recommendations for Further Study
Each of these studies could have follow-up research. For example, dozens of species could be evaluated similar to
the Single Species Study. Additional growing media blends incorporating zeolite at finer scales, i.e. at 5% increments
to approximately 65%, or different amendments could be investigated to determine if these blends improve the
moisture-holding and nutrient-holding capacity of green roof growing media. Irrigation frequencies for additional
species grown in an extensive green roof system could also be evaluated. As zeolite amendments appear to benefit
some plant species while not improving or even hindering others, further studies of species specific reactions to
zeolite amendment are required, particularly as it pertains to overwintering success or loss. Also, studies of zeolite
amendments as it applies to a mixture of plantings, which is the typical practice in green roof applications, are
necessary.
For the Single Species Study, the intent for the selection of the species evaluated in this study was to add diversity to
the list of species suitable for extensive green roof cultivation. These species survived the low moisture conditions
of semi-arid region with annual average precipitation less than 400 mm. Other conditions of the Front Range
including high solar radiance, high wind velocities and high number of sunny days contribute to high
evapotranspiration rates and a need to irrigate the plants in this environment. The calculated evapotranspiration
reference rate (ET0) was approximately 960 mm for the period between March and October, 2009, while total
rainfall and irrigation for this period was 660 mm. The species evaluated in this study can be recommended for more
widespread use on extensive green roofs in regions with less overall harsh conditions, and correspondingly lower
ET0. The succulent species should be tested for non-irrigated roofs in areas with annual growing season
precipitation exceeding 500 mm which corresponds to approximately half the observed ET0 during months for which
ET0 could be calculated in this study.
The low overwintering success or eventual die-off of several species in the study, i.e., S. acre, S. album, A. parvifolia,
B. dactyloides and E. umbellatum aureum may be an indication desiccation of roots due to limited snow cover and
rainfall during the winter. Measurement of moisture content and development of a limited irrigation regime to
prevent desiccation during winter months may improve plant survival in green roofs in arid regions or areas of
limited snow cover.
Due to the observations that plant success in the Single Species Study and Mixed Species Study varied by species,
with either adverse and beneficial interactions most likely due to differing water usage rates by the species
observed, additional studies of both individual plants and mixed plantings is warranted. A similar recommendation
was made recently by Cook-Patton and Bauerle (2012). Tracking mixed plant species studies may identify species
that mutually benefit survivability and plant growth and may yield other benefits like increased runoff uptake or
cooler roofs. The potential for increased stormwater control due a mixed stand planting was observed by Lundholm
et al. (2010) as well cooler temperatures in the substrate.
Additional analysis of plant biomass may be warranted. Sedums have shallow rooted systems which would seem
more well suited for overhead rotary irrigation while some herbaceous species have tap roots which might benefit
from drip irrigation if the emitter is placed next to plant. Assessing root mass in addition to top growth may provide
further insight into choosing the right irrigation system for the individual type of plants, especially herbaceous plants
which have more extensive root systems than succulents.
2-4
-------
Chapter 3 Materials and Methods
Environmental Conditions of Study Sites
Environmental conditions were monitored at five minute intervals on the EPA Region 8 green roof in Denver,
Colorado (studies 1-3) by use of Campbell Scientific weather monitoring equipment (Table 3-1). The dry down
characteristics of extensive green roof growing media and the impact of dry down on plant species (study 4) were
evaluated in greenhouse and outdoor trials at the CSU campus in Fort Collins, CO. Environmental conditions were
monitored in Fort Collins by HOBO weather monitoring equipment at 15 minute intervals (Table 3-2).
Table 3-1: Weather Monitoring Equipment on the EPA Region 8 Green Roof
Campbell Scientific Equipment |
(Model #) ||
Infrared Radiometer (IRR-P) ||
Temperature and Relative Humidity Probe |
(HMP45C) ||
Young Wind Sentry set (03001-L) |
Tipping Bucket (TE525WS-L) ||
Snowfall conversion adaptor (CS705) ||
Silicon Pyranometer (LI200X) ||
Datalogger (CR1000) ||
Description
Surface temperature of vegetation
Measures temperature and relative humidity
at 0.3m (1ft) height
Wind speed and direction
at 1 m (3 ft) height
Precipitation gage
Converts snowfall into rain equivalent
Solar radiation sensor
Range of
Tolerance
-55° to +80°C
-40° to +60°C
0 to 50 m/s
0° to +50°C
to -20°C
-40° to +65°C
Data storage device ||
Table 3-2: Weather Monitoring Equipment at the Fort Collins, CO Research Location.
Onset/ Apogee (Model #)
HOBO® Temperature and Relative Humidity Probe (U12)
Apogee Precision Pyranometer (SP-1 10)
Description
Measures temperature and RH
Solar radiation sensor
Range of Tolerance
-20° to +70°C
-40° to +55°C
HOBO® Datalogger (U12-013) || Data storage device ||
For studies 1, 2 and 3, after determining which plants survived, plant area data (plant cover) measurements were
taken over time to determine success. Digital images were taken throughout the growing season to determine
growth rate by measuring change in plant cover over time. Whenever digital images were taken, two plant widths
and plant height were also recorded in cm. Growing media VMC data were collected to determine relative water use
of plants for all three studies. At the end of the experiments, study 1 plants were harvested so that top growth or
above ground biomass could be determined for each plant.
3-1
-------
Figure 3-1 shows the layout of individual plants for studies 1-3. A label was pasted on one end of each module and
all modules were oriented the same direction. Figure 3-2 shows the layout of modules in blocks for studies 1-3.
8
I
•
I
i
f
7
3 1 4
I
•
For Study 1 (species)
and
Study 3 (mixed)
Study 2 (media)
t
t
Labels for modules facing southwest
Figure 3-1 Layout of plants in individual modules for studies 1, 2 and 3.
Plant Cover by Digital Images
As a measure of plant growth rate and success, plant cover (cm2) digital images were taken every two weeks. The
data was analyzed by DIA and presented for four dates (at 6-week intervals) in the growing season of 2008 and 2009
as results were most demonstrative at these intervals (the same intervals were used for studies 2 and 3 and C2D).
A Fuji Film S3000 3.2 mega pixel camera with a six times optical zoom lens was mounted to a Bogen Manfrotto
190xprob tripod (Ramsey, NJ) with an extendable horizontal arm. A plum bob was used to ensure that all photos are
taken from a preset distance, and a bubble level on the back of the camera ensured the photo orientation was
consistent for every picture. The same camera and image settings were used to keep constant any differences these
factors could make in image quality.
3-2
-------
The digital images were analyzed using SigmaScan Pro 5.0 image analysis software (SPAA Science, Chicago, IL). This
image analysis was used to draw outlines for each plant in each digital image. Durham et al. (2007) successfully used
this method in their trials to measure growth rates of green roof species.
KEY
s
o
3T
1-t
S10
S3
50%
66
100
S3
0%
33
an
SIS
p
SI A
33
100
SIB
Sib
Ot6
66
SID
2
0
^r
w
SIS
S3
0%
SI A
66
33
SIB
Ot2
100
Ot7
66
33
100
S3
50%
B=Block T=Treatment
S=Study L=Location
FYI, Purple flags at corners of blocks
S2
SJ
A=Antennaria, B=Bouteloua. D=Delospermar E=Eriogonom, O=Opuntia, S=Sedum
0=no Zeolite in media, 33=one third Zeolite, 66=two thirds Zeolite, 100=a!l Zeolite
0%-no Zeolite in media, 50%=half of media (by volume) is Zeolite
Penthouse Area
Roughly North/"
03
*
UJ
S3
50%
S1E
SID
33
100
S3
0%
S1O
ota
66
SI A
SIB
SIS
100
66
33
Ot3
03
I
f.
S10
SIS
SI A
100
66
S3
50%
33
Ot9
SID
S1E
33
66
S3
0%
Ot4
100
SIB
03
O
£
t/i
100
66
SI A
33
66
33
Ot5
100
OHO
S1E
SIB
S3
50%
S3
OK
SIS
SID
S1O
Figure 3-2 Initial layout of blocks for studies 1, 2 and 3.
Plant Cover by Converted Two-dimensionsal
Concurrent to the DIA, individual plant widths were measured four times in the growing season in of 2008 and 2009
for each of the three studies. Two widths, one parallel to the short end of the module (0.61 m [2 ft]) and the other
was perpendicular to it, were measured using a ruler down to 1 mm to achieve C2D.
Biomass
All above growing media portions of each plant in study 1 were harvested at the end of the experiment. Root
weights were not measured because neighboring plant roots grew together and would be difficult to separate.
Plants were cut at growing media level, rinsed in water to remove growing media debris, patted dry with a paper
towel and fresh weight mass was recorded. Samples were inserted into a pre-labeled 13 x 8 x 27 cm (5 x 3 x 10 in.)
3-3
-------
brown paper bag (Rite Aid, Harrisburg, PA) to allow air and water movement through the paper. The samples were
dried in an oven at 70^C for 72 hr and weighed for biomass.
Growing Media Volumetric Moisture Content
The commercially available extensive green roof growing media is a proprietary blend created for use in the
GreenGrid® product line. The mix is 80% inorganic and 20% organic materials by volume. Growing media physical
properties were analyzed by Hummel & Co, Inc. Laboratory in Trumansburg, NY, USA. All physical properties were
tested per American Society for Testing and Materials (ASTM) E2399. Analytical methods included organic matter
(ASTM F1647, method 1, loss on ignition), dry density, particle density (ASTM D5550), saturated hydraulic
conductivity (permeability), total porosity, and air and water filled porosity at maximum water capacity and field
capacity.
A Delta-T ThetaProbe ML2X (Delta-T Devices, Cambridge, UK) was used to take instantaneous readings of growing
medium VMC. ThetaProbe devices were previously used successfully in extensive green roof research (Monterusso
et al., 2005; VanWoert et al., 2005; Durhman et al., 2006). Accuracy of the ThetaProbe is ±0.01 m3/m3 jn 0 - 40?C.
The sensor is factory calibrated after manufacturing and prior to the sensor being sealed. Accuracy was tested at
least once per month by dipping the probe in a cup of water and getting a reading of 100% VMC. The probe was
inserted into the growing medium to a depth equal to the length of the probe (5 cm) and at least 1 cm from the
edge of the container; VMC readings appeared instantly on the analogue output screen of the attached handheld
meter and readings were written down on a data sheet. While there are standard gravimetric procedures for
determining VMC, there is no standard published method for calibrating the sensor beyond factory. A correction
curve for a particular media could be derived by developing ThetaProbe response to gravimetrically tested VMC;
however, this does not account for individual species plant roots. The advantage of using the probe is the non-
destructive means of testing. As such however, the VMC data presented here are to be considered relative and not
absolute VMC.
For studies 1 (determination of suitability of plant species for extensive green roof use in the semi-arid, high
elevation Front Range of Colorado), 3 (evaluation of trial plant species when grown in mixed stands modified
extensive green roof growing media), and 2 (determination of suitability of zeolite amendments for supporting plant
growth when grown in an extensive green roof system), growing media measurements of VMC were analyzed at
the beginning, middle and end of September in 2008 and 2009. Additionally, VMC was analyzed for concurrent plant
cover measurements for the Zeolite Amendment Study.
Figure 3-3 shows an example of media moisture measurement locations for the two different module sizes. Each
plant studied was centered on a 93 cm2 (1.0 ft2) square represented the module edges and the dotted lines. Seven
total measurements (represented by the black squares) per measurement date were taken in the 61 x 122 cm (2 x 4
ft) modules for studies 1 and 3 and three total measurements (again, represented by the black squares) were taken
in each of the 61 x 61 cm (2 x 2 ft) modules for study 2. For the larger modules, three measurements were taken
down the center of the module and two on each side of the module to get an even distribution of growing media
moisture within the module. Similarly, two measurements were taken down the center and one on the side of the
smaller modules.
3-4
-------
n
•
n
—
• n
n
i
i
i
i
i
i
i
-4-
i
j
1
i
i
i
i
i
H
H •
n
•
n
n T n
i
n » n
Black squares represent sampling locations
Circular symbols represent single plantings
Figure 3-3 Examples of planting and volumetric moisture content sampling locations.
For Study 4 (determination of dry down characteristics of extensive green roof growing media and the impact of dry
down on plant species), growing media VMC was recorded daily for each plant using the ThetaProbe. Values were
collected daily until they remained constant. Relative water use for each species was estimated from VMC data by
subtracting the growing media VMC of the non-vegetated control for each day. For the outdoor study, a rainproof
cover was used during threats of rain.
Analytical Methods
The digital image data was analyzed using Sigma Scan Pro 5.0 image analysis software (SPAA Science, Chicago, IL).
This program measures growth rates by analyzing predetermined ranges of pixel colors on digital images.
Data sets were analyzed using a repeated measures analysis of variance procedure (GLIMMIX) in SAS® version 9.02
(SAS Institute Inc., Gary, NC). The GLIMMIX procedure was performed using t-tests for multiple comparisons of
means to show differences in plant cover and VMC. The DIA data were transformed for analysis to the log scale to
equalize and normalize the residuals; no transformation was performed on the VMC data. Means, standard
deviations, standard errors and correlations between C2D and DIA data sets, as well as between DIA and biomass
and C2D and biomass data sets, were determined in Excel (Microsoft Office Excel, 2007).
Statisca (StatSoft, Inc. 2003, version 6) was used for statistical analysis and graph development in Chapter 7.
Factorial multivariate and univariate analysis of variance (MANOVA, StatSoft, Inc. 2003, version 6) were used to test
for effects of independent variables VMC and total irrigation (rainfall and irrigation) versus categorical variables.
Follow up univariate factorial ANOVAs were used to further probe any significant multivariate effects. Statistical
significance was fixed at p < 0.05.
3-5
-------
Chapter 4 Single Species Study
Extensive green roofs have not been scientifically evaluated in the high elevation, semi-arid climate of Colorado.
Elsewhere in North America, research on species that can succeed on extensive green roofs has revealed that
succulents, predominantly Sedum taxa, out-perform most non-succulents (Monterusso et al., 2005; Rowe et al.,
2006; Durhman et al., 2007). However, the non-succulents tested were typically native to areas with high annual
precipitation and relatively deep soil profiles. Plants native to the Rocky Mountain region, especially those that
inhabit areas with shallow, rocky, well-drained soils, may be suited for use in extensive green roof systems (Getter
and Rowe, 2006). With the exception of D. cooper/, five of the six species used in this study are native to the western
United States in general and Colorado in particular (Table 4-1).
Many plant-related research projects require quantification of plant area covered (plant cover) or, more specifically,
rate of change in plant cover over time. Quantification of plant cover is valuable for studies pertaining to green roof
plantings because plant species that can cover an area quickly are preferred for green roof applications for both
aesthetics and performance (White and Snodgrass, 2003). The use of such species can reduce the cost associated
with denser plantings of species that grow slower and cover less area.
There are several methods for quantifying plant cover and rate of change in plant cover. However, most reported
methods are subjective and not based on quantitative measurements. Typically, visual assessment or visual ratings
are used to evaluate plant cover. Manually measured plant growth indices are frequently used as a measure of plant
performance. Typically, measurements of plant diameters are used to estimate plant cover. The current research
converts two plant diameters into the area of a circle to estimate plant cover (C2D). DIA is another method used for
quantification of plant area which requires periodic photographing of plants and then digitally analyzing the images
to quantify plant cover. DIA can also be used to estimate or validate biomass accumulation in plants.
During 2008 and 2009, two methods of quantifying plant cover were utilized to evaluate the performance of the six
species on an extensive green roof located in a semi-arid, high elevation region. For each of six species in the study
(Table 4-1), approximate plant cover was obtained by manually measuring diameters of each plant and then
converting those diameters into approximate plant cover (C2D). In addition, digital images of these same plants
were taken periodically throughout the growing season; these images were then digitally analyzed to quantify plant
cover (DIA). The DIA data were compared to the C2D data.
The specific objectives of the research for the Single Species Study were to:
1. determine species plant cover via DIA and C2D methods
4-1
-------
2. determine the correlation between the DIA and C2D methods
3. determine the correlation between DIA and plant biomass
4. determine the correlation between C2D and plant biomass.
In this study, a treatment was one of six species (Table 4-1). A series of modules, each containing one of the six
species per module, were placed into one of five blocks. Thus a total of six species, each replicated in five blocks
resulted in 30 modules. In each module, eight individual plants were planted, each with 93 cm2 (1.0 ft2) of growing
space. See Chapters for greater details on the materials and methods and layout of the modules (Figure 3-2).
Table 4-1: Plant Species Evaluated in the Single Species Study
Species
Common
name
Growth
habit
Antennaria
parvifolia
small-leaf
pussytoes
groundcover
Bouteloua
gracilis
blue grama
upright (grass)
Delosperma
cooperi
hardy ice plant
Groundcover
Eriogonum
umbellatum aureum
Kannah Creek®
buckwheat
groundcover
Opuntia fragilis
brittle prickly
pear
decumbent
(cactus)
Sedum lanceolatum
spearleaf stonecrop
Groundcover
Results
Every individual plant of each of the six species (n=240) survived the 2008 growing season. During the 2008-2009
season four of the six species had 100% survival rate. A parvifolia, which had a 65% survival rate for the first
measurement on May 13, 2009, was included in the data analysis. However, E. umbellatum aureum 'Psdowns',
which had only a 12.5% survival rate, was not included in the data analysis. Plant cover is reported in terms of days
from trial initiation; with Day 1 being the day the modules were placed on the green roof (March 26, 2008), and Day
49 being the first date of comparison (May 14, 2008). Table 4-2 contains the days of data collection during the study
and their corresponding calendar dates. Table 4-3 contains weather data during the period of study.
Table 4-2 Days of Study and Corresponding Calendar Dates
Day of study
49
91
133
174
413
455
497
538
Calendar date
May 14, 2008
June 25, 2008
August 6, 2008
September 16, 2008
May 13, 2009
June 24, 2009
August 5, 2009
September 15, 2009
Table 4-3 indicates that there was only 550.6 mm (21.68 in.) of rainfall from July, 2008 through October, 2009.
Average monthly rainfall was 34.5 mm (1.36 in). From November, 2008 through January, 2009, there was less than
10 mm (0.4 in) per month this sustained period amounted to 25.7 mm (1.0 in). Limited precipitation and cold
temperatures may have led to overwintering stress of the plants, especially the herbaceous plants, i.e., A. parvifolia
and E. umbellatum aureum. On January 5, 2009, ambient temperatures dropped below -10 C° and temperatures at
the green roof membrane dropped to -5 C°, implying frozen temperatures existed throughout the root zone. During
growing season, roughly mid-April through mid October, only May through September had non-freezing
temperatures. The green roof plants were supplemented with irrigation; this irrigation system is turned off during
winter to prevent breakage due to freezing. Parts of the irrigation system were damaged in the 2008/2009 winter
leading to replacement of the irrigation system in June, 2009. Rainfall in March, 2009 was half that of ET0, while
4-2
-------
April, 2009 exceeded ET0 (see Table 7-1). Though rainfall rates may not have been sufficient through May, which
was less than VT. ET0, overwintering stress appears to be the main cause for loss of the herbaceous species. As the
first measurement for 2009 was May 13, 2009, an earlier assessment of plant survivability may have more clearly
demonstrated that it was due to over winter stress and not due to lack of rainfall at the start of the growing season
or problems with the irrigation system.
Table 4-3 Weather Data from Region 8 Green Roof
1
oo
o
0
-------
May 13, 2009
June 24, 2009
Augusts, 2009
September 15, 2009
1
Figure 4-1: Example of 2009 recovery as series of the same four D. cooper/ plants on Days 413, 455, 497 and 538.
800 -,
700
£ 600 H
<±
$ 500 -I
o
O
c 400 H
RJ
Q_
300
200 -
100 -
0
. parvifolia
O. fragilis
^^B. gracilis
-CHS. lanceolatum
i i i i r
cooperi
i r
N. N. N
Date
Figure 4-2 Plant cover determined by DIA analysis for the five experimental species over period of study.
4-4
-------
t
•£•
2
-------
800 n
700 -
•A. pan'ffofra —•— B. gracilis —r>-i>. cooperi
-O. fragiiis -O-S. ianceotalum
49 91 133 174 413 455 497
Days After Study Initiation
533
Figure 4-4 Plant cover determined by DIA analysis for the five experimental species for eight measurements.
CM
E
o
0
o
O
"c
^
1800 -|
1600 -
1400 -
1200 -
1000 -
800 -
600 -
400 -
200 -
0 -
•O. fragilis
-+-B. gracilis
-o-S. lanceoiatum
•D. cooper/
49 91 133 174 413 455 497 538
Days After Study Initiation
Figure 4-5: Plant cover determined by C2D analysis for the five experimental species eight measurements.
Comparison between Trials
Recovery of many of the species after winter dormancy yielded irregular regrowth patterns (Figure 4-1 and Figure
4-6). Since C2D data measure plant diameters at the widest points of the plant axes, areas of dieback within those
diameters are included in the analysis, giving an overestimation of actual plant cover. Therefore, most discrepancies
4-6
-------
between the DIA and C2D data sets could be attributed to overestimation of plant cover by the C2D measurements.
Figure 4-6 shows the following examples of irregular growth patterns: a) A. parvifolia (on Day 455) irregular growth
habit after overwintering; b) O. fragilis (on Day 91) after physical damage and c) S. lanceolatum (on Day 455) post-
bloom center dieback. Figure 4-7 shows results of correlation analysis between DIA and C2D.
a %#&&" \fffi'S\%
Figure 4-6: Three examples of regrowth patterns.
•A. parvifolia
•B.gracilis
S. lanceolatum
•D. cooperi
0.3
04/17/08 07/26/08 11/03/08 02/11/09 05/22/09 08/30/09
Date
Figure 4-7 Correlation analysis between DIA and C2D plant cover methods.
4-7
-------
Both A. parvifolia and D. cooper/ showed continued decline on 6/24/09 (Day 455) relative to 5/13/09 (Day 411).
Partly this is overwintering stress, as one would expect increase growth with beginning of growing season. As the
irrigation system was damaged in the winter and replaced mid June, 2009, these declines may also be partly
attributable to decreased efficiency of the existing irrigation system, however, there was sufficient rainfall in April,
2009 and irrigation with the new system was started immediately after completion of the installation.
On 06/25/08, Day 91, each O. fragilis plant in two of the five blocks had pads removed by extension cords that were
dragged over the plants. All of the pads that were removed from the parent cactus were replanted near the parent
plant. The measurements of those plants on Day 91 potentially yielded superficially larger results (69% for DIA and
77% for C2D) than they would have if the entire plant were intact (b in Figure 4-6). This is because the individual
cactus pads could not be placed as closely to the parent plant at replanting as they were while on the plant.
Therefore, a wider set of diameters were recorded after replanting. Rooting and regrowth occurred rapidly but may
have also increased assessment of overwintering success (64% and 44%, respectively) as multiples plants were being
measured, rather than initial plantings. Note also the low correlation value for O. fragilis for Day 91 (6/25/08) in
Figure 4-7.
In 2009, bloom on S. lanceolatum occurred early in the season in three of the five blocks and after the inflorescence
senesced (prior to 6/24/09, Day 455), the center of each plant died out leaving an irregular circular area of green
around the perimeter of the plant (c in Figure 4-6). Therefore the C2D measurements showed the plant to be much
larger than what the DIA quantified, hence the reduced correlation values for 2009 as shown in Figure 4-7.
Biomass Accumulation
Biomass accumulation from harvested plants (Table 4-4) was correlated with the last date of DIA and C2D to
evaluate how well plant cover corresponded with individual plant biomass accumulation (Table 4-5). Correlations
between DIA and biomass, and C2D and biomass, on final date of DIA and C2D data collection (Day 538 [9/15/ 09])
for the five species (n = 40 except A parvifolia where n = 26). In general, correlations between the last date of DIA
and biomass data were high (mean r = 0.83) for the three groundcover plants: A. parvifolia, D. cooper/ and S.
lanceolatum. Bouteloua gracilis, with a more upright growth habit had a lower correlation (r = 0.64) likely because
images taken from directly above would not account for biomass as if taken from the vertical as in Tackenberg
(2007). Correlations for O. fragilis were the lowest among the species in this study (r = 0.41 for DIA and 0.18 for
C2D); this low correlation was attributed to the decumbent growth habit of this species and pads aligned both
vertically and horizontally. Thus, similar to B. gracilis, vertical biomass was not accounted for by either plant cover
analysis (r = 0.64 for DIA and 0.19 for C2D). In general, DIA had higher correlations to biomass than did C2D.
Table 4-4 Biomass Accumulation by Species
Species
A. parvifolia
B. gracilis
D. cooperi
O. fragilis
S. lanceolatum
Wet biomass (g)
39.2
54.0
258.9
522.7
181.1
Dry biomass (g)
14.7
35.3
42.3
148.4
46.0
Water content (%)
61.1
33.1
83.5
72.1
75.1
4-8
-------
Table 4-5 Final Correlations between Plant Cover Analysis Methods and Dry Biomas:
Species
A. parvifolia
B. gracilis
D. cooperi
O. fragilis
S. lanceolatum
DIA correlations (r)
0.79
0.64
0.87
0.41
0.84
C2D correlations (r)
0.54
0.19
0.79
0.18
0.40
Conclusions
In general, all species increased in plant cover during 2008 for both DIA and C2D data sets. However, during 2009,
four of the five species showed temporary declines in plant cover, the exception being O. fragilis. This reduction in
plant cover is likely a result of overwintering stress. A similar phenomenon was observed by Monterusso et al.
(2005) in the growth index graphs for Agastachefoeniculum, Aster laevis, Coreopsis lanceolate/ and several other
species.
On the final date of plant cover comparisons (Day 538 [9/15/09]), the two species with the highest plant cover were
B. gracilis and D. cooperi, with the remaining three species (A. parvifolia, O. fragilis and S. lanceolatum) closely
grouped in plant cover. O. fragilis had the highest biomass accumulation after two seasons. Based on evaluations
over the two consecutive growing seasons, B. gracilis and D. cooperi were more successful than A. parvifolia, O.
fragilis and S. lanceolatum, but all of these species resulted in a net increase in plant cover.
Using DIA to evaluate plant cover and biomass accumulation is especially appropriate for groundcover species
(Bousselotetal., 2010).
4-9
-------
Chapter 5 Mixed Species Study
Introduction
The Mixed Species Study was set up like the Single Species Study (Chapter 4) except eight different species (Table
5-1) were planted together in each of ten 61 x 122 x 10 cm (2 x 4 x 1/3 ft) modules. One of each species was planted
in the modules and plantings were evenly spaced (see Figure 3-3). Five of the modules were planted with the
existing green roof growing media (GreenGrid®) and the other five were planted with a 50% by volume zeolite
(ZeoPro™ H-Plus) mixed in with the existing growing media. Incorporating zeolite as an amendment was intended to
improve the moisture-holding and nutrient-holding capacity of the green roof growing media. One module of each
growing media type was placed in each of the five blocks (see Figure 3-2). Similar data as in the previous study was
collected.
Table 5-1: Plant Species in the Mixed Species Study
Scientific name | Common name
Allium cernuum | nodding onion
Antennaria parvifolia || small-leaf pussytoes
Bouteloua gracilis || blue grama
Delosperma cooperi
Eriogonum umbellatum aureum
Opuntia fragilis
Sedum lanceolatum
Sempervivum 'Royal Ruby '
hardy ice plant
Kannah Creek® buckwheat
brittle pricklypear
lanceleaf stonecrop
hens and chicks, houseleek
Results
All plants of all species in both growing media treatments survived the 2008 growing season with an exception of
one A. cernuum plant lost to bird predation. The overwintering success of all remaining plants is documented in
Table 5-2 and Table 5-3. In general, Table 5-2 indicates winter survival in the 50% zeolite amendment modules was
similar or lower than in the existing growing media modules for each species. Table 5-3 indicates that when
herbaceous plants did survive, plant cover (as measured by C2D) of these species was less in the zeolite amended
modules. Similar to the Single Species Study (study 1), the two species with the lowest overwintering rates were A.
parvifolia and E. umbellatum. Unlike the Single Species Study where B. gracilis had 100% survival, in this study 6.
gracilis showed reduced overwintering survival, just 60% in both media treatments, possibly due to competition
5-1
-------
from the mixed stand of species. Over winter, succulents increased in plant cover, as there is a slight increase in
plant cover for the GreenGrid® media over the 50% zeolite amended media, 44% compared to 36%, though results
varied per species. O. fragilis and Sempervivum 'Royal Ruby' had higher plant cover in the 50% zeolite
amendment; the 50% zeolite amendment reduced (60% compared to 100%) the survivorship of D. cooper/. Though
not as robust, as in the Single Species Study O. fragilis increased in plant cover over the winter (Table 5-3).
Table £
5-2: Overwintering Results for the Mixed Species Study Evaluated on Day 413 (May 1 3, 2009)
Species
Allium cernuum
Antennaria parvifolia
No amendment
100%
40%
50% zeolite amendment
75%
40%
Bouteloua gracilis \ 60% | 60%
Eriogonum umbellatum aureum | 60% | 40%
Herbaceous Mean || 65% || 54%
Delosperma cooperi || 100% || 60%
Opuntia fragilis
Sedum lanceolatum
Sempervivum 'Royal Ruby'
Succulent Mean
100%
100%
100%
100%
100%
100%
100%
90%
Table £
5-3: Change in Percent Plant Cover for the Mixed Species Study from September 1 9, 2008 to May 1 3, 2009
Species || No amendment || 50% zeolite amendment
Allium cernuum \ 25% | - 43%
Antennaria parvifolia | 10% | -63%
Bouteloua gracilis | - 66% | - 75%
Eriogonum umbellatum aureum || 64% || 0%
Herbaceous Mean | 8% | - 45%
Delosperma cooperi
Opuntia fragilis
Sedum lanceolatum
Sempervivum 'Royal Ruby'
Succulent Mean
22%
50%
74%
32%
44%
0%
64%
37%
46%
36%
Plant species cover as measured by C2D is outlined in Figure 5-1 through Figure 5-4. The mean and error bars were
calculated based on the surviving plants (error bars represent standard error). Similar to the Single Species Study
(Chapter 4), the two species that had the highest plant cover by the end of the study were 6. gracilis and D. cooperi.
All other species were much lower in plant cover.
Table 5-4 compares peak plant cover for the amended and non-amended modules. With the exception of A.
parvifolia plant cover increased in the zeolite amended modules over that of the modules with green roof media
during the 2009 growing season. S. lanceolatum which is not presented in Table 5-4 actually declined in plant cover
from 570 cm2 on May 13, 2009 to 260 cm2 on 9/15/09 for the GreenGrid® media, while it increased from 470 cm2 to
710 cm2 with the 50% zeolite amendment.
5-2
-------
Table 5-4: Comparison of peak plant cover in 2009 in the Mixed Species Study
Species
Allium cernuum
Antennaria parvifolia
Bouteloua gracilis
Eriogonum umbellatum aureum
Herbaceous Mean
Delosperma cooperi
Opuntia fragilis
Sempervivum 'Royal Ruby'
Succulent Mean
No amendment
Date
8/5/09
9/15/09
8/5/09
9/15/09
9/15/09
8/5/09
8/5/09
Plant cover (cm2 )
180
320
1300
490
3950
150
1040
50% zeolite amendment
Date
8/5/09
9/15/09
8/5/09
8/5/09
9/15/09
9/15/09
8/5/09
Plant cover (cm2 )
250
190
1900
530
4120
540
1550
Percent
Difference
25%
-40%
34%
7.7%
26%
4.1%
72%
33%
36%
•A. cernuum
•A. parvifolia
•B. gracilis
•E. umbellatum
4000
3500
3000
2500
2000
1500
05/07/08 08/15/08 11/23/08 03/03/09 06/11/09 09/19/09
Figure 5-1 Plant cover for herbaceous plants of mixed species study in existing green roof growing media
5-3
-------
•D. cooper/
•O. fragilis
•S. lanceolatum —•— S. 'Royal Ruby'
4000
1000
500
0
05/07/08 08/15/08 11/23/08 03/03/09 06/11/09 09/19/09
Figure 5-2 Plant cover for succulent plants of mixed species study in existing green roof growing media
0 A cernuum • A. parvifolia A 6. graciiis A £. umbellatum
0 J
05/07/08 08/15/08 11/23/08 03/03/09 06/11/09 09/19/09
Figure 5-3 Plant cover for herbaceous plants of mixed species study in 50% zeolite amended green roof growing media
5-4
-------
•D. cooper/
•O. fragilis
•S. lanceolatum
•S. 'Royal Ruby1
0
05/07/08
08/15/08 11/23/08 03/03/09 06/11/09 09/19/09
Figure 5-4 Plant cover for succulent plants of mixed species study in 50% zeolite amended green roof growing media
Conclusions
In general, species increased in plant cover during 2008 though the 50% zeolite amendment impacted the over
wintering success in some plants. In 2009, typically the plants in the zeolite had greater peak plant cover later in the
season with increases of 26% for the herbaceous plants and over 36% for the succulents. Specifically for the two
herbaceous species, while E. umbellatum aureum had greater survivorship in this study than study 1, overall the
addition of zeolite in this amount does not appear to be beneficial and the 50% zeolite amendment appears to be
detrimental to A. parvifolia, decreasing over winter plant cover and plant cover in general for 2009. For the
succulent species, the 50% zeolite amendment was detrimental to survivability of D. cooper/ and did not improve
overwinter plant cover.
There is disparity in survivorship from one study to another as demonstrated by E. umbellatum aureum which had
greater survivorship in the Mixed Species Study at 60% (no zeolite) compared to 12.5% survival rate for the Single
Species Study, and A. parvifolia, which had 80% survivorship in the Single Species Study but only achieved 40%
survivorship (with and without zeolite amendment) in the Mixed Species Study. This shows that there is much
variability in the response of the plants to the differing environmental conditions. E. umbellatum aureum potentially
benefitted from differing water usage rates of the species used in this study, while during the Single Species Study,
water usage and needs of E. umbellatum aureum may have hindered survivability.
Some plants can reduce stress of neighboring plants and improve survivability of neighboring plants in harsh
habitats. Butler and Orians (2011) observed that Sedum species may reduce water loss from green roof media
5-5
-------
thereby allowing other species to benefit especially during periods of summer water deficit. Cook-Patton and
Bauerle (2012) recommended testing in both single and mixed stands as both survivability (e.g. drought tolerance)
and beneficial function (e.g. evapotranspiration for stormwater management) of individual species may depend on
plant diversity of a green roof.
5-6
-------
Chapter 6. Zeolite Amendment Study
Introduction
For the Zeolite Amendment Study, three percentages of zeolite (ZeoPro™ H-Plus) (33%, 66% or 100%) were
incorporated into a commercially available extensive green roof growing medium (GreenGrid®). Four taxa (three
species, with one species represented by two cultivars) of Sedum already on the EPA Region 8 green roof (S. acre, S.
album, S. spurium 'Dragons Blood' and S. spurium 'John Creech') were planted into each of the zeolite amended
mixtures of growing media and a growing media control to determine which composition is most suitable for plant
growth.
Ten replicates of each media mix were set up in a randomized complete-block design, similar to the Single Species
Study (Chapter 4), but the primary variables was amount of zeolite amendment in media and four taxa. A main
difference between the Single Species Study and the Zeolite Amendment Study is that smaller sized, 61 x 61 x 10 cm
(2 x 2 x 1/3 ft) modules were used. Four planting media (one each of the three percentages of zeolite amended
growing media and one growing media control) made up one module (see Figure 3-1). Each module was randomly
assigned to one of the four positions in each block to minimize environmental variability (see Figure 3-2). Physical
and chemical properties of the blends are outlined in Table 6-1.
Table 6-1 Chemical and Physical Characteristics of the Four Growing Media
Growing media characteristic ||_
Organic natter content by mass (loss on ignition )||_
NO3.Nitrogen (N)* |[
Phosphorus (P) |[
Potassium (K) ||_
Bulk density ||
Particle density ||_
Saturated hydraulic conductivity ||_
Air content ||
Water content ||_
At pF1 = 1.8 (proportion of Air content |[
large pores FLL, 2008) Water content |[
Control
4.9%
105 ppm
19ppm
251 ppm
0.66 g/cc
1.96g/cc
0.0102cm/s
17.7%
48.6%
35.7%
30.6%
J 33% zeolite
!1.8%
197 ppm
!21 ppm
1215 ppm
S0.75 g/cc
2.01 g/cc
! 0.0108 cm/s
13.6%
!48.9%
32.8%
J 29.7%
J 66% zeolite || 100% zeolite
!0.6%
158 ppm
! 26 ppm
1456 ppm
S0.90 g/cc
2.26 g/cc
! 0.0101 cm/s
14.9%
!45.1%
32.3%
0.3%
21 ppm
14 ppm
1597 ppm
0.97 g/cc
2.35 g/cc
0.0154 cm/s
26.8%
32.0%
39.4%
J 27.7% || 19.5%
pF is the logarithmic value (base 10) of the water column in cm; soil moisture measurement to define soil suction. At pF = 1.8 is
equal to field capacity on the green roof substrate moisture retention curve.
6-1
-------
Nitrogen was analyzed as nitrate (NO3.N) but the zeolite contains nitrogen as ammonium ion-N and therefore the
100% treatment showed very little nitrogen content. Nitrogen content increased in the 33% and 66%, likely because
the form of nitrogen in the zeolite changed with mixture of organic matter, i.e., changed some ammonium ion-N to
into NO3.N. All plants were fertilized at initiation of the study; however, the zeolite treatments had higher nutrient
levels, especially K, than the treatment with no zeolite as indicated by the 1597 ppm K in the sample of pure zeolite
(Table 6-1).
Results
All four Sedum taxa responded to the addition of zeolite, however, responses varied in growing season,
overwintering or mixtures of zeolite (Figure 6-1, error bars represent standard error). For example, by the end of
2008, S. acre had the highest plant cover in the 33% and 66% amendments and was lower in control and 100%
zeolite, while S. album increased in plant cover with increasing zeolite content of the growing media.
2000
1800 •
1600
1400 -
1200 •
1000
800 -
600
400 •
200 -
0
2000 -,
1800 -
1600
1400
1200 -
1000 -
800 -
600 -
400 -
200 -
S. acre
—B— control
- • - 33% zeolite
»»©-••• 66% zeolite
—•-- 100% zeolite
2000
1800 -
1600
1400 -
1200 -
1000
800 -
600
400 -
200 -
0
S. album
,-'.'.$.'•,
91
133 174 413 455 497
Days after study initiation
538
91
133 174 413 455
Days after study initiation
497 538
S. spurium 'Dragon's Blood'
S. spurium 'John Creech'
49 91 133 174 413 455
Days after study initiation
497 538
91
133 174 413 455
Days after study initiation
497 538
Figure 6-1: Plant cover as determined by DIA over eight dates during two growing seasons.
However, both S. acre and S. album overwintered poorly as few individual plants survived; this was more
pronounced as zeolite content of the growing media increased (Table 6-2). While winter survival as a percentage
was higher in the treatment with no zeolite than the treatments with zeolite, the plants that did survive had very
low plant cover (Figure 6-1) and were small. This is consistent with research that showed plants that were not
fertilized were smaller in size but survived over the winter compared to those that were fertilized (Rowe et al.,
2006).
6-2
-------
Table 6-2: Overwinter Survival for Each Sedum Taxa for Controls and Zeolite Amendments as Determined on May 13, 2009
Taxa
Sedum acre
Sedum album
Sedum spurium 'Dragon's Blood'
Sedum spurium 'John Creech'
Control
80%
90%
100%
100%
3 3% zeolite
40%
90%
100%
100%
66% zeolite
10%
50%
100%
100%
100% zeolite
0%
10%
100%
100%
Researchers in Michigan have noted good overwintering success for these two species of Sedum, even in some cases
noting the dominance of these two species specifically (Durhman et al., 2004; Monterusso et al., 2005; Durhman et
al., 2007). Minimum ambient air temperatures below freezing were recorded between October 2008 and April, 2009
with the lowest temperature of-21.6 C° occurring in December, 2008. Due to the contrasting results, apparently
there are enough climactic differences between regions to influence survivability of these Sedums. However,
minimum temperature alone may not be the only problem as another Sedum, S. spectabile did not survive
temperatures of -3.0°C in September but, depending on the cultivar, can survive conditions at less than -20°C in
January (lies and Agnew, 1995).
There are many possible factors which could have affected survivability of these green roof plants in these different
growing media blends, especially during the winter season. The minimal precipitation during the winter along with
the water holding capacity of the zeolite may have contributed to the desiccation of these two species. Winter VMC
and diurnal temperature fluctuation related to media color and albedo may have also influence plant survival. Figure
6-2 shows a block of media treatments clearly showing lighter color of zeolite in relation to GreenGrid® medium.
The mean daily minimum temperature of the GreenGrid® growing media during the winter months (December,
2008 through March, 2009) was -3.0°C. The minimum recorded temperature on the surface of the media was -
18.76°C which occurred on December 15, 2008 at 7:50 AM while the minimum temperature on the membrane
under the substrate of the module was -13.09°C coming two hours later at 9:50 AM.
a) media
b) 33% zeolite
c) 66% zeolite
d) 100% zeolite
Figure 6-2 Example block on July 1, 2008 showing four media.
6-3
-------
Durham et al. (2007) noted that snow cover protects the shoots and buds of alpine plants against water loss during
winter. Lack of snow cover in dry areas can desiccate plants (Savonen, 2012). Plants subject to desiccation may leaf
out but die later. Watering or irrigating during winter may prevent desiccation (Savonen, 2012); there was only
approximately 10 mm of rain per month from November 2008 through February 2009 (Table 4-3). Additionally, the
root hardiness of these species is unknown in this type of shallow, well-drained system; while it has not been
formally documented, root size in relation to top growth for some of these species, i.e., S. acre and S. album, has
been found to be noticeably less in higher nutrient and moisture content situations compared to drier and lower
fertility growing media. Additionally, the two S. spurium taxa which were native to Colorado apparently were
competitively better the non-native Sedum species at obtaining water resources during the winter.
The two S. spurium taxa ('Dragon's Blood and 'John Creech') showed much different results than S. acre and S.
album. At the end of the 2008 growing season, all treatments for both of the S. spurium cultivars had similar plant
cover. Although overwintering survival was 100% for all amendments and the controls for both S. spurium cultivars
(Table 6-2), plants in the 100% zeolite were reduced in size at the beginning of the second season (note the
decrease in plant cover on Day 413 in Figure 6-1), which is clearly an effect of overwintering and potential
desiccation. This was especially noteworthy for the 100% zeolite treatment for both S. spurium cultivars.
Conclusions
The survivability of some of the plants in the 100% zeolite indicates that other forms of green roof media may be
utilized. However, interaction effects with other plants remain to be tested. As in the Mixed Species Study, some
plant species perform poorly in zeolite. Competition for water resources also impacts survivability. Whether S. acre
and S. album survivability could be increased if planted with species other than S. spurium cultivars remains to be
tested. As the S. acre and S. album decreased in plant cover even in the control, the introduction of a limited
irrigation regime during winters of low rainfall and limited snow cover may also improve survivability of these
species in these environmental conditions.
6-4
-------
Chapter 7 Analysis of Volumetric Moisture Content
Calculated Evapotranspiration Rates
The evapotranspiration (ET) rates were calculated from monthly data using the Penman-Monteith equation and
guidelines of the Food and Agriculture Organization of the United Nations (FAO) (Allen et al., 1998). As per FAO
guidelines, the monthly reference evapotranspiration rate (ET0) was calculated for the 15th day of the month. The
monthly value presented in Table 7-1 was derived by multiplying the representative 15th day of the month by the
number of days in the month. Values of ET for specific crops can be derived by multiplying a crop coefficient (Kc) to
ET0 to derive individual evapotranspiration rates per crop (ETC). For many crop and forage plants listed in the FAO
guidelines (Allen et al., 1998), Kc > 1; however, one of the plants listed in the FAO guidelines that has a Kc < lis the
pineapple.
The pineapple, which has its stomata closed during the day, has a Kc of 0.5 for most conditions though this can be as
low as 0.3 for mature crop in bare soil (Allen et al., 1998). Crassulacean acid metabolism (CAM) is an adaption by
plants to arid conditions to close the stomata during the day while opening their stomata at night, taking up carbon
dioxide and storing it as malic acid for photosynthesis (Ting, 1985). Because CAM plants open their stomata at night
and close them during the day to minimize water loss, these plants have very high water use efficiency, therefore Kc
for CAM plants is expected to be < 1. Similarly many Sedums planted on green roofs exhibit CAM particularly when
water stressed. The pinapple, of the Bromeliaceae family, the Sedums of the Crassulaceae family and O. fragilis of
the Cactaceae family have all been previously identified as CAM plants (Sayed, 2001). Therefore, Yi ET0 is also
presented in Table 7-1 as potential ETC value for CAM plants used on green roofs.
Smeal et al. (2010) developed a plant or landscape coefficent for western landscape xeriscape plants which can be
applied to ET0 calculated from the FAO Penman-Monteith equation. This was for a range of plants tested in New
Mexico under varying irrigation regime. The mean landscape coefficient measured by Smeal et al. (2010) was 0.3.
This value is applied to a specific equation that looks at ET since last day of irrigation and includes a more complete
from of analysis (e.g., plan canopy area) (Smeal et al., 2010) However, this has been nominally applied to the
normalized monthly ET0 values as a minimal reference value in Table 7-1. Additionally, irrigation and rainfall rates
are presented, along with a qualitative indicator whether irrigation rates and rainfall totals were greater than ET0.
7-1
-------
Table 7-1 Calculated Evapotranspiration Rates and Irrigation and Rainfall Totals
^H
C3
OO
o
0
ts
OS
o
0
ts
Month
May
Jun
Jul
Aug
Sep
Oct
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Stephan
Boltzma
n
(MJ/m2-
d)
—
37.6
40.2
38.5
35.8
33.3
31.3
31.7
35.2
36.8
38.4
38.4
35.8
33.3
Vapor
pressure
deficit
(kPa)
~
2.15
2.99
2.02
1.46
1.08
0.96
0.74
1.32
1.56
1.91
2.13
1.46
1.08
Evapotranspiration (mm)
Calculated
monthly
reference
(ET0)
-
183
196
148
106
75.6
99.5
91.4
139
145
162
154
113
59.4
!/2 monthly
reference
(0.5xET0)
-
91.6
97.9
74.1
53.0
37.8
49.7
45.7
69.3
72.3
81.0
77.0
56.3
29.7
Monthly
xeriscape
reference
(0.3xET0)
—
55.0
58.7
44.4
31.8
22.7
29.8
27.4
41.6
43.4
48.6
46.2
33.8
17.8
Monthly
rainfall
(mm)
64.33
16.83
4.3
55.6
28.2
17.5
44.2
108.2
53.6
55.4
60.2
20.8
16.0
45.2
Monthly
irrigation
rate2
(mm)
160.5
97.6
114.1
87.6
68.4
40.8
3.1
3.9
4.8
13.9
56.6
82.1
66.0
28.2.
Total
irrigation
and
rainfall
224.5
114.4
118.4
143.2
96.6
58.3
47.3
112.1
58.4
69.3
116.8
102.9
82.0
73.7
Qualitative
monthly
deficit1
«4
<
<
<
<
<
<>
«
<>
<>
<
<
<
«
Normalized to monthly total based on meter readings.
2 « = exceeds monthly ET0; < = between monthly monthly ET0, and l/i monthly ET0; <> = between
monthly ETC, for xeriscaping; > below monthly ET0, for xeriscaping (not observed); ~ = no data.
3 National Weather Service station (ID: 052223) at Denver Water (1600 W. 12th Avenue, Denver, CO)
from Region 8 green roof.
4 Based on May 2009 ET0.
!/2 monthly
collected 2.
ET0, and
.6 km away
Table 7-1 indicates that irrigation rates applied to the roof were generally less than that of the reference ET0 rates
with the potential exception of May, 2008. As May, 2008 was the start of the study, i.e. plants put in position on the
roof, irrigation was increased for success of these transplants. Otherwise, when the combination of irrigation and
rainfall exceeded ET0 for the month, it was due to large rainfall totals rather than excessive irrigation. The
combination of irrigation and rainfall never fell below the 0.3 x ET0 rate for xeriscaping. This would be considered
the minimal irrigation requirement, while the target for green roof irrigation would be between 0.3 x ET0 and Yi x
ET0 as many of these plants exhibit CAM. However, herbaceous plants that do not exhibit CAM may have ET rates
closer to the ET0. These plants may require additional irrigation, though as previously noted mixed plantings of
Sedums and other species may allow the other species to benefit as the Sedums reduce water loss from green roof
media (Butler and Orians, 2011). Any applied irrigation should also take into account actual rainfall. As data and ET0
have been normalized to monthly values, irrigation should also take into account periods without rainfall and other
intra monthly variation.
From June through October, 2008, ET0 was 710 mm (28 in.) while from June to October, 2009 ET0 was 630 mm (25
in.). For the same period, the total irrigation and rainfall was 530.9 (20.9 in.) for 2008 and 444.7 (17.5 in.) for 2009.
The combined irrigation and rainfall total exceed the Yi ET0 for 2008, 350 mm (14 in.), and 2009, 320 mm (13 in.),
respectively.
7-2
-------
Relative Volumetric Moisture Content Analysis
The cumulative daily rainfall and cumulative daily irrigation and rainfall total for September 2008 and 2009 are
shown in Figure 7-1. Cumulative irrigation rates were assumed to smooth the graph as actual irrigation was applied
early in the morning only on Mondays and Thursdays.
Cummulative rain 2008 (in.)
•Cummulative rain 2009 (in.)
Cummulative Rain and Irrigation 2008 (mm)
•Cummulative Rain and Irrigation 2009 (mm)
120
100
£
c 80
o
4->
'E GO
1
15 40
H-
1
20
10 15 20 25
Day of month
30
Figure 7-1 Irrigation and rainfall totals for September.
The VMC was measured for studies 1, 2 and 3 on the EPA Region 8 green roof during the month of September in
both 2008 and 2009. Delta-TTheta Probe ML2X (Delta-T Devices, Cambridge, UK) were used to take instantaneous
readings of growing media VMC. As zeolite is reputed to have good micro-pore space available for holding water, at
least compared to other extensive green roof growing media materials, moisture holding capacity of the growing
media was expected to increase with zeolite content in the mix. The ability to measure the exact water content due
to the micropore structure, in the order of 10"10 m diameter, with macroscopic physical probes may be limited.
Measurements were made on three separate dates, representing the beginning, middle and end of the month.
Additional measurements were made concurrent to plant cover data collection for study 2. Figure 3-3 describes
locations of VMC data collection.
Figure 7-2 and Figure 7-3 show average VMC measurements for study 1 for three dates in September for 2008 and
2009. There is a higher VMC around the middle of the month for 2008 (Figure 7-2), but there was a large rainfall
event, 17 mm (0.67 in.) on September 12, 2008 (Figure 7-1). In 2009, there were several small rain fall events and
7-3
-------
the VMC measurements appear more consistent. Besides rainfall patterns, the major difference between 2008 and
2009 is the change in irrigation systems from drip to overhead rotary (spray).
~A. parvifolia • B. gmcilis & D. cooperi A O. fragi/is D 5. lanceolatum 2009
08/30/08 09/04/08 09/09/08 09/14/08 09/19/08 09/24/08 09/29/08 10/04/08
Date
Figure 7-2 Average volumetric moisture content for each species for three dates in 2008.
7-4
-------
-A. parvifolia
•B. gracil/s
•D. coo peri A O. fragilis
•5. lanceo/atum
16
14
I12
.1 10
•4-*
re
on
73
ra
4—
•I
0
8/30/09 9/4/09 9/9/09 9/14/09 9/19/09 9/24/09 9/29/09 10/4/09
Date
Figure 7-3 Average volumetric moisture content for each species for three dates in 2009.
Results of the factorial MANOVA/ANOVA for the Single Species Study (study 1) are summarized in Table 7-2.
Average VMC was used rather than individual data points per module and only four species were analyzed as A.
parvifolia had missing data and E. umbellatum aureum had low survivorship as discussed previously.
Table 7-2 Factorial Analysis of Volumetric Moisture Content for Study 1
Categorical Factors
Species
Block
Type of Irrigation
Species - Block
Species - Type of Irrigation
Block - Type of Irrigation
Species - Block - Type of Irrigation
Multivariate Analysis (VMC and
cumulative irrigation and rainfall)
F value
6.94
2.90
3.57
1.27
2.04
0.92
0.69
Rvalue
0.000001
0.005
0.033
0.19
0.063
0.50
0.85
Univariate Analysis (VMC)
F value
11.7
4.6258
6.9661
2.0696
3.1524
1.4027
1.0844
P value
0.000002
0.002
0.010
0.028
0.029
0.24
0.38
Figure 7-4, Figure 7-5 and Figure 7-6 show the graphical results for species, block and type of irrigation respectively
for the multivariate analysis. Figure 7-7 and Figure 7-8 shows the interaction effects between block and species, and
species and type of irrigation.
7-5
-------
14
5? 13
B 12
11
Species; LS Means
Wilks lambda=.62642, F(6, 158)^6.9382, p=.00000
Effective hypothesis decomposition
Vertical bars denote 0.95 confidence intervals
1 9
o
B. gracilis
D.cooperi O.fragilis
Species
S. lanceolatum
Figure 7-4 Box and whisker plot of effect of species on volumetric moisture content for study 1.
Block; LS Means
Wilks lambda=.76056, F(8,158)=2.8965, p=.00487
Effective hypothesis decomposition
Vertical bars denote 0.95 confidence intervals
13
£
8
P. 11
O ! !
2
'o
o 9
=3
0 Q
-------
Type I; LS Means
Wilks lambda=.91710, F(2, 79)=3.5706, p=.03277
Effective hypothesis decomposition
Vertical bars denote 0.95 confidence intervals
£
^
£Z
o
O
1
o
•=;
d>
3
o
>
a>
ni
JU
(D
5
Q ^
Q n
p n
7 ^
7 n
ft ^
<
Drip
Spray
Type of Irrigation
Figure 7-6 Box and whisker plot of effect of block on volumetric moisture content for study 1.
20
18
16
8 14
10
0)
^- 4
(Jj T1
2
0
Species*Block; LS Means
Current effect: F(12, 80)=2.0696, p=.02831
Effective hypothesis decomposition
Vertical bars denote 0.95 confidence intervals
Block 1 Block 2 Blocks Block 4
Block
Block 5
~9~ Species
B.gracilis
~D~ Species
D. cooperi
~5~ Species
0. fragilis
~$~ Species
S. lanceolatum
Figure 7-7 Box and whisker plot of effect of block and species on volumetric moisture content for study 1.
7-7
-------
Species'Type I; LS Means
Current effect: F(3, 80)=3.1524, p=.02937
Effective hypothesis decomposition
Vertical bars denote 0.95 confidence intervals
16
14
o -19
0 IZ
s
~?~ Drip irrigation
~g~ Spray irrigation
B. gracilis
D. cooperi
O.fragilis
S. lanceolatum
Species
Figure 7-8 Box and whisker plot of effect of species and irrigation type on volumetric moisture content for study 1.
O. fragilis, a cactus, consistently has the highest VMC; this is clear from multivariate analysis (Figure 7-4) and
univariate analysis (both Figure 7-7 and Figure 7-8). Block 5 had the highest VMC (Figure 7-5) which is a possible
indication of shading in the late afternoon by the upper floors of the building, as block 5 was located on the east side
of the building (Figure 3-2). Irrigation was always applied early in the morning. Block 1, which was on the west side
of the building, would have had the lowest VMC but O. fragilis had its highest VMC for block 1 (Figure 7-7). While
Figure 7-6 indicates there is a difference in VMC due to the change in irrigation systems, Figure 7-8 might indicate
this difference is mostly due to the reduction of VMC measured in the B. gracilis modules. B. gracilis had the second
lowest biomass at the end of the Single Species Study and lowest measure of water content (Table 4-4). This may
imply that overhead rotary irrigation is appropriate for Sedums but may not be appropriate for herbaceous species
planted in a single stand. As VMC for O. fragilis were consistently higher than other species, it would appear that
irrigation rates were more than this species need. O. fragilis had the highest biomass (Table 4-4 ) and similar water
content to the Sedums; there was no difference in VMC for O. fragilis (Figure 7-8).
The plant loss of A. parvifolia corresponded to block 1 which had the lowest VMC reading (except for O. fragilis). A
multivariate analysis for cumulative irrigation and rainfall and VMC for the same three categorical variables with A.
parvifolia and without blocks 1 and 5 yielded only species as significant (F = 2.28, p = 0.026), with O.fragilis having
highest VMC while other species were similar. A univariate analysis of VMC indicated effects for species (F = 3.90, p
= 0.0069) and type of irrigation (F = 4.63, p = 0.035).
Results of the factorial MANOVA/ANOVA for the Mixed Species Study (study 3) are summarized in Table 7-3. There
are significant effects for zeolite, which is shown graphically in Figure 7-9, while the interaction effect of zeolite -
type of irrigation nearly registered significance for the univariate analysis and this is shown graphically in Figure
7-10. There is no effect based on block location for the Mixed Species Study (study 3). This may be due to varying
water usage rates of mixed plants in the modules versus single plant species (studyl) where effect was observable,
i.e., in the Single Species Study (study 1), all plants in a module would assumably use water at the same rate.
7-8
-------
Table 7-3 Factorial Analysis of Volumetric Moisture Content for Study 3
Categorical Factors
Zeolite
Block
Type of Irrigation
Zeolite - Block
Zeolite - Type of Irrigation
Block - Type of Irrigation
Zeolite - Block - Type of Irrigation
Multivariate Analysis (VMC and
cumulative irrigation and rainfall)
F value
8.35
0.858
1.27
0.99
2.28
0.98
1.09
Rvalue
0.00096
0.55
0.29
0.45
0.12
0.46
0.38
Univariate Analysis (VMC)
F value
14.4
1.55
0.168
1.79
3.94
1.77
1.98
P value
0.00049
0.21
0.69
0.15
0.054
0.15
0.12
Figure 7-9 shows that the 50% zeolite amendment decreased measured VMC. Figure 7-10 shows that drip irrigation
had higher measured VMC for no amendment while overhead rotary irrigation had higher VMC for the 50%
amendment. As noted earlier in Chapter 5, the zeolite 50% amendment had a slight adverse effect on the
overwintering success of plants (Table 5-2 and Table 5-3) as measured on May 13, 2009, while there was a mean
increase in plant cover of 26% for herbaceous plants and >36% increase for succulents (Table 5-4) for plants in the
50% zeolite mixture over those in GreenGrid® medium. Block 1 had the lowest survivability in study 3 with a loss of
7 plants (4 in 50% mixture). This potentially implies that zeolite contributed to desiccation of the root zone.
Zeolite; LS Means
Wilks lambda=.70026, F(2, 39)=8 3469, p= 00096
Effective hypothesis decomposition
Vertical bars denote 0.95 confidence intervals
12
11
10
None Half
Zeolite Amendment
Figure 7-9 Box and whisker plot of effect of zeolite on volumetric moisture content for study 3.
7-9
-------
Type of IrrigatiorfZeolite; LS Means
Current effect: F(1,40)=3.9359, p=.05416
Effective hypothesis decomposition
Vertical bars denote 0.95 confidence intervals
,-s
^
i*t
13
12
11
10
9
8
7
6
5
4
^
-|-
[
— '—
] ~
[
£
]
,
~$~ Type of Irrigation
Drip
None Half ^ Type of Irrigation
Spray
Zeolite Amendment
Figure 7-10 Box and whisker plot of effect of bock on volumetric moisture content for study 3.
Results of the factorial MANOVA/ANOVA for the Zeolite Amendment Study (study 2) are summarized in Table 7-4.
The only significant effect for the multivariable analysis is the type of irrigation, shown in Figure 7-11. As in the
Mixed Species Study (study 3), there is not an observation of statistical difference due to block or location on the
roof; this study (2) had different plant species planted in individual plant modules as did the Mixed Species Study
(study 3). For the univariate analysis, type irrigation once again has a strong effect, while interaction effects of
zeolite amendment -block, Figure 7-12, and zeolite amendment - type of irrigation, Figure 7-13 were statistically
significant.
Table 7-4 Factorial Analysis of Volumetric Moisture Content for Study 2
Categorical Factors
Zeolite
Block
Type of Irrigation
Zeolite - Block
Zeolite - Type of Irrigation
Block - Type of Irrigation
Zeolite - Block - Type of Irrigation
Multivariate Analysis (VMC and
cumulative irrigation and rainfall)
F value
1.08
0.87
21.0
1.46
1.69
1.03
0.56
Rvalue
0.38
0.54
0.000000
0.090
0.13
0.42
0.95
Univariate Analysis (VMC)
F value
2.19
1.78
41.6
3.21
3.47
2.1
1.17
P value
0.000000
0.14
0.000000
0.00086
0.020
0.089
0.32
7-10
-------
11.0
10.5
Q 9.0
B
7-5
o
> 7.0
6.5
6.0
Type of Irrigation; LS Means
Wilks lambda=.65318, F(2, 79)=20.973, p=.00000
Effective hypothesis decomposition
Vertical bars denote 0.95 confidence intervals
Drip
Spray
Type of Irrigation
Figure 7-11 Box and whisker plot of effect of type of irrigation on volumetric moisture content for study 3.
16
15 -
14 -
13
12
10
9
8
7
5
4
3
Zeolite Amendment*Block; LS Means
Current effect: F(12,80)=3.2120,p=.00086
Effective hypothesis decomposition
Vertical bars denote 0.95 confidence intervals
No arnrnendrnent
0.33 0.66
Zeolite Amendment
Block 1
Block 2
Blocks
Block 4
'Blocks
Figure 7-12 Box and whisker plot of effect of block and amendment on volumetric moisture content for study 2.
7-11
-------
13
12
g 11 •
| 10 •
o
y
£ 9
3
LO
U
' —
CD /
E
"o e-
5 r
4
Zeolite Amendment*Type of Irrigation; LS Means
Current effect: F(3, 80)^3.4698, p=.01992
Effective hypothesis decomposition
Vertical bars denote 0.95 confidence intervals
No ammendment
0.33 0.66
Zeolite Amendment
Type of Irrigation
Drip
Type of Irrigation
Spray
Figure 7-1 3 Box and whisker plot of effect of amendment and irrigation type on volumetric moisture content for study 2.
Figure 7-12 shows that for the control, GreenGrid® medium, block 5 has the highest VMC; this is similar to
interaction effects of Figure 7-7 for the Single Species Study which shows increasing VMC of block 5. Figure 7-13
shows that for the drip irrigation, i.e., the first year of the study in 2008, the trend is that the least amount of
moisture was present in the 0% zeolite treatment and the highest was in the 100% zeolite treatment, which is
inconsistent with the data provided in Table 6-1. As noted earlier the albedo of the zeolite amendment is noticeable
(see Figure 6-2) and this may play a role in increasing VMC by lowering temperature of the media. Also, a slight crust
tended to form on the surface of substrates containing zeolite, potentially further reducing evaporation.
Figure 7-13 also shows that with overhead rotary irrigation, measured VMC is the same with the control, which is
consistent with Figure 7-10, but there is reduced VMC in the zeolite amended modules, which is not consistent with
Figure 7-10. The different plant mixtures, zeolite mixtures and varying survivability of plants may have played a role
in these differences. Figure 7-13 shows that there is once again an increasing trend of VMC with increasing zeolite
content in 2009.
Studies 1, 2 and 3 had significant effects on VMC due to type of irrigation. VMC decreased with overhead rotary
irrigation; however, year to year, there was a total 14.6 mm (0.57 in.) less rainfall (12.2 mm, 0.48 in.) and irrigation
(2.4 mm, 0.09 in.) in September, 2009 than in September, 2008 (Table 7-1). As shown in Figure 7-2, the VMC in
September, 2008 increased in the middle of the month due to rainfall (Figure 7-1), while in Figure 7-3, the response
to rainfall in September, 2009 is minimal as measured VMC remains more constant throughout the month.
The drip irrigation system was used in 2008 and an overhead rotary system was used after June, 2009. There were
additional VMC data collections for the Zeolite Amendment Study (study 2). Figure 7-14 shows there is much more
variability in measured VMC in 2008, particularly 6/25/08 and 8/8/08 date, when the drip irrigation was in use, than
7-12
-------
in 2009, particularly 8/19/09, when overhead rotary irrigation was in use. Additionally, in August and September of
2008, there was significantly greater rainfall (twice as much) than in August and September of 2009. There was also a
noticeable increase in plant cover (Figure 4-2 through 4-5) in August, 2008 for study 1, which had much higher
rainfall than July, 2008; July, 2008 was predominantly dependent on irrigation (>90%, Table 7-1).
20
18
16
0.05) were similar, as type of irrigation is the dominant effect,
7-13
-------
while interaction effects of zeolite - type of irrigation had an effect in the comparison of September 2008 to
September 2009 (Table 7-4), it did not an effect in the longer term portion of the study, i.e., May 2008 through
September 2009 (Table 7-5). Interaction results of block - type of irrigation was significant in Table 7-5; block 1 had
the highest VMC for the drip irrigation and lowest VMC for overhead rotary irrigation (Figure 7-15). In study 1, which
did not use zeolite amendment, block 5 had the highest VMC regardless of irrigation system (Figure 7-5). Figure 7-15
implies the overhead rotary irrigation is less variable as VMC ranges from about 6 to 10 while the means for drip
irrigation range from 9 to 16.
o
U
-------
18
16
E 14
Oi
o
12
10
Block*Zeolite Amendment; LS Means
Current effect: F(12, 240)=2.0055, p=.02447
Effective hypothesis decomposition
Vertical bars denote 0.95 confidence intervals
g _^
Block 1
Block 2
Block 3
Block
Block 4
Blocks
Zeolite Amendment
none
P Zeolite Amendment
_ 0.33
4- Zeolite Amendment
0.66
Zeolite Amendment
1
Figure 7-16 Box and whisker plot of effect of amendment and block on volumetric moisture content for study 2.
Comparison of Volumetric Moisture Content and Plant Cover
The MANOVA/ANOVA multivariate analysis in Table 7-6 compared DIA and VMC analysis performed on the same
dates (6/25/2008, 8/6/2008, 9/16/2008, 8/19/2009 and 9/15/2009). Results indicate that all interaction effects
involving plant and block or type were not significant. While there are several significant interaction effects, most
graphical presentation are similar to many of the previous graphs. Figure 7-17 through Figure 7-20 present the
interaction effects between plants or block with type of irrigation and zeolite amendment.
7-15
-------
Table 7-6 Factorial Analysis of Volumetric Moisture Content and Plant Cover for Study 2
Categorical Factors
Plant
Type of Irrigation
Block
Zeolite
Plant - Type of irrigation
Plant -Block
Type of Irrigation - Block
Plant -Zeolite
Type of Irrigation - Zeolite
Block -Zeolite
Plant - Type of Irrigation - Block
Plant - Type of Irrigation - Zeolite
Plant -Block -Zeolite
Type of Irrigation - Block - Zeolite
Plant - Type of Irrigation - Block - Zeolite
Multivariate Analysis
F value
212.068
237.413
6.085
23.463
217.995
0.793
10.750
7.010
11.213
4.114
0.680
8.067
0.848
2.012
0.660
P value
0.000
0.000
0.000
0.000
0.000
0.75
0.000
0.000
0.000
0.000
0.88
0.000
0.81
0.0026
0.99
PlantType of lrrigation*Zeolite Amendment; LS Means S. acre
Wilks lambda=.84016, F(18, 1596)=8.0674, p=0.0000 -31 S. album
Effective hypothesis decomposition -HE Dragons Blood
Vertical bars denote 0.95 confidence intervals 3i John Creech
Drip Spray
Drip Spray
Drip Spray
Drip Spray
Zeolite Amendment: 0% Zeolite Amendment: 33% Zeolite Amendment: 66% Zeolite Amendment: 100%
Figure 7-17 Box and whisker plot of effect of plant, irrigation type and zeolite amendment on volumetric moisture content.
7-16
-------
2000
1800
1600
1400
1200
1000
800
600
400
200
0
-200
-400
PlantType of lrrigation*Zeolite Amendment; LS Means 5E S. acre
Wilks lambda=.84016, F(18.1596)^8.0674. p=0.0000 3E S. album
Effective hypothesis decomposition 3E Dragons Blood
Vertical bars denote 0.95 confidence intervals 3E John Creech
i
I
Drip Spray
Drip Spray
Drip Spray
Drip
Spray
Zeolite Amendment: 0% Zeolite Amendment: 33% Zeolite Amendment: 66% Zeolite Amendment: 100%
Figure 7-18 Box and whisker plot of effect of plant, irrigation type and zeolite amendment on plant area.
Type of Irrigation'Block'Zeolite Amendment; LS Means
Wilkslambda=.94213.F(24, 1596)=2.0121, p=.00260
Effective hypothesis decomposition
Vertical bars denote 0.95 confidence intervals
: Drip Irrigation
: Spray Irrigation
20
18
16
!«
(D
3 12
Si
1 10
o
S 8
O)
I 6
O
4
2
Block! Block3 Block 5 Block 1 Blocks Block 5 Block 1 Block 3 Block 5 Block 1 Block 3 Blocks
Block 2 Block 4 Block 2 Block 4 Block 2 Block 4 Block 2 Block 4
Zeolite Amendment: 0% Zeolite Amendment: 33% Zeolite Amendment: 66% Zeolite Amendment: 100%
Figure 7-19 Box and whisker plot of effect of block, irrigation type and zeolite amendment on volumetric moisture content.
7-17
-------
1000
900
800
700
"I" GOD
O3
S 500
<
| 400
300
200
100
0
Type of lrrigation*Block*Zeolite Amendment; LS Means
Wilkslambda=.94213,F(24.1596)=2.0121.p= .00260
Effective hypothesis decomposition
Vertical bars denote 0.95 confidence intervals
•
3E Drip Irrigation
32 Spray Irrigation
Block 1 Block3 Block 5 Block 1 Block 3 Block 5 Block 1 Blocks Block 5 Block 1 Block 3 Blocks
Block 2 Block 4 Block 2 Block 4 Block 2 Block 4 Block 2 Block 4
Zeolite Amendment: 0% Zeolite Amendment: 33% Zeolite Amendment: 66% Zeolite Amendment: 100%
Figure 7-20 Box and whisker plot of effect of block, irrigation type and zeolite amendment on volumetric moisture content.
Figure 7-17 shows there is a less observed VMC for the overhead spray irrigation for all levels of zeolite amendment,
while Figure 7-18 shows there is increased plant cover for S. spurium cultivars 'Dragons Blood' and 'John Creech'.
Particularly of interest is that while the VMC drops for the 33% and 66% zeolite amendment, there is
correspondingly higher plant cover for S. spurium cultivars 'Dragons Blood' and 'John Creech'. Figure 7-19 shows
there is less VMC for the 33% through 100% zeolite amendment for the overhead rotary irrigation across all five
blocks. Figure 7-20 shows there is increased plant cover in all cases except block 1, 100% zeolite amendment, with
much higher plant cover for 33% and 66% zeolite amendment for the overhead rotary irrigation across the five
blocks. This may help explain some of the disparity in comparing reduced VMC observed for Zeolite Amendment
Study (study 2) in Figure 7-13 compared to increased VMC for 50% zeolite amendment for Mixed Species Study
(study 3).
The implication is that overhead rotary irrigation overall increased plant cover even though there were reductions in
observed VMC. The original drip irrigation system was fitted with emitters spaced roughly 30 cm (1 ft) apart.
Observations indicated only a small cone of moisture formed beneath each emitter, with dry media intervening
between emitters, and considerable amounts of water draining through the medium to discharge drains. This is in
part due to the soil-less media which preferentially allows water to move though it vertically rather than laterally.
The replacement overhead rotary irrigation system more uniformly distributed moisture across the green roof
planting media. The overhead rotary irrigation was more readily available to the shallow rooted Sedums. This
implies the overhead rotary irrigation is better suited to the green roofs.
The analysis also implies that the zeolite amendments i.e. 33%, 50% and 66% appear to increase plant growth for
certain plants, so lower VMC measurements in these cases imply this moisture is available to the plants for uptake.
The overhead rotary irrigation values were taken at the end of the 2009 year implying that there will be greater time
7-18
-------
for growth and this is being compared to the drip irrigation which included measurements earlier in the season for
2008, i.e., May and June when plant growth would have been minimal. However, there was the extreme die-off of
S. acre and S. album which is also included in this overall analysis and plant cover still dramatically improved with
overhead rotary irrigation as displayed in Figure 7-18 and Figure 7-20. Less irrigation was applied in 2009 with the
overhead rotary irrigation, than in 2008 with the drip irrigation system. Year to year for the months July through
September, there was 10% more rainfall in 2009 and 32% less irrigation.
Conclusions
The overhead rotary irrigation system appears more appropriately suited to an extensive green roof system planted
with Sedums because it effectively supplies irrigation to the media over a wider area than the drip irrigation system.
This observation is in agreement with observations discussed in other regions of North America (Beattie and
Berghage, 2004; Friedrich, 2005).
While both varieties of S. spurium in study 2 benefitted from change to overhead rotary irrigation across all levels of
zeolite amendment, S. acre and S. album had low overwintering survival and had continued decline though the 2009
with application of overhead rotary irrigation even for the control (no zeolite amendment). Even though desiccation
during the winter has been identified as one probable the cause for die-off of S. acre and S. album, this may also
imply that some plants might be affected by changing irrigation type. Further work needs to be done to confirm
effects of zeolite amendment in affecting VMC and confirming this moisture is available to a variety of plants.
Additional analysis of plant biomass may be warranted. Sedums have shallow rooted systems which would seem
more well suited for overhead rotary irrigation while some herbaceous species tend to have more of a tap root
which might benefit from drip irrigation, i.e., if properly placed next to plant. Assessing root mass in addition to top
growth may provide further insight into choosing the right irrigation system for the individual type of plants.
7-19
-------
Chapter 8 Moisture Deficit Study
Introduction
The growing media used for extensive green roofs is extremely porous, very well drained, and prone to extreme
fluctuations in moisture content. Due to the characteristics of the growing media, plant species utilized in extensive
green roof systems must be able to withstand periods of low moisture availability in their root zones. The survival
and growth of plants in an extensive green roof located in a semi-arid region require irrigation, and predictions have
been made that success of extensive green roofs in areas with infrequent precipitation events is improbable unless
supplemental irrigation is provided (Miller, 2003). Additionally, a diversified plant community on an extensive green
roof may be able to respond to variable moisture conditions and maximize the evaporative cooling benefit, thus
extending the benefits of extensive green roofs (Compton and Whitlow, 2006).
Due to the porous and well-drained nature of the typical growing media used in extensive green roof systems,
plants species considered for use in such systems need to be evaluated for their response to gradual and long-term
drying of the growing media. Thus, relative rate of dry down for plants species considered for use in such systems is
an important characteristic to assess. In semi-arid regions, such knowledge will help to determine the need for
irrigation and the frequency of irrigation events for these species. The goal of this study was to determine the
impact of gradual drying of extensive green roof growing medium (Table 8-1) on the growth of fifteen plant species,
and to determine the relative water use for each of the fifteen species (Table 8-2).
Table 8-1 Physical Characteristics of the Growing Medium Used in all Three Trials.
Growing Media Characteristic
Composition by volume
Five parts heat-expanded shale
Two parts sphagnum peatmoss
Two parts perlite
One part vermiculite
Bulk density
Particle density
Saturated hydraulic conductivity
At maximum water capacity
Air content
Water content
Value
50% 63 - 95 mm diameter
30% 20 - 63 mm diameter
20% < 20 mm diameter
-
~
~
0.77 g/cc
2.20 g/cc
0.0087 cm/s
13.8%
51.1%
The media used in this study was based on mixing specs from GreenGrid®. Greenhouse and outdoor studies used
the same media. Media physical properties were analyzed at Hummel & Co, Inc. Laboratory in Trumansburg, NY,
8-1
-------
USA, and reported on March 02, 2010. All physical properties were tested per ASTM E2399. Analytical methods
included organic matter (ASTM F1647, method 1, loss on ignition), dry density, particle density (ASTM D5550),
saturated hydraulic conductivity (permeability), total porosity, and air and water filled porosity at maximum water
capacity.
Table 8-2 Species Evaluated in Greenhouse and Outdoor Trials.
Species
Allium cernuum Roth, (nodding onion)
Antennaria parvifolia Nutt. (small-leaf pussytoes)
Artemisia frigida Willd. (fringed sage)
Bouteloua gracilis (Kunth) Lag. ex Griffiths (blue grama)
Buchloe dactyloides (Nutt.) Engelm. (buffalograss)
Carexflacca Schreb. (heath sedge)
Delosperma cooperi (Hook, f.) L. Bol. (hardy ice plant)
Delosperma nubigenum (Schltr.) L. Bol. (yellow ice plant)
Penstemon pinifolius Greene (pineleaf penstemon)
Sedum acre L. (goldmoss stonecrop)
Sedum album L. (white stonecrop)
Sedum lanceolatum Torr. (lanceleaf stonecrop)
Sedum spurium Marsch-Bieb. (two-lined stonecrop) 'John Creech'
Sempervivum (hens and chicks) 'Royal Ruby'
Thymus pseudolanuginosus Ronn. (woolly thyme)
Non-vegetated control
] Trials
{greenhouse
greenhouse, outdoors
{greenhouse, outdoors
greenhouse
{greenhouse, outdoors
greenhouse
{greenhouse, outdoors
greenhouse
{greenhouse, outdoors
greenhouse
{greenhouse, outdoors
greenhouse, outdoors
{greenhouse, outdoors
greenhouse, outdoors
{greenhouse
greenhouse, outdoors
The green house studies were performed in 2008 and 2009 extending 151 days until dieback conditions for all plants
were observed. The outdoor trial was performed only in 2009 and was truncated to 43 days due to freezing
temperatures. The greenhouse studies used individual plantings in circular green plastic containers 15.2 cm
diameter by 10.8 cm deep pots with media depths of 10 cm. Containers were randomly placed on wire mesh
greenhouse benches equidistantly apart at 2.5 cm (1 in.) as part of a complete block design of 24 replicates. The
outdoor studies used the same type of modules used on the Region 8 green roof and like the Single Species Study
there were 8 replicates in each module.
All plants were established for 10 weeks in a greenhouse. Irrigation was to saturation every 48 h until ten days
before the start of the dry-down study. Irrigation was reduced to every 72 h and then 96 h just before the final
irrigation. For final irrigation, 450 ml was applied and allowed to freely drain. The first VMC was at least 12 hr after
the final irrigation. In addition, each plant received 5 g of Fertilizer (Scotts Osmocote Pro 19-5-8; Scotts-Sierra
Horticultural Products Co., Marysville, OH) four weeks before dry down study began.
Greenhouse Trials
Results for the greenhouse trials show change of VMC for up to 18 days after initiation of dry down period,
depending on the herbaceous (Figure 8-1) and succulent (Figure 8-2) species (error bars represent standard error).
Figure 8-lincludes the non-vegetated control. A dry down period of 18 days is a much longer period of time when
compared with a study conducted in a Michigan greenhouse trial, which found that VMC of a mixture of Sedums
ceased changing after only seven days, with some species reaching 0% VMC in as little as one day (VanWoert et al.,
2005). The dissimilarity between studies is most likely due to differences among species, differences in
developmental stages of plants, differences in growing media depth, solar radiation intensity, container type,
growing media moisture holding capacities and measurement techniques.
8-2
-------
Artemisia frigida
12345678 9101112131415161718 12345678 9101112131415161718
30 -i Boutelouaaracilis 30 1 Buchloedactyloides
I
25 -
12345678 9101112131415161718
Carex flacca
12345678 9101112131415161718
30 -\ Penstemon pinifolius
12345678 9101112131415161718
30 i Thymus pseudolanuginosus
•£ 25 -
0>
M 20 -
I
o
15 -
10 -
5 -
0
30 -,
25 -
20 -
15 -
10 -
5 -
0 -
12345678 9101112131415161718
Non-vegtated Control
123456789 101112131415161718
Days after study initiation
12345678 9101112131415161718
Days after study initiation
Figure 8-1: Mean volumetric moisture content measurements of growing media for herbaceous plants in greenhouse trials
8-3
-------
30 n
Delosperma cooper!
30 i
123456789 101112131415161718
Delosperma nubigenum
123456789 101112131415161718
30 i Sedum album
£ 25 -
£
5 20 •{
I 15
'o
I 1°-
| 5 -
"o
> 0
12345678 9101112131415161718
Sedum spurium 'John Creech'
12345678 9101112131415161718
Days after study initiation
30 -,
25 -
20 -
15 -
10 •
5 -
0 -
123456789 101112131415161718
Sedum acre
12345678 9101112131415161718
30 -,
25 -
20 -
15 -
10 -
5 -
0 -
Sedum lanceolatum
30 -,
25 -
20 -
15 -
10 -
5 -
0 -
12345678 9101112131415161718
Sempervivum 'Royal Ruby'
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Days after study initiation
Figure 8-2: Mean volumetric moisture content measurements of growing media for succulent plants in greenhouse trials
8-4
-------
Figure 8-3 shows the differences in plant appearance from the beginning of the study compared to 12 days into the
study.
:- . t i
m^*M ti^fti
*1P^^i:»;.' >
--' - - , ^^^
Figure 8-3: Example block showing the change in plant appearance a) the day afterthe trial began and b) day 12.
The mean relative water use for each species was estimated from substrate VMC data by subtracting the VMC of the
non-vegetated control from the container VMC for each day and is presented in Table 8-3 along with number of
days to top growth dieback and percent of plants revived after watering. Top growth die back was defined as when
there was no viable green tissue, i.e. foliage and stems, above the substrate surface. When plants reached this state,
plants were watered with 450 ml every 48 hrs to measure to determine if the plant had entered into dormancy or
died. If plants had not died during the 151 day study, 450 ml of water was applied at the end of the study to
evaluate if they could recover from an extended period of drought.
Table 8-3 Mean Relative Water Use, Days to Top Growth Dieback and Percent Revival after Watering in Greenhouse Trials
Species
Antennaria parvifolia
Artemisia frigida
Bouteloua gracilis
Buchloe dactyloides
Carexflacca
Penstemon pinifolius
Thymus pseudolanuginosus
Plant type
herbaceous
herbaceous
herbaceous
herbaceous
herbaceous
herbaceous
herbaceous
Herbaceous mean
Allium cernuum
Delosperma cooperi
Delosperma nubigenum
Sedum acre
Sedum album
Sedum lanceolatum
Sedum spurium 'John Creech'
Sempervivum 'Royal Ruby'
succulent
succulent
succulent
succulent
succulent
succulent
succulent
succulent
Succulent mean
Mean relative water use (SE)
-1.53% (0.19) c1
-6.23% (0.74) k
-4.63% (0.46) hi
-3.56% (0.35) f
-4.77% (0.50) hij
-3.63% (0.32) fg
-2.26% (0.20) de
-3.80%
-1.13% (0.25) b
-5.24% (0.60) ij
-5.56% (0.69) ijk
-2.72% (0.3 l)e
-4.48% (0.60) gh
-1.22% (0.27) be
-2.00% (0.22) cd
+2.04% (0.36) a
-2.54%
Days to dieback (SE)
22.79 (0.65) d
16.08 (0.32) a
18.23 (0.71) ab
20. 19 (0.90) be
20. 13 (0.90) be
20.09 (0.67) be
20.75 (0.87) c
19.75
59.25 (1.77) f
52.25(1.44)e
107.06 (3.46) g
107.67 (6.46) g
151.00 (0.00) j
138.71 (2.53) i
127.87 (3.72) h
151.00 (0.00) j
111.75
Revival
31.25%
8.33% 1
22.92%
37.50%
27.08%
0.00%
31.25%
22.62%
91.67%
0.00%
2.08%
2.08%
58.33%
54.17%
56.25%
69.44%
41.75%
Lower case letters show significant differences at the p< 0.05 level
8-5
-------
Table 8-3indicates that only one species, Sempervivum 'Royal Ruby', had lower water usage than the non-vegetated
control which potentially implies CAM. Results for number of days to top growth die back show a clear division
between the herbaceous and succulent species. There was a nearly six-fold difference in days to dieback for the
herbaceous plants versus the succulent species (Table 8-3). The herbaceous plants had a mean revival of 22.62%
while the succulent species had a mean revival of 41.75% (Table 8-3).
Two of the succulent species (S. album and Sempervivum 'Royal Ruby') did not have any replications that died back
during any of the 151-day trials (Table 8-3); however, while S. album and S. 'Royal Ruby' survived the initial dry
down period, once rewatering commenced, some individuals died. These are similar results to a study in Michigan
where the succulent species of Sedums remained viable for the entire four month study period (Durhman et al..
2004). Another Sedum, S. rubrotinctum has been shown to remain alive for up to two years in a greenhouse without
irrigation (Teeri etal., 1986).
Outdoor Trial
Results for the outdoor trials show change of VMC for up to 18 days after initiation of dry down period, depending
on the herbaceous (Figure 8-4) and succulent (Figure 8-5) species (error bars represent standard error). The results
of the outdoor trial had similar trends to the greenhouse trials concerning rate of dry down of the herbaceous and
succulents. In general, the growing media of succulent plants dried down more slowly than the herbaceous plants,
although exceptions did occur. This can been seen in Table 8-4 which shows results of simple rate analysis (liner
regression). It indicates that the succulants loose water at a slightly slower rate than the herbaceous plants and at a
rate similar to that of media. Butler and Orians (2011) observed that sedums may reduce water loss from the green
roof media.
30 n
Antennaria parvifolia
Artemisia frigida
c
o
U
d>
_
o
30 n
2345678 9101112131415161718
Buchloe dactyloides 30 -,
25 -
20 -
12345678 9101112131415161718
Non-vegetated Control
23456789 101112131415161718
Days after study initiation
12345678 9101112131415161718
Days after study initiation
Figure 8-4: Mean volumetric moisture content measurements of growing media for herbaceous plants in outdoor trials
8-6
-------
£
c
o
U
I
'o
^o
'l_
'S
"o
30 -i
25 -
Penstemon pintfolius
30 n
25 -
20 -
5 15 -
10 -
_
o
12345678 9101112131415161718
Sedum album
12345678 9101112131415161718
Sedum lanceolatum
12345678 9101112131415161718
Sedum spurium 'John Creech'
30 -,
25 -
20 -
15 -
10 -
5 -
0
12345678 9101112131415161718
Sempervivum 'Royal Ruby'
123456789 101112131415161718
Days after study initiation
12345678 9101112131415161718
Days after study initiation
Figure 8-5: Mean volumetric moisture content measurements of growing media for succulent plants in outdoor trials
8-7
-------
Table 8-4 Mean Relative Water Use, Days to 50% Volumetric Moisture Content Loss and Rate of Loss in Outdoor Trials
Species
Antennaria parvifolia
Artemisia frigida
Bouteloua gracilis
Plant type
Herbaceous
Herbaceous
Herbaceous
Herbaceous mean
Delosperma cooperi
Penstemon pinifolius
Sedum album
Sedum lanceolatum
Sedum spurium 'John Creech'
Sempervivum 'Royal Ruby'
Succulent
Succulent
Succulent
Succulent
Succulent
Succulent
Succulent mean
Non-vegetated control
Days to 50% or
greater VMC loss
3
3
3
3.0
o
J
o
J
o
J
o
J
o
J
4
3.2
3
Rate of waterless
3 day (r2)
-7.1(0.98)
- 7.6 (0.97)
- 6.6 (0.97)
-7.1(0.98)
- 6.4 (0.88)
- 7.0 (0.97)
- 6.9 (0.99)
-8.1(0.98)
-7.5(0.98)
- 5.7 (0.97)
- 6.9 (0.84)
-6.9(0.98)
6 day (r2)
- 4.2 (0.92)
-3.9(0.86)
- 4.0 (0.92)
- 4.0 (0.90)
-3.5(0.88)
- 4.0 (0.90)
-3.6(0.86)
-4.6(0.91)
-3.9 (0.86)
-3.5(0.90)
-3.9(0.89)
-3.8(0.90)
The number of days to dieback and revival rates were quantified for the outdoor trial (Table 8-5). This table shows
the mean difference in growing media VMC from the non-vegetated control, days to top growth dieback and
percent revival after re-watering. The growing media of A. parvifolia retained more moisture for a longer period of
time than did the growing media of most of the succulent species, except Sempervivum 'Royal Ruby', which is similar
to what occurred in the greenhouse trials though S. lanceolatum retained slightly more moisture in the greenhouse
trials. Once again only the species Sempervivum 'Royal Ruby' had a mean lower water usage than the non-
vegetated control. The freezing temperatures prematurely truncated the study preventing the longer surviving
succulent species from completing the dieback process as was done in the greenhouse trials. Therefore, results for
the succulent species are not applicable except for the fact that they all remained viable for greater than the 43 days
of the trial prior to exposure to freezing temperatures.
"able 8-5 Mean Relative Water Use, Days to Top Growth Dieback and Percent Revival after Watering for Outdoors Trial
Species
Antennaria parvifolia
Artemisia frigida
Buchloe dactyloides
Plant type
Herbaceous
Herbaceous
Herbaceous
Herbaceous mean
Delosperma cooperi
Penstemon pinifolius
Sedum album
Sedum lanceolatum
Sedum spurium 'John Creech'
Sempervivum 'Royal Ruby'
succulent
succulent
succulent
succulent
succulent
succulent
Succulent mean
Mean relative water use (SE)
-0.71% (0.2 l)b:
-3.21% (0.31) de
-1.57% (0.15) be
-1.83%
-2.62% (0.25) d
-1.82% (0.08) c
-3.35% (0.36) e
-0.94% (0.23) be
-2.57% (0.20) d
+0.23% (0.15) a
-1.85%
Days to dieback (SE)
3 1.4 (0.24) c
20.0 (0.31) a
27.7 (0.25) b
27.58
NA2
3 1.2 (0.42) c
NA
NA
NA
NA
NA
Revival"
54.17%
50.00%
41.67%
41.67%
NA
20.83%
NA
NA
NA
NA
NA
Lower case letters show significant differences at the p< 0.05 level
NA = Not applicable (due to truncation of study from freezing temperatures.)
Comparison between Trials
A visual comparison of the two sets of dry down curves between the greenhouse and outdoor trials shows
qualitative differences. These differences can be explained by divergent environmental conditions. Greenhouse
growing conditions had lower solar radiation due to filtration through the greenhouse covering (Table 8-6). Lower
8-8
-------
solar radiation in the greenhouse would lower ET rates as compared to higher solar radiation and outdoor winds
which would increase ET rates, especially in a semi-arid climate such as Colorado. A rooftop environment could
potentially have an even higher ET rate than either the greenhouse or the outdoor trial conditions in this study due
to higher temperatures and lower relative humidity in urban areas (Schmidt, 2006).
Table 8-6 Environmental Conditions Daily Means Derived from Measurements in the Greenhouse and Outdoor Trials
Trial
Greenhouse
Greenhouse
Outdoor
Dates
9/03/2008 to 9/30/2008
9/03/2009 to 9/30/2009
8/20/2009 to 9/30/2009
Temperature (°C)
21.9(0.05)
21.7 (0.04)
16.7(0.13)
Relative humidity (%)
57.7(0.13)
56.8(0.16)
58.8 (0.28)
Maximum solar radiation (W-m~2)
162 (1.87)
163 W-m"2 (1.91)
311 W-rn2 (4.36)
Standard errors in parenthesis.
The number of days to dieback took longer outdoors than in the greenhouse. There were differences in the amount
of growing media not covered by plant canopy. Also, it is likely that the cooler nighttime temperatures outdoors
(than in the greenhouse) would reduce nighttime evaporation. As the modules used outdoors had a greater rooting
volume to draw moisture from (Figure 8-6), theoretically additional moisture from those areas of the module
without vegetation between the plants was available. Figure 8-6 shows Sempervivum 'Royal Ruby' in greenhouse
containers (a) and outdoor containers (b). In general, revival rates were also greater outdoors than indoors,
potentially due to increased root zone.
Figure 8-6 Photo examples of different containers used in green house (a) and outdoor (b) trials
Note: images are not of the same scale
Recommendations
Due to the differences in dry down rates and number of days to dieback between succulents and herbaceous
species, the frequency of irrigation recommendations are different. For succulent species, it has been recommended
that irrigation be provided at 28 day intervals for growing media at a depth of 6 cm (2.5 in.) (VanWoert et al., 2005).
This study would concur with that recommendation as all succulents remained viable for at least 28 days following
8-9
-------
an irrigation event. Additionally, while it is difficult to establish permanent wilting points for many succulent species
because they retain moisture in their foliage (Berghage et al., 2007), irrigating at least 10 days after VMC ceases to
change (Day 18 in this study) appears to be an appropriate and resourceful management tactic for extensive green
roofs (Bousselot et al., 2011).
The herbaceous plants in this study will require more frequent irrigation than the succulents. If days to dieback are
an indication of tolerance of low VMC, then irrigation should be provided more often than every 16 days for
herbaceous species, which was the mean days to dieback for the earliest species to dieback, A. frigida. Even if VMC
drops below wilting point (point where water is not longer available to plants), these species should be able to
remain viable temporarily until moisture is again supplied. Therefore, irrigation frequency recommendations for the
herbaceous species in this study are at least every 14 days (Bousselot et al., 2011).
8-10
-------
Chapter 9 References
Allen, R. G., L. S. Pereira, D. Raes and M. Smith (1998). Crop evapotranspiration - Guidelines for computing crop
water requirements - FAO Irrigation and drainage paper 56. FAQ Irrigation and Drainage Papers. Rome,
Italy, Food and Agriculture Organization of the United Nations.
Beattie, D. and R. Berghage (2004). Green roof media characteristics: the basics. In Proc. of 2nd North American
Green Roof Conference: Greening Rooftops for Sustainable Communities. Portland, OR. 2-4 June 2004, The
Cardinal Group, Toronto.: 411-416.
Berghage, R., D. Beattie, A. Jarrett and F. Rezaei (2007). Green roof plant water use. In Quantifying evaporation and
transpirational water losses from green roofs and green roof media capacity for neutralizing acid rain. R.
Berghage, A. Jarrett, D. Beattieet al. State College, PA, The Pennsylvania State University: 18-38.
Bousselot, J. M., J. E. Klett and R. D. Koski (2010). "Extensive Green Roof Species Evaluations Using Digital Image
Analysis." HortScience 45(8): 1288-1292.
Bousselot, J. M., J. E. Klett and R. D. Koski (2011). "Moisture Content of Extensive Green Roof Substrate and
Growth Response of 15 Temperate Plant Species during Dry Down." HortScience 46(3): 518-522.
Butler, C. and C. M. Orians (2011). "Sedum cools soil and can improve neighboring plant performance during water
deficit on a green roof." Ecological Engineering 37(11): 1796-1803.
Compton, J. S. and T. H. Whitlow (2006). A zero discharge green roof system and species selection to optimize
evapotranspiration and water retention. In Proc. of 4th North American Green Roof Conference: Greening
Rooftops for Sustainable Communities. Boston, MA. 10-12 May, 2006, The Cardinal Group, Toronto.
Cook-Patton, S. C. and T. L. Bauerle (2012). "Potential benefits of plant diversity on vegetated roofs: A literature
review." Journal of Environmental Management 106(0): 85-92.
Durhman, A. K., D. B. Rowe and C. L. Rugh (2006). "Effect of watering regimen on chlorophyll fluorescence and
growth of selected green roof plant taxa." HortScience 41: 1623-1628.
Durhman, A. K., D. B. Rowe and C. L. Rugh (2007). "Effect of substrate depth on initial growth, coverage, and
survival of 25 succulent green roof plant taxa." HortScience 42: 588-595.
9-1
-------
Durhman, A. K., N. D. VanWoert, D. B. Rowe, C. L. Rugh and D. Ebert-May (2004). Evaluation of Crassulaceae
species on extensive green roofs. In Proc. of 2nd North American Green Roof Conference: Greening
Rooftops for Sustainable Communities. Portland, OR. 2-4 June 2004, The Cardinal Group, Toronto.: 504-
517.
Passman, E. A. and R. Simcock (2008). Development and Implementation of Locally Sourced Extensive Green Roof
Substrate in New Zealand. World Green Roof Congress. London, England: 14.
FLL (2008). Guidelines for the Planning, Construction and Maintenance of Green Roofing, Forschungsgesellschaft
Landschaftsentwicklung Landschaftsbau e. V. (FLL): 122.
Friedrich, C. R. (2005). Principles for selecting the proper components for a green roof growing media. In Proc. of
3rd North American Green Roof Conference: Greening Rooftops for Sustainable Communities. Washington,
DC. 4-6 May 2005, The Cardinal Group, Toronto: 262-274.
Getter, K. L. and D. B. Rowe (2006). "The role of extensive green roofs in sustainable development." HortScience 41:
1276-1285.
lies, J. and N. Agnew (1995). "Seasonal cold-acclimation patterns ofSedum spectabile x telephium L. 'Autumn Joy'
and Sedum spectabile Boreau. 'Brilliant'." HortScience 30(6): 1221-1224.
Lundholm, J., J. S. Maclvor, Z. MacDougall and M. Ranalli (2010). "Plant Species and Functional Group
Combinations Affect Green Roof Ecosystem Functions." PLoS ONE 5(3): e9677.
Miller, C. (2003). Moisture management in green roofs. In Proc. of 1st North American Green Roof Conference:
Greening Rooftops for Sustainable Communities. Chicago, IL. 29-30 May 2003., The Cardinal Group,
Toronto.: 177-182.
Miller, C. (2011). "Extensive Vegetative "Green" Roofs." from http://www.wbdg.org/resources/greenroofs.php.
Miller, G. (2000). "Physiological response of bermudagrass grown in soil amendments during drought stress."
HortScience 35(2): 213-216.
Monterusso, M. A., D. B. Rowe and C. L. Rugh (2005). "Establishment and persistence ofSedum spp. and native taxa
for green roof applications." HortScience 40: 391-396.
Murphy, J., H. Samaranayake, J. Honig, T. Lawson and S. Murphy (2005). "Creeping bentgrass establishment on
amended-sand root zones in two microenvironments." Crop Science 45(4): 1511.
Philippi, P. M. (2011). "Introduction to the German FLL-Guideline for the Planning, Execution and Upkeep of Green-
Roof Sites." Retrieved 11/16/11, 2011, from
http://www.epa.gov/region8/greenroof/pdf/IntroductiontotheGermanFLL2.pdf
Rowe, D. B., M. A. Monterusso and C. L. Rugh (2006). "Assessment of heat-expanded slate and fertility
requirements in green roof substrates." HortTechnology 16: 471-477.
Savonen, C. (2012, 6/27/12). "Snow or lack thereof - effects on landscape plants." from
http://extension.oregonstate.edu/gardening/snow-or-lack-thereof-effects-landscape-plants.
Sayed, O. H. (2001). "Crassulacean acid metabolism 1975-2000, a check list." Photosynthetica 39(3): 339-352.
9-2
-------
Schmidt, M. (2006). The evapotranspiration of greened roofs and facades. In Proc. of 4th North American Green Roof
Conference: Greening Rooftops for Sustainable Communities. Boston, MA. 10-12 May, 2006, The Cardinal
Group, Toronto.
Smeal, D., M. K. O'Neill, K. A. Lombard and R. N. Arnold (2010). Climate-Based Coefficients for Scheduling
Irrigations in Urban Xeriscapes. 5th National Decennial Irrigation Conference. Sponsored jointly by
American Society of Agricultural and Biological Engineers (ASABE) and the Irrigation Association.
Phoenix, Arizona, ASABE: 10.
Tackenberg, O. (2007). "A new method for non-destructive measurement of biomass, growth rates, vertical biomass
distribution and dry matter content based on digital image analysis." Annals of Botany 99(4): 777-783.
Teeri, J. A., M. Turner and J. Gurevitch (1986). "The response of leaf water potential and Crassulacean acid
metabolism to prolonged drought in Sedum rubrotinctum" Plant Physiology 81(2): 678-680.
Ting, I. P. (1985). "Crassulacean Acid Metabolism." Annual Review of Plant Physiology 36(1): 595-622.
USDA. (2012). "The PLANTS Database." NRCS, National Plant Data Team, Greensboro, NC 27401-4901 USA.
Retrieved 8/31/12, 2012, from http://plants.usda.gov.
VanWoert, N. D., D. B. Rowe, J. A. Andresen, C. L. Rugh and L. Xiao (2005). "Watering regime and green roof
substrate design affect Sedum plant growth." HortScience 40: 659-664.
Virta, R. L. (2001). Zeolites. Minerals Yearbook - 2001. USGS: 84.81 - 84.84.
Virta, R. L. (2009). Zeolites. 2009 Minerals Yearbook. USGS: 83.81-83.83.
White, J. W. and E. Snodgrass (2003). Extensive greenroof plant selection and characteristics. In Proc. of 1st North
American Green Roof Conference: Greening Rooftops for Sustainable Communities. Chicago, IL. 29-30 May
2003., The Cardinal Group, Toronto.: 166-176.
9-3
------- |