Colorado State University Green Roof Research
PROJECT/TASK DESCRIPTION
The data collected through this project were used to achieve three objectives:
- Determine herbaceous plant species suitable for green roof use in the semi-arid, high elevation Front Range of Colorado.
- Determine media types or mixes suited to supporting extensive green roof plants.
- Identify additional areas for expanded research.
Objective #1: Determine herbaceous plant species suitable for green roof use in the semi-arid, high elevation Front Range of Colorado.
Plants were selected for experimentation based on certain criteria: ability to grow in dry soils with only limited amounts of available water (such plants tend to be highly drought resistance), growth habit - groundcovers (beneficial for extensive green roofs to obtain good coverage) or accent plants (good for contrast with groundcovers in heights and bloom times), foliage chartacteristics - evergreen or deciduous, and bloom characteristics - flower color and length of bloom season (Table T-1). Species which are native to Colorado were of particular interest because these are adapted to the extreme conditions of the climate characterized throughout much of Colorado.
Table T-1: GREEN ROOF PLANT SPECIES TESTED
Species |
Benefits |
|||||
|
drought |
groundcover |
evergreen |
native |
long bloom |
bloom color |
Antennaria parvifolia |
X |
X |
X |
X |
X |
white |
Bouteloua gracilis |
X |
X |
|
X |
|
turf |
Delosperma cooperi |
X |
X |
X |
|
X |
deep pink |
Eriogonum umbellatum |
X |
X |
X |
X |
X |
yellow |
Opuntia fragilis |
X |
accent |
X |
X |
|
varies |
Sedum lanceolatum |
X |
X |
X |
X |
|
yellow |
Plants that are drought resistant typically employ one of three methods: avoid, escape, or tolerate. Not all types of drought resistance in plants are well-suited to green roofs. For example, some plants avoid drought by rooting deeply to access a more stable supply of water; this will obviously not be possible on a green roof. Plants that use the escape tactic are not ideal for green roofs either because they have short life cycles, timed to grow and reproduce during the rainy season, and green roofs should ideally be green at least throughout the growing season, if not year round. True drought tolerance is not a trait commonly found in plants; these plants merit further investigation for use in green roofs.
One example of true drought tolerance that does fit in with green roof systems is Crassulacean acid metabolism ( CAM). Many Sedum species are CAM plants. Crassulacean acid metabolism refers to the family name of these plants (Crassulaceae) and the way they metabolize or utilize carbon. CAM plants are truly drought tolerant because they keep their stomata (pore-like structures where gas exchange and transpiration occur) closed during the day when transpiration rates are high, and open them at night when transpiration rates are significantly lower. While it is good that CAM plants keep their stomata closed during the day, they have to manage their CO 2 intake much differently than other plants. Carbon dioxide is needed while photosynthesis is taking place (daytime) so CAM plants have to convert the CO 2, which is brought in only at night, into a useable form for use during the day. This process takes additional energy, slowing the growth of the plant. Still, the tradeoff of higher energy use for water savings is beneficial to the plant in drought conditions.
Environmental conditions were monitored by use of Campbell Scientific weather monitoring equipment (Table T-2). These data were used to help interpret plant performance on the research green roof.
Table T-2: WEATHER MONITORING EQUIPMENT
Campbell Scientific Equipment (Model #) |
Description |
Range of Tolerance |
Accuracy/ Precision |
Infrared Radiometer (IRR-P) |
Surface temperature of vegetation |
-55º to +80ºC |
±0.2ºC @ -10º to +65ºC; ±0.5ºC @ -40º to 70ºC |
Temperature and Relative Humidity Probe (HMP45C) |
Measures temperature and RH at 12 inch height |
-40º to +60ºC |
±2% over 10-90% RH; ±3% over 90-100% RH |
Young Wind Sentry set (03001-L) |
Wind speed and direction |
0 to 50 m/s |
±0.5 m/s |
Tipping Bucket (TE525WS-L) |
Precipitation gage |
0° to +50°C |
Up to 1 in/hr = ±1%, 1-2 in/hr = +0, -2.5%; 2-3 in/hr = +0, -3.5%; |
Snowfall conversion adaptor (CS705) |
Converts snowfall into rain equivalent |
to -20°C |
assumes 1:0 starting ratio of antifreeze to water |
Silicon Pyranometer (LI200X) |
Solar radiation sensor |
-40° to +65°C |
Absolute error in daylight is ±5% max; ±3% typical |
Datalogger (CR1000) |
Data storage device |
|
Battery: 12 volt PS100 Campbell Scientific |
Objective #2: Determining media types or mixes suited to support extensive green roof plants
Most extensive green roof media is made up predominantly of expanded slate, shale or clay. While these materials are very well-drained, lightweight (but do not blow away) and do not break down like organic materials, they do have some limitations. They typically drain too quickly (too much macro-pore space, not enough micro-pore space) and do not hold nutrients very well (low cation exchange capacity - CEC).
Zeolites are materials that have all of the benefits of expanded slate, shale and clay, plus have more micro-pore space and higher CEC. Zeolites are currently being utilized as amendments for shallow, well-drained golf greens.
SAMPLING PROCESS DESIGN
For all three studies, there were five blocks or replications. The five areas were zoned as 3 x 8 tray areas on the existing EPA Region 8 headquarters green roof. See Figure F-1 for a graphical depiction of the design layout on the roof for all three studies. Twelve of the 24 existing trays in each block were randomly chosen and assigned to one of the three studies. Six of the trays were for the Species Study, the area of four trays was for the Media Study, and two trays were for the Mixed Species Study. The 12 remianing trays remained as the existing green roof to provide an in situ environment (see Figure F-2).
Figure F-1: TRAY CONFIGURATION FOR ALL THREE STUDIES
Figure F-2: PHOTO DEPICTION OF TRAY LAYOUT (04/09/08)
Data (See Section B2 for descriptions of data capturing methods.)
For all three studies, initial data were taken on species survivability, and of those species that survived, growth rate measurements were used to determine success. Photographs to determine growth rate by measuring change in area covered per week were also taken. The photographs were taken once a week, during the growing season (~April 1 st - November 1 st) in a predetermined order at approximately the same time of day. The predetermined order helped to ensure that the plants were photographed at about the same time of day each week. Plant widths and heights were also measured and recorded. Top growth dry weights of each plant in the experiment were taken at the end of the experiment.
The ability of a species to survive in the shallow, well-drained media of extensive green roofs is often reflective of the species degree of drought tolerance. Because different plant species use water at different rates; a species water-use efficiency is valuable information when attempting to determine how appropriate a species will be for use in green roof applications. Soil moisture content was measured to compare relative drought resistance of the plant species. Delta-T Theta Probe ML2X (Delta-T Devices, Cambridge, UK) was used to take instantaneous readings of volumetric soil moisture content (Figure F-3). Researchers at Michigan State University have successfully used Delta-T Theta Probes ML2X in green roof research with similar media types (Durhman et al. 2006, Monterusso et al. 2005, VanWoert et al. 2005).
Figure F-3: DELTA-T THETA PROBE ML2X SOIL MOISTURE SENSOR
Overwintering success of each species evaluated was of particular importance, as Front Range Colorado winters are typically characterized by warm sunny days (frequently up to 60ºF [15ºC] or above) and freezing nights, with frequent high winds, and unpredictable precipitation and snow cover duration. These moisture-limitating conditionas create challenging environments for plant species. Plants still require moisture during the winter to prevent winter desiccation and maintain adequate root metabolism. Between November 1st and April 1st of each year of the study, the green roof was monitored monthly to check on plant health and collect weather data.
The treatment were one of six species (Table T-1). A series of the six species/treatments (one species per tray) were put into one of the five blocks described above. Thus a total of six species, replicated five times equaled 30 trays. In a tray, eight individual plants were planted, each with a square foot (~0.09m 2) of growing space (2ft x 4ft [0.61m x 1.22m] trays) (Figure F-4).
Figure F-4: EXAMPLE OF TRAY LAYOUT FOR SPECIES STUDY
Plants are represented by ¤, solid lines correspond to tray edges and dotted lines show the imaginary lines between the 1 ft 2 (0.093 m 2) areas.
For the green roof media trial, 4 taxa of Sedum already on the green roof (Sedum acre, S. album, S. spurium ‘Dragons Blood’ and S. spurium ‘John Creech’) were planted into one of four percentages of zeolite mixes (0%, 33%, 66% or 100%) to determine which concentration was most suitable for plant growth. The percentages of media were made up of the media already in use on the EPA Region 8 green roof. The media mix (80% inorganic and 20% organic materials) was a proprietary blend owned by Weston Solutions for use in the Green Grid product line.
The experiment was replicated 10 times and set up as a randomized complete block design, similar to the Species Study, but the variable was media type instead of species. Another difference between the Species Study and the Media Study was that the trays were a smaller; only 2ft by 2ft (0.61m x 0.61m) (Figure F-5). Four trays together were a replication. All four percentages of media were randomly assigned to one of the four areas in each replication. The Media Study was set up in this manner to keep environmental variability to a minimum.
Figure F-5: EXAMPLE OF TRAY LAYOUT FOR the MEDIA STUDY
Plants are represented by ¤, solid lines correspond to tray edges and dotted lines show the imaginary lines between the 1 ft 2 (0.093 m 2) areas.
Also similar to the Species Study, the Media Study used Delta-T Theta Probe ML2X (Delta-T Devices, Cambridge, UK) to take instantaneous readings of soil moisture content in millivolts (mV). Soil moisture holding capacity of the media was predicted to increase with zeolite content in the mix.
The Mixed Species Study (study 3) will be set up like the Species Study except eight different plant species (Table T-4) were planted together in each of 10 2ft x 4ft (0.61m x 1.22m) trays. Five of the trays were planted with the existing green roof media and another five trays were planted in a 50% by volume zeolite mixed with the existing media. One tray of each was placed in each of the five blocks. Special attention was paid to plant interactions and changes in available soil moisture compared to the Species Study.
Table T-4: PLANT SPECIES IN the MIXED SPECIES STUDY)
Species, Scientific Name |
Common Name |
Allium cernuum |
Nodding Onion |
Antennaria parvifolia |
Small-leaf Pussytoes |
Bouteloua gracilis |
Blue Grama |
Delosperma cooperi |
Hardy Ice Plant |
Eriogonum umbellatum |
Kannah Creek® Buckwheat |
Opuntia fragilis |
Brittle Pricklypear |
Sedum lanceolatum |
Lanceleaf Stonecrop |
Sempervivum rubrum |
Hens and Chicks, Houseleek |
As a measure of plant growth rate and success, plant expansion (grown rate) were measured each week during the growing season (~April 1 st - November 1 st) by using a series of digital photographs. The camera 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 were 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 (Fuji Film S3000 with a 6x optical zoom 3.2 mega pixels lens) and image settings were used to keep constant any differences these factors could make in image quality.
The digital images were analyzed with SigmaScan Pro 5.0 image analysis software (SPAA Science, Chicago). This image analysis was used to draw outlines for each plant in each digital image. Changes from one week to the next in area covered indicated growth rate. Researchers at Michigan State University successfully used this method in their trials to measure growth rates of green roof species (Durhman et al. 2007).
Individual plant widths and heights were measured weekly for each of the three studies. Plants were numbered as described in Figure F-6. Two widths were taken; the first parallel to the short end of the tray (2ft [0.61m]) and the second perpendicular to it. Plant height was taken at the center of the plant.
All above media portions of each plant in both studies were harvested at the end of the experiment. Root weights were not measured because neighboring plant roots grew together, thus making it difficult to separate each plant. Plants were cut at media level, rinsed in water to remove media debris, patted dry with a paper towel, and then weighed for fresh weight mass. Harvested plants were then inserted into a prelabeled 13x7.9x27cm brown paper bag (Rite Aid, Harrisburg, PA, USA) to allow air and water movement through the paper. The samples were dried in an oven at 70ºC for 72 hours and weighed for dry weights. The balance used to measure fresh and dry weights was a Sartorius model number R200D (Sartorius Bohemia, New York, USA). Scale calibration was performed twice each year by the Vivanco laboratory group to 0.00 mg with calibration weights.
End of experiment dry weights were compared to initial dry weights taken at the beginning of the experiment. Plants for initial dry weights planted on the same date as the experimental trays were harvested at the time the experimental trays were delivered to the green roof. Therefore initial dry weight plants were treated the same as experimental plants in the greenhouse during the establishment period. The treatment of fresh and dry weight samples (handling, temperature, etc.) was the same for initial dry weights as for end of experiment dry weights.
Delta-T Theta Probe ML2X (Delta-T Devices, Cambridge, UK) were used to take instantaneous readings of soil moisture content in millivolts (mV). The probe was inserted into the media until the probe ends rest on the bottom of the tray. A reading was taken via the attached meter instantly and copied down on a data sheet.
Soil moisture content was measured once a week during the growing season (~April 1 st - November 1 st) in a predetermined order at approximately the same time of day keeping constant the number of hours since watering. Seven total measurements were taken in the larger 2ft x 4ft (0.61m x 1.22m) trays and three total measurements were taken in each of the smaller 2ft x 2ft (0.61m x 0.61m) trays. For the larger trays, three measurements were taken down the center of the tray and two on each side of the tray to get an even distribution of media moisture within the tray. Similarly, two measurements were taken down the center and one on the side of the smaller trays. (Figure F-7)
Figure F-7: EXAMPLE OF MEDIA MOISTURE MEASUREMENT LOCATIONS
An example of media moisture measurement locations (represented by the black squares) for the two different tray sizes. Plants are represented by ¤, solid lines correspond to tray edges and dotted lines show the imaginary lines between the 1 ft 2 (0.093 m 2) areas.
Accuracy of the Theta Probe is ±0.01 m 3/m 3 in 0-40ºC. Soil specific calibrations were performed to ensure accuracy at this level. There is no method for calibrating the sensor; however, accuracy was tested weekly by dipping the probe in a cup of water and getting a reading of 100% volumetric moisture content.
An official Quality Assurance Project Plan (QAPP) for this research project can be found on the EPA Region 8 website at: http://www.epa.gov/region8/greenroof/documents/index.html.