tesis impacto de cubiertas verdes en habitación

download tesis impacto de cubiertas verdes en habitación

of 10

Transcript of tesis impacto de cubiertas verdes en habitación

  • 8/13/2019 tesis impacto de cubiertas verdes en habitacin

    1/10

    Roofenvelope ratio impact on green roof

    energy performance

    Ryan Martens &Brad Bass &Susana Saiz Alcazar

    Published online: 15 April 2008# Springer Science + Business Media, LLC 2008

    Abstract This paper addresses the impact of roof-to-envelope ratio on overall energy savings of

    a green roof design over conventional roof designs. Simulations were performed using a

    modified version of the Environmental System Performance program simulator, developed at the

    University of Strathclyde. The modified design employed a model developed by Columbia

    University and the Goddard Institute of Space Science which models the evapotranspiritive

    effect of a green roof calculated using the Bowen ratio; that is, the ratio of sensible heat flux to the

    surrounding air to the latent heat flux resulting from evapourative energy losses. The resultingheat flux term is proportional to the external surface convection, but inversely proportional to the

    surface Bowen ratio, which is held constant and chosen to match experimental results obtained

    for a given roof design. The present study performed simulations for the month of July in a

    Toronto climate on square warehouse style one, two, and three-story buildings, with windows

    occupying 10% of the area of each wall. For the first set of simulations, the internal building load

    of each story was set to zero, and the roofenvelope ratio was increased by increasing the

    building width and length. For the final simulations, several roofenvelope ratios were chosen,

    and the internal load of each story was increased from 0 to 50,000 W. As the roofenvelope ratio

    increases, the cooling load of the upper floor for multi-story designs approaches the entire

    building cooling load. This indicates the importance of upper zone cooling in total building

    energy reductions. Furthermore, the total energy savings of a green-roofed building over a

    conventional roofed building were far more significant for single-story structures. A 250 250 m

    green-roof design with 50,000 W internal loading was found to have percentage energy savings

    of 73%, 29%, and 18%, for a one, two, and three-story design, respectively.

    Keywords Green roof. Building energy simulation. ESP-r. Building envelope . Latent flux

    Urban Ecosyst (2008) 11:399408

    DOI 10.1007/s11252-008-0053-z

    R. Martens (*) :B. Bass :S. S. Alcazar

    Adaptation and Impacts Research Division, Environment Canada, 33 Willcocks Street, Toronto, ONM5S 3E8, Canada

    e-mail: [email protected]

    B. Bass

    e-mail: [email protected]

    S. S. Alcazar

    e-mail: [email protected]

  • 8/13/2019 tesis impacto de cubiertas verdes en habitacin

    2/10

    Introduction

    Historically green roofs have been used as a medium for providing insulation and protection in

    cold climates, mainly in northern parts of Europe, while in hot dry climates they have been used

    to cool the indoor air and increase its moisture content through evaporative processes.Nowadays, green roofs are being implemented not only due to their thermal insulation

    properties, but also because they are considered as a way of recovering the benefits provided by

    the lost green space of cities. Various studies which have analyzed the effects of green roofs have

    focused on a number of key benefits, including reduction in energy consumption, reduction in

    greenhouse gas emissions, reduction in heat island effect, improvement in air quality, reduction

    of urban noise, storm water management, and various social and recreational opportunities.

    Green roofs reduce the heat transfer through the roof, thus reducing the heating and

    cooling energy consumption in the building. The heat transfer processes in a common flat

    roof, that is, convection, conduction and radiation, are affected by the green roof not only

    through a change in the thermal characteristics of the materials and surface properties, butalso through the evapotranspiration and the plant metabolic processes occurring within the

    plant system. It is therefore necessary to analyze the thermal performance of both the soil

    and vegetation layers added to a conventional roof design.

    Niachou et al. (2001) conducted a measurement of surface and air temperatures on green

    roofs to examine their thermal properties and potential energy savings. The authors

    analyzed the effect of green roofs with different levels of insulation, reporting reductions in

    energy consumption ranging from 40% for the non-insulated roof to 2% for the well

    insulated roof (considering roof conductance values of 0.4 W/m2C). This work employed

    the green roof mathematical model developed by Palomo del Barrio (1997) to evaluate if agreen roof could provide a cooling effect to the buildings. The conclusion of this study was

    that green roofs increase the insulating value of the roof, but do not provide any additional

    cooling effect on the building.

    In another study, several experiments were conducted comparing insulated and non

    insulated green roofs (Eumorfopoulou and Aravantinos 1998). It was concluded that green

    roofs must be used as a complementary system to the common insulated roof, and cannot

    substitute the insulation layer to achieve thermal comfort inside the buildings. The study

    identified the shading effect of the green roof as the main factor in the improvement of the

    thermal performance.

    Wong et al. (2002) analyzed the effect of different types of green roofs on the energyconsumption of a five storey commercial building in Singapore. They compared different

    types of vegetation layers (e.g., shrubs, trees and turfing) and various soil thickness. They

    reported total energy savings of 15% in relation to the energy consumption of the building

    with a common flat roof when a green roof composed by shrubs and 300 mm of clay soil with

    40% moisture content was installed in a building with a roof to wall surface ratio of 0.2.

    Although green roof technology has been under development for some time, a reliable and

    general-purpose model does not yet exist for performing energy-based simulations of a generic

    green roof. Such a model would allow simulations to be performed on a variety of design

    alternatives, allowing intelligent design decisions to be made without the high cost of obtaining

    experimental data on a given design alternative. In many fields of study, computer modeling

    and simulation has proven to be a low-cost means of obtaining a wealth of information quickly

    and efficiently. What would previously require considerable effort and cost in construction,

    instrumentation, and time for monitoring data, can now be performed and analysed in minutes

    on a typical computer of today. In order to perform this work, however, a numerical model must

    be developed and fully validated under a variety of circumstances.

    400 Urban Ecosyst (2008) 11:399408

  • 8/13/2019 tesis impacto de cubiertas verdes en habitacin

    3/10

    Although this study does not develop a general-purpose model of a green roof for use with

    energy-based building simulation, it does demonstrate the ease in which a great deal of data

    may be obtained efficiently and easily using a building energy simulation tool, Environmental

    System Performance program (ESP-r; Energy Systems Research Unit (ESRU)). In this study, a

    rooftop energy balance model for a green roof is integrated into an existing full building-energy simulation tool in order to determine the performance of a simple green roof model.

    In a study conducted by Gaffin et al. (2005, 2006), based on experimental results

    obtained from test buildings at Pennsylvania State University (DeNardo 2003), a rooftop

    energy balance model was used to model the thermal characteristics of a green roof. The

    evaporative heat loss from a green roof, or latent heat, was modeled as proportional to the

    convective heat transfer from the rooftop surface to the surrounding air. This factor of

    proportionality, the Bowen ratio, is the ratio of the sensible heat flux to the latent heat flux.

    Thus the latent heat flux from a green roof, may be expressed as:

    Qlatent Qconvectionb

    whereis the Bowen ratio. For a control roof, this term is assumed to be zero. Unlike the

    previous study, however, no equilibrium conditions have been assumed, resulting from the

    ability of ESP-r to simulate transient heat transfer. Furthermore, the entire test-building was

    modeled in ESP-r, rather than just the roof.

    The base-case model was then extended to model a larger warehouse-style building in a

    Toronto climate. Various sized buildings were simulated, with varying internal loading, in

    order to determine various trends in the overall green roof energy savings over a

    conventional roof design. It is important to note that not all components of a green roof,such as heat storage and insulation from the soil medium, were accurately modeled, as the

    green roof improvements were captured simply by adding the latent heat flux term

    containing the pre-determined Bowen ratio value.

    Implementation in ESP-r

    ESP-r is a general-purpose, transient building simulation tool developed at the University of

    Strathclyde in Glasgow, Scotland, which accurately models the heat, air, moisture, and

    electrical power flows of a building under the control of a plant system (ESRU). This multi-domain problem is solved at user-configurable time-steps. ESP-r is an open-source project,

    which allows the implementation of new features, as well as a full understanding of the

    theory and assumptions made in the simulation. Further information on the theoretical

    formulation of ESP-r is documented by Clarke (Clarke 2001).

    In order to perform the present study, the simple one-dimensional model employed in

    (Gaffin et al. 2005,2006) was integrated in ESP-r. A full validation was performed using

    data obtained from (Gaffin et al. 2005, 2006), as well as experimental data used in

    (DeNardo 2003). Several complications arose in this comparison, namely due to

    inconsistent climate data. A larger set of climate data was required for the full building

    model used by ESP-r, as compared to the one-dimensional model of the roof. Not all of the

    required data was logged at the Pennsylvania State test site. As a result, it was necessary to

    obtain some of the required solar radiation data from logging performed at a nearby site.

    This inconsistency, as well as the inability to obtain full thermal properties of all portions of

    the building modeled, proved problematic, and many assumptions were made in the

    comparison. The building modeled in this study maintains only the same roof properties as

    Urban Ecosyst (2008) 11:399408 401

  • 8/13/2019 tesis impacto de cubiertas verdes en habitacin

    4/10

    those used in the Pennsylvania State test building, which were in fact known. The wall

    properties were changed to account for requirements of the new building design, discussed later.

    Furthermore, the climate data used was that for the city of Toronto (43 N, 79 W). This climate

    is similar to the Pennsylvania State (40 N, 77 W) climate, and all quantities in this data set are

    in fact consistent, eliminating the problems previously discussed. The climate for Toronto andPennsylvania State are shown in Tables1 and2, respectively (Weatherbase2007).

    The convection model employed was the same as used in (Gaffin et al. 2005, 2006),

    based on a model by Terjung and ORourke (Terjung and ORourke1980b), but modified

    at low wind speeds to prevent the convection term from vanishing.

    Qconvection g1 u0:8 TsurfTair ; u >1:75

    Qconvection +2 Tsurf Tair ; u e1:75

    In the above,g1and g2were set to 6.5 and 10.3, respectively, as obtained for the control

    roof using statistical analysis (Gaffin et al. 2005,2006). Convective heat-transfer from the

    green roof was assumed to be similar to that of the control roof. Any changes in heat-

    transfer for the green roof will be caused by the latent transfer, which as previously

    discussed, is proportional to the convective transfer.

    The zone air temperature was assumed well-mixed and held constant at 22C through

    the use of an ideal air-conditioning and heating model. A control time-step of 2 min was

    chosen to ensure that the zone air temperature remained constant at all times. As the

    simulations were performed during the month of July, the focus of analysis was on total

    cooling loads. It should be noted that on some nights, a small amount of heating wasrequired to maintain this fixed internal air temperature. The effect of heating on building

    energy consumption was ignored in this study.

    Model of a simple warehouse

    A simple warehouse building was modeled using ESP-r. As previously noted, the roof

    properties of the green roof used were identical to that of the Pennsylvania test site to

    maintain compatibility with the parameters used in the model. All other properties were

    Table 1 Average climate for Toronto, ON

    Quantity Units Annual Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

    Average temperature C 7 6 5 6 12 17 21 20 15 8 3 2

    Average high

    temperature

    C 12 2 1 3 11 18 22 26 25 20 13 6

    Average low

    temperature

    C 2 9 9 4 1 7 12 15 14 10 3 6

    Average precipitation Cm 76 4 4 5 6 6 6 7 8 7 6 6 6Average relative

    humidity (morning)

    % 83 83 83 82 78 76 78 79 86 89 87 86 84

    Average relative

    humidity (evening)

    % 64 76 73 68 58 55 54 54 57 61 65 74 78

    All simulations in this study were performed using climate data from Toronto. A summary of the Toronto

    climate is provided for comparison with State College, PA, USA

    402 Urban Ecosyst (2008) 11:399408

  • 8/13/2019 tesis impacto de cubiertas verdes en habitacin

    5/10

    modified, but were consistent between the green-roofed building and the control building.

    For all cases, the height of each story was fixed to 3.0 m, and windows were assumed to

    occupy 10% of the total area of each wall, except the north wall, which contained a single

    door.

    Table 3 shows the material properties used in the simulated buildings. The individual

    materials for each construction are specified from the outside to the inside. With the

    exception of the roof, which is based on Pennsylvania test site constructions, the properties

    were obtained from the ESP-r construction database. As a result, some of the materials usedmay not be familiar to North American readers.

    It is important to note that the green roof model does not contain any material for

    increasing the thermal capacitance. The sole effect of the green roof is obtained through the

    latent heat loss term, resulting from water vapour evaporation. Further experimental

    comparisons should be performed to examine the effect of including thermal capacitance in

    the model. Since the soil medium employed at the test site studied was quite thin, the

    overall effect is expected to be small. Although simple to add in the simulations performed,

    this was not modeled to maintain consistency with the parameters obtained at the

    Pennsylvania State test site.

    The study was performed for the month of July in a Toronto climate. It should be notedthat the chosen Bowen ratio, 0.12, is highly dependent on the green roof itself, as well as

    the climate in which the roof is implemented. The previous study (Gaffin et al. 2006) used

    Bowen ratios ranging from 0.21 to 0.35. The lower value was chosen as the simulated

    model is already somewhat conservative in nature. As mentioned previously, no thermal

    capacitance was modeled for the roof. Also, the model calculates the temperature of the

    roof-top surface, not the temperature below the soil surface in contact with the roof

    insulation, which is slightly cooler. The combination of the above factors justifies the

    selection of 0.12 for the Bowen ratio. Although the Bowen ratio employed was determined

    for a Pennsylvania climate, the Toronto climate should result in similar values. A Toronto

    climate was chosen for several reasons. First, and most important, is the fact that the Penn-

    state data obtained was pieced together from several sources. Inconsistencies in data sets

    proved problematic in ESP-r simulations. In addition, the Toronto climate proved of

    immediate interest for other research, although it would be beneficial to repeat the study

    under other climates. Recall, however, that the Bowen ratio is strongly dependant on

    humidity, and other climatic factors.

    Table 2 Average climate for State College, PA, USA

    Quantity Units Annual Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

    Average temperature C 9 2 2 2 8 14 19 21 20 16 10 4 1

    Average hightemperature

    C 15 1 2 7 14 21 25 27 26 22 16 8 2

    Average low

    temperature

    C 4 6 6 2 3 8 13 16 15 11 5 5

    Average precipitation cm 98 7 6 8 8 10 10 9 8 7 7 6 6

    Average relative

    humidity

    % 74 77 73 72 68 72 72 74 76 76 73 75 77

    Previous studies developed a green roof model using a test site in Pennsylvania. Due to inconsistencies in

    available PA data, and a requirement for results in the city of Toronto, simulations were conducted for a

    Toronto climate

    Urban Ecosyst (2008) 11:399408 403

  • 8/13/2019 tesis impacto de cubiertas verdes en habitacin

    6/10

    Roof

    envelope ratio analysis

    In order to assess the impact of roofenvelope ratio on overall energy savings resulting

    from a green roof, a variety of simulations were performed on the previously described

    warehouse-type building. One, two, and three story buildings were examined in order to

    determine the drop in energy savings with increasing building height. Building-envelope

    ratio was varied by changing the length of each wall, and the internal loading was assumed

    to be zero, unless noted otherwise. The benefit of a green roof over a conventional-roof was

    obtained by comparing the total building cooling loads required for each case. The

    percentage savings is defined as the ratio of the difference in cooling load required for a

    green roof to the total cooling load of the control roof.

    Figure1shows the total cooling load for both green and control roofs as a function of

    building roofenvelope ratio, for a constant internal loading of both 0 and 5,000 W. It

    should be noted that the latter load employed is somewhat low for the building sizes

    investigated in this study. Nevertheless, the increase in cooling load required is still

    significant, and provides further evidence of the advantage gained through the use of a

    Table 3 Summary of simulated building material properties

    Thickness Conductivity Density Specific

    heat

    Longwave

    emissivity

    Solar

    absorptivity

    Diffusion

    resistance

    (m) (W/mC) (kg/m3) (J/kgC)

    Wall

    Brick 0.100 0.960 2000 650 0.90 0.70 25

    Glasswool 0.075 0.040 250 840 0.90 0.30 4

    Air 0.050 0.000 0 0 0.99 0.99 1

    Breeze block 0.100 0.440 1500 650 0.90 0.65 15

    Roof

    Aluminum 0.003 210.0 2700 880 0.22 0.20 19200

    Air 0.025 0.000 0 0 0.99 0.99 1

    Glass fiber quilt 0.080 0.040 12 840 0.90 0.65 30

    Aluminum 0.003 210.0 2700 880 0.22 0.20 19200

    Green roofPlywood 0.019 0.200 560 1000 0.90 0.70 576

    Fiberglass 0.089 0.049 300 1000 0.90 0.50 5

    OSB 0.006 0.110 600 1210 0.80 0.65 12

    Ground floor

    Earth 0.250 1.280 1460 879 0.90 0.85 5

    Gravel 0.150 0.520 2050 184 0.90 0.85 2

    Heavy-mix concrete 0.150 1.400 2100 653 0.90 0.65 19

    Air 0.050 0.000 0 0 0.99 0.99 1

    Chipboard 0.019 0.150 800 2093 0.91 0.65 96

    Wilton 0.006 0.060 186 1360 0.90 0.60 10

    Suspended floor

    Wilton 0.006 0.060 186 1360 0.90 0.60 10

    Chipboard 0.019 0.150 800 2093 0.91 0.65 96

    Air 0.050 0.000 0 0 0.99 0.99 1

    Heavy-mix Concrete 0.140 1.400 2100 653 0.90 0.65 19

    Steel 0.004 50.000 7800 502 0.12 0.20 19200

    Roof properties used in the simulation are identical to those of the Pennsylvania test site. All other properties

    used in the simulation are taken from the ESP-r construction database

    404 Urban Ecosyst (2008) 11:399408

  • 8/13/2019 tesis impacto de cubiertas verdes en habitacin

    7/10

    green roof. The reduction in cooling load for the 5,000 W case decreases steadily for both

    the green and control roof due to the increased total building size required to obtain larger

    roofenvelope ratios. What is more important is the increase in savings obtained by a green

    roof at these higher ratios.

    Figures2and 3extend these findings to a two and three story building, respectively. In

    addition, these figures break down the cooling load required into that required to cool only

    the upper floor of the building, as well as that required to cool the total building. Again, for

    small roofenvelope ratios, the green roof provides no benefit, as expected. At the oppositeend of the spectrum, however, the reduction in cooling load with the use of a green roof is

    significant. More importantly, it becomes clear how important a roof is in dictating the

    overall energy load of a building. As is clearly shown in Figs. 2and3, the cooling load of

    the upper floor alone approaches the total cooling load at higher roofenvelope ratios. In

    fact, above a roofenvelope ratio of 0.7 for a two-story building, the upper floor value

    equals the total building cooling load. This effect is slightly less pronounced in the three-

    story design. In both cases, the increase in cooling load with increasing roofenvelope ratio

    Fig. 1 Variations in cooling load

    for a single-story building with

    the control and green roofs. The

    simulations were performed with

    a modified version of the ESP-r

    model. Each simulation is run

    with a constant internal loadingof 0 and 5,000 W, with 0 repre-

    senting a baseline situation where

    cooling is the only demand for

    electricity in an unoccupied

    building

    Fig. 2 Variations in cooling loadfor a two-story building with the

    control and green roofs. Cooling

    load is shown for the entire

    building, and also for the upper

    floor alone, demonstrating that

    for large roofenvelope ratios,

    cooling of the upper floor

    accounts for the entire building

    cooling load. Green roofs provide

    significant savings over the con-

    trol roof with increasing roof

    envelope ratios

    Urban Ecosyst (2008) 11:399408 405

  • 8/13/2019 tesis impacto de cubiertas verdes en habitacin

    8/10

    is due to the larger building sizes of the high roofenvelope ratio buildings. Of more

    importance is the energy savings of the green roof case over the conventional roof design.

    Figure 4 summarizes the percentage savings in cooling load of a green roof over a

    conventional roof for all cases previously considered. The reduction in performance for

    multi-story buildings quickly becomes apparent, although the gain in performance for high

    roofenvelope ratios is significant in all cases.

    The results obtained for the single story building represent an extreme case. The highroofenvelope ratio was obtained for a building of 250 250 m. Although this is not

    uncommon with some warehouse building constructions, the convection model employed

    may not provide accurate results for a rooftop of this size. Further studies should investigate

    the use of other convection models in order to fully understand the flow over such a large

    surface. Furthermore, a building with no internal loading does not represent a practical case.

    The next section will deal with the addition of internal loading, and the overall effect on

    reducing the performance gain of a green roof.

    Fig. 3 Variations in cooling load

    for a three-story building with the

    control and green roofs. Cooling

    load is shown for the entire

    building, and also for the upper

    floor alone. Slightly larger roof

    envelope ratios are required forthe upper floor cooling load to

    reach the total building cooling

    load, when compared to the two-

    story building

    Fig. 4 Percentage savings in

    building energy usage of greenroof design over the control roof

    for a one, two, and three-story

    structure. Overall savings from a

    green roof are reduced for an

    increasing number of stories, but

    become more substantial for

    higher roofenvelope ratios

    406 Urban Ecosyst (2008) 11:399408

  • 8/13/2019 tesis impacto de cubiertas verdes en habitacin

    9/10

    Internal load analysis

    In addition to examining the impact of roofenvelope ratio, the effect of internal loading

    was also examined. For several representative roofenvelope ratios, the internal loading

    was varied from 0 to 50,000 W, a load more representative of buildings of this dimension.

    Figure5 shows the effect of increasing the internal load for each of the test buildings, with

    a roof area of 100 100 m. As shown in the plot, the net energy savings increases

    approximately linearly with increased internal loading. This increase, however, reaches amaximum value, after which an increase in internal loading results in a drop in net energy

    savings. For a 100100 m square roof, the effect is not noticed for two and three story

    buildings.

    Similar plots may be generated for larger roof areas. Although the energy savings are

    larger, notably a net energy saving of up to 13,000 kW h for the three story building with a

    roof area of 250250 m, the plots are considerably more flat, indicating the reduced effect

    of total internal load on such a large building. Furthermore, the peak in energy savings is

    not present in the case of larger buildings.

    Fig. 5 Net energy savings

    obtained from a green roof as a

    function of internal loading for a

    one, two, and three-story ware-

    house structure measuring 100

    100 m. Energy savings increase

    approximately linearly with in-creasing internal load. A maxi-

    mum value is shown in the

    single-story case, after which net

    energy savings decrease with in-

    creasing internal load

    Fig. 6 Percentage energysavings for a single-story with

    varied internal loading. Increased

    internal loading results in an

    overall reduction in energy

    savings, although savings are still

    substantial

    Urban Ecosyst (2008) 11:399408 407

  • 8/13/2019 tesis impacto de cubiertas verdes en habitacin

    10/10

    Figure 6 demonstrates the percentage energy savings of a single story building with

    varying internal loading. As shown in the figure, more realistic values for energy savings

    are obtained when realistic internal loading is assumed. The values approaching 100%

    obtained for all cases with 0 W internal loading are clearly not obtainable in a practical

    structure. For loadings more typical of a practical structure, however, the savings are stillsignificant. A 250250 m green-roof design with 50,000 W internal loading was found to

    have percentage energy savings of 73%, 29%, and 18%, for a one, two, and three-story

    design, respectively.

    In addition to addressing the issue of improved convection modeling, future studies

    should investigate several further issues. This model accounts for the evapotranspiration

    present on a green roof. The model does not directly account for changes in convection

    resulting from the green roof. Furthermore, a green roof with a larger mass should also be

    studied using the present model, in order to investigate the effect of thermal capacitance on

    energy savings. The present study has assumed a constant value for the Bowen ratio.

    Additional studies should identify the Bowen ratio as a function of climate conditions, as

    well as any additional required green roof parameters. This would allow an investigation of

    green roof performance at other sites and under other climates.

    References

    Clarke JA (2001) Energy simulation in building design, 2nd edn. Butterworth-Heinemann, Oxford

    DeNardo J (2003) Green roof mitigation of stormwater and energy usage, Masters Thesis. Pennsylvania

    State University, Centre County

    Eumorfopoulou E, Aravantinos D (1998) The contribution of a planted roof to the thermal protection ofbuildings in Greece. Energy Build 27:2936

    Gaffin S, Rosenzweig C, Parshall L, Beattie D, Berghage R, OKeefe G, Braman D (2005) Energy balance

    modeling applied to a comparison of white and green roof cooling efficiency. Third Annual Greening

    Rooftops for Sustainable Communities Conference, Awards and Trade Show, Washington, D.C.

    Gaffin S, Rosenzweig C, Parshall L, Hillel D, Eichenbaum-Pikser J, Greenbaun A, Blake R, Beattie D,

    Berghage R (2006) Quantifying evaporative cooling from green roofs and comparison to other land

    surfaces. Fourth Annual Greening Rooftops for Sustainable Communities Conference, Boston

    Niachou A, Papakonstantinou K, Santamouris M, Tsangrassoulis A, Mihalakakou G (2001) Analysis of the

    green roof thermal properties and investigation of its energy performance. Energy Build 33:719729

    Palomo del Barrio E (1997) Analysis of the green roofs cooling potential in buildings. Energy Build 27:179193

    Terjung WH, ORourke PA (1980a) Simulating the causal elements of urban heat islands. Boundary Layer

    Meteorology 19:13118

    Terjung WH, ORourke PA (1980b) Energy exchanges in urban landscapes: selected climatic models.

    Publications in Climatology. vol. 33(1). University of Delaware, Center for Climatic Research, Newark

    Weatherbase (2007) Canty and Associates LLC. Accessed athttp://www.weatherbase.com

    Wong NH, Cheong DKW, Yan H, Soh J, Ong CL, Sia A (2002) The effects of rooftop garden on energy

    consumption of a commercial building in Singapore. Energy Build 35:353364

    408 Urban Ecosyst (2008) 11:399408

    http://www.weatherbase.com/http://www.weatherbase.com/