Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Scaling urban energy use and greenhouse gas emissions through LiDAR van der Laan, Michael Tije 2011

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2012_spring_vanderlaan_michael.pdf [ 15.63MB ]
Metadata
JSON: 24-1.0072500.json
JSON-LD: 24-1.0072500-ld.json
RDF/XML (Pretty): 24-1.0072500-rdf.xml
RDF/JSON: 24-1.0072500-rdf.json
Turtle: 24-1.0072500-turtle.txt
N-Triples: 24-1.0072500-rdf-ntriples.txt
Original Record: 24-1.0072500-source.json
Full Text
24-1.0072500-fulltext.txt
Citation
24-1.0072500.ris

Full Text

Scaling Urban Energy Use and Greenhouse Gas Emissions through LiDAR by Michael Tije van der Laan B.EnD. (Hons), The University of British Columbia, 2009  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in The Faculty of Graduate Studies (Geography)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) December 2011 © Michael Tije van der Laan 2011  Abstract Although models to quantify CO2e emissions in urban areas exist, they are within isolated disciplines, and are targeted at specific scales, emissions processes, and end-users — not a priori compatible with planning needs. Furthermore, the majority of existing models rely on inventory data, which is typically only available at aggregate space and time scales. It is necessary however, that neighborhood-scale CO2e emissions estimates are provided to determine the key relationships between urban form and emissions — which can than be applied to future planning strategies. This this thesis developed a new methodology to integrate LiDAR data, building simulation software and a building typology database to rapidly model energy and emissions for a large number of buildings. To adjust building energy demand to local urban-context, building morphology, and population density a scaling approach is proposed. This methodology was applied to a study area of 7.4 km2 in Vancouver, BC, consisting of 7812 buildings ranging in moderate to high density. Modeled building energy use in this transect was sensitive to local conditions (average variation in building energy use due to urban-context 2.8%, building morphology 2.8%, and population density 3.2%) resulting CO2e emissions of 14.2 kg CO2e m-2 yr-1 (1309 kg CO2e Inh.-1 yr1  ) varying dramatically between the central business district (40.1), mixed-  use (12.7), and residential (9.0) neighbourhoods. Spatial and temporal patterns of building energy use, CO2e emissions and anthropogenic heat release by buildings are presented and discussed in relation to urban form.  ii  Table of Contents  1  Abstract  ii  Table of Contents  iii  List of Tables  v  List of Figures  vi  List of Abbreviations  viii  Acknowledgements  ix  Dedication  x  Introduction 1.1  Cities and energy  1  1.2  The urban environment  3  1.2.1  1.3  Building energy use 1.3.1 1.3.2 1.3.3 1.3.4  1.4  6 6 8 9 10  Modeling methodologies  10  Inventory methodology Direct measurement methodology Building energy modeling methodology  Research goals 1.5.1 1.5.2  2  3  Urban context Building morphology Archetype Occupant behavior  1.4.1 1.4.2 1.4.3  1.5  Characterization  Research aims Research objectives  10 11 12  13 13 14  Materials and Methods 2.1 2.2  Study site  16  Geospatial integration  19  2.2.1 2.2.2 2.2.3  2.3  Spatial data Aspatial data Building threshold and the ‘edge effect’  An archetype approach 2.3.1 Overview 2.3.2 Typology structure 2.3.3 Elements database  2.4  Building-scale population down-scaling  20 21 22  24 24 25 26  28 iii  2.5  Building-scale morphology extraction 2.5.1 Building morphology defined 2.5.2 Deriving building morphology  2.6  Urban context extraction 2.6.1 Urban context defined 2.6.2 Sky view factor ySVF 2.6.3 Fractional shading  2.7  Modeling building energy use 2.7.1 2.7.2 2.7.3 2.7.4  3  30 31  33 33 33 35  37 38 40 43 45  Results and Discussion 3.1  3.2  4  Heat balance External radiation Shading module Hot water demand  30  Scaling approach  47  3.1.1 Set-up 3.1.2 Influence of the urban context 3.1.3 Influence of building morphology 3.1.4 Influence of population density 3.1.5 Up-scaling energy and emissions  48 50 55 63 65  Upscaled energy and emissions  68  3.2.1 Modeled energy use 3.2.2 Modeled CO2e emissions 3.2.3 Modeled QFB 3.2.4 Comparing modeled and measured 3.2.5 Scaling sensitivity  68 74 81 86 87  Conclusion 4.1  Practical significance   90  4.2  90  Recommendations for future work  References  93  Appendices A.1 Maps  101  A.2 Morphometric QFB plots  110  iv  List of Tables Table 2-1  Neighbourhood metrics  19  Table 2-2  Property assessment statistics  22  Table 2-3  Archetype attributes: Part 1  27  Table 2-4  Population regression output  29  Table 3-1  Archetype attributes: Part 2  50  Table 3-2  Energy use intensity EUIV vs. sky view factor  53  Table 3-3  Energy use intensity EUIV vs. fractional shading  54  Table 3-4  Energy use intensity EUIV vs. volume  60  Table 3-5  Energy use intensity EUIV (space heating) vs. volume  61  Table 3-6  Energy use intensity EUIV vs. characteristic length LC  62  Table 3-7  Energy use intensity EUIV vs. population density  65  Table 3-8  Archetype attributes: Part 3  67  Table 3-9  Neighbourhood scaling sensitivity  68  Table 3-10  Building energy use  74  Table 3-11  Building CO2e emissions  81  Table 3-12  Anthropogenic heat flux case studies  82  Table 3-13  Anthropogenic heat emissions  86  Table 3-14  Energy and emissions comparison  87  v  List of Figures Figure 1-1  Canadian energy use trends (1990-2008)  2  Figure 1-2  LiDAR products example  5  Figure 1-3  Thesis overview diagram  15  Figure 2-1  Study site  17  Figure 2-2  Subset overview maps  18  Figure 2-3  Geospatial integration  20  Figure 2-4  Building ‘edge effect’  23  Figure 2-5  Single detached dwelling rank order  23  Figure 2-6  Typology vs LiDAR attributes  27  Figure 2-7  Obstruction angle  34  Figure 2-8  Sky view factor calculation  35  Figure 2-9  Azimuth-elevation matrix  37  Figure 2-10  Heat balance components  40  Figure 2-11  Overlapping shadows  43  Figure 2-12  Solar radiation reflections  44  Figure 3-1  Simulation strategy  48  Figure 3-2  Shading array example  49  Figure 3-3  Energy use intensity EUIV vs. sky view factor  53  Figure 3-4  Energy use intensity EUIV vs. fractional shading  54  Figure 3-5  Building morphology: footprint AF vs. volume V  57  Figure 3-6  Building morphology: depth vs. footprint  58  Figure 3-7  Energy use intensity EUIV vs. volume  60  Figure 3-8  Energy use intensity EUIV (space heating) vs. volume  61  Figure 3-9  Energy use intensity EUIV vs. characteristic length LC  62  Figure 3-10  Energy use intensity EUIV vs. population density  64  Figure 3-11  Modeled building energy use map  70  Figure 3-12  Transect raster of thermal energy  71  Figure 3-13  Transect plot of energy use  72  Figure 3-14  Plot of thermal and thermal energy  72  Figure 3-15  Neighbourhood energy density  74  Figure 3-16  Transect raster of CO2e emissions  77  Figure 3-17  Transect raster of CO2e emissions (per capita)  78  Figure 3-18  Transect plot of CO2e emissions  79  Figure 3-19  Transect plot of CO2e emissions (per capita)  79  vi  Figure 3-20  Neighbourhood CO2e density  81  Figure 3-21  Transect raster of QFB emissions  84  Figure 3-22  Transect plot of QFB emissions  85  Figure 3-23  Neighbourhood QFB density  85  Figure A-1 Transect aerial map  102  Figure A-2 Transect figure ground  103  Figure A-3 Transect land use map  104  Figure A-4 Transect obstruction angle map  105  Figure A-5 Transect digital elevation model  106  Figure A-6 Transect raster of sky view factor  107  Figure A-7 Transect raster of fractional shading  108  Figure A-8 Transect raster of population density  109  Figure A-9 Anthropogenic heat flux density vs. Building volume  110  Figure A-10 Anthropogenic heat flux density vs. Building height  111  Figure A-11 Anthropogenic heat flux density vs. Sky view factor  111  Figure A-12 Anthropogenic heat flux density vs. Population density  112  Figure A-13 Anthropogenic heat flux density vs. Fractional shading  112  Figure A-14 Anthropogenic heat flux density vs. Plan area  113  vii  List of Abbreviations BEM:  Building energy model  CEEI:  Community energy and emissions inventory  DA:  Dissemination area  DEM:  Digital elevation model  DHW:  Domestic hot water  DT:  Downtown study area  EUIV,FA,LA : Energy use intensity (per building volume, per floor area, per gross land area) DES:  District energy system  GHG: Greenhouse gas GIS:  Geographical information system  HDD:  Heating degree day  LiDAR: Light detection and ranging LCZ:  Local climate zone  MP:  Mount Pleasant study area  SS:  Sunset study area  SDD:  Single detached dwelling  TEUILA: Thermal energy use intensity (per gross land are) UT:  Urban transect study area  UBL:  Urban boundary layer  UCL:  Urban canopy layer  UHI:  Urban heat island  viii  Acknowledgements This thesis would not be possible without the guidance, insight and support of Dr. Andreas Christen, our meetings were always enlightening. I am grateful for the many opportunities provided to me which allowed for a fulfilling research experience. Thanks also to Prof. Ronald Kellett, for your advice, invaluable feedback and continuous support, I have benefited enormously from your scope of expertise. Special thanks to Rory Tooke and Ben Crawford for the support, data and perceptive conversation, providing a glimpse into the future of the research field. Thanks to Prof. Cynthia Girling for providing invaluable work experience and a vibrant environment to undertake it. Thanks also to Jessica Webster at NRCan’s CANMET Energy Technology Centre for providing EcoEnergy retrofit data, to Scott Albrechtsen and Dennis Nelson from BC Hydro for providing modeled electricity data. To Dr. Timothy Oke, Dr. Nicholas Coops, Dr. James Voogt, Scott Krayenhoff, Carmen Emmel, Inna Olcholvski, Eli Heyman, Phil Riley and David Peacock for providing guidance, feedback and fruitful discussion. Thanks to my Mom and Dad, I am who I am today because of you, thanks for being my biggest fans. Essential funding and support of this research was provided by the Canadian Foundation for Climate and Atmospheric Sciences (specifically the EPiCC network), elementsdb, and Natural Resources Canada.  ix  Dedication To Andrea, a smart, funny and supportive partner.  x  Introduction  1  Introduction  This section will first introduce energy use patterns at the global, national and municipal scales, noting the trend in energy use and influence on global climate change. It will then discuss our current understanding of the influence of urban form on building energy demand, highlighting the current methodologies for characterizing urban form and modeling energy use at the urban-scale. It will conclude with thesis research objectives, aimed at integrating LiDAR data into urban form characterization for the purposes of building energy and emissions modeling.  1.1  Cities and energy  Urbanization has lead to the modification of both local and regional climate, drastically changing the supporting landscape. In 2011, the world’s population reached seven billion, over half of which now live in cities (UN, 2011; IPCC, 2007). Additionally, global reliance on fossil-fuels has dominated the world energy supply, accommodating 81% of primary energy (IEA, 2011a). Although the share of fossil-fuels in the world energy supply has decreased marginally since 1973 — from 87% in 1973 to 81% in 2009 — world energy use has subsequently doubled from 6111 Mtoe to 12150 Mtoe during the same time period (IEA, 2011b). This has lead to the release of substantial amounts of greenhouse gases (GHG1), contributing to global climate change. As a consequence, average global temperatures are projected to rise between 1.8 and 6.4 K by 2100 (IPCC, 2007). It is estimated that at least 30-40% of global anthropogenic GHG emissions are directly attributable to cities and as much as 80% indirectly (Satterthwaite, 2008). Furthermore, the magnitude of GHG emissions in a city is greatly influenced by urban form choices, including density, land  1  Change in radiative forcing of the troposphere and hence contributing to global climate change. 1  Introduction  use mix, spatial pattern and building type. When considered alongside the global warming potential of GHG released during fossil fuel combustion, global reliance on them, and the nearing peak in oil production (Aleklett et al., 2010; deAlmeida et al., 2009), efficient low-energy cities that utilize urban form are an important part of GHG emissions-reduction strategies. If we narrow our scope to energy use trends in the Canadian building stock (31% of total National energy use in 2008), we can follow the progress of energy efficiency related to urban form choices (Figure 1-1). In the Residential-sector we can see that energy intensity (per floor area) is dropping, however, this coincides with an increase in floor area per household (and per person), which offsets much of the efficiency gain. In contrast, the Commercial-sector has seen an increase in total floor area per person but unlike the Residential-sector there has been no decrease in energy intensity (per floor area) and therefore a substantially larger increase in total energy use.  Year  08  06  20  04  20  02  20  00  20  98  20  96  -30%  19  08  06  20  04  20  02  20  00  20  98  20  96  19  94  19  19  19  92  -30%  -20%  90  Energy use / m2  Energy use / m2  19  People / Household  -20%  0% -10%  94  Energy use / Person  Floor area / Person  19  0% -10%  Energy use / Person  10%  19  Floor area / Household  Total energy use  20%  92  Total energy use  10%  30%  19  Floor area / Person  20%  19  Commercial Buildings Percent change from 1990-2008  30%  90  Percent change from 1990-2008  Residential Buildings  Year  Percent change from 1990-2008  Figure 1-1 Canadian energy use trends (1990-2008) Transportation Percent change in Residential and Commercial Building Sectors from 1990 levels. Energy use / m240% is per gross floor area. Data source: Natural Resources Canada Energy Use Data Handbook 30%  Passenger km  20%  Population  10% In the city of Vancouver, Energy BC,/ kmCanada (where this thesis will focus) the 0% Building-sector alone is responsible for 53% of total GHG emissions (reported -10% -20%  08  20  06  20  04  20  02  20  00  20  98  19  96  19  94  19  92  19  19  90  -30%  Year  2  Introduction  in CO2e1), with 77% of these emissions originating from Residential and Commercial buildings (BCa, 2008). While a general pattern of Buildingsector emissions can be mapped according to population, this attribution does not account for local differences in building type and urban form. For instance, access to the sun and natural ventilation, both key components in building energy consumption, are strongly influenced by building shape, orientation and surrounding urban form. Although methods exist to reduce GHG emissions, such as technology advancements, fuel switching, and behavioral changes, stabilizing emissions is likely to be a complicated task. Furthermore, energy conservation alone, although crucial, is insufficient for meeting challenging emission reduction aspirations and targets. It is therefore important we consider urban form when developing appropriate decision support methods for evaluating the impact, and opportunity of future planning scenarios. However, the relationship between urban form and energy use is not yet well defined. This has been in large part due to the difficulty in characterizing urban form, a task that has been both time consuming and labor intensive.  1.2  The urban environment  1.2.1  Characterization  Over the last 50 years, land use planning has brought about major computation advancements in urban characterization. In 1960, a geographical information system (GIS) was introduced by the Canadian Government that could store, analyze and manipulate data from the Canadian land inventory (Coppock and Rhind, 1991). Since then, vast improvements have been made in geospatial tools, greatly increasing the resolution at which we can spatialize and integrate multiple data sets. 1  The carbon dioxide equivalency describes, for a given mixture of GHG, the amount of CO2 with the same global warming potential (e.g. including effects of CH4 and N2O). 3  Introduction  It is now the norm for municipalities to upkeep a variety of overlapping cadastral data sets. Many of these data sets — such as Vanouver’s OpenData Catalogue — are making their way onto publicly accessible online applications (City of Vancouver, 2011). The amount and type of data available varies between municipalities and may include building parcel boundaries, roadways and other landuse features. Furthermore, cadastral data can be combined with aspatial data sources, such as property assessments and census statistics. Obtaining particular data sets however, or merging data sets that span several political boundaries remains a challenge. Light detection and ranging (LiDAR), a recent advancement in three-dimensional surface characterization, provides promise where manual classification of urban form attributes is difficult. The use of remotely sensed LiDAR in urban areas is a relatively new field of study. LiDAR makes possible the automatic classification of urban form attributes, capturing not only the spatial distribution of city elements (e.g. buildings and trees) but also their morphology — for example height, area, and volume (Figure 1-2). Products that can be output from remotely sensed LiDAR, such as city digital elevation models (DEM), provide a promising input for building morphology analysis (Neidhart and Sester, 2004; Steadman et al., 2009; Tooke et al., 2011), energy and urban form analysis (Ratti et al., 2002, 2005, 2006; Yu et al., 2009; Lindberg and Grimmond, 2011), and when combined with satellite imagery, urban surface characterization (Goodwin et al., 2009; Tooke et al., 2009; Yu et al., 2010). Additionally, the automatic classification possible through LiDAR is advantageous over conventional survey methods which can be time consuming, laborious and subjective.  4  Introduction  Figure 1-2 LiDAR products example Three different neighbourhoods in Vancouver, BC, Canada; a) Light detection and ranging (LiDAR) derived figure ground, b) Horizon obstruction angle, and c) Digital elevation model.  5  Introduction  1.3  Building energy use  Building energy demand is the sum of space conditioning (heating and cooling), domestic hot water (DHW), lighting, appliance, and system equipment1 loads. The energy required for the above proposes can be supplied through a variety of passive and active sources. For example, conventional non-renewable sources include: oil, natural gas, coal, and nuclear. Alternatively, building energy can be provided by renewable sources such as biomass, wind, hydro, solar thermal and voltaic, geoexchange, geothermal and heat recovery (Hegger et al., 2008). The amount of energy consumed by a building depends on its use, materials, construction, system and equipment efficiencies, shape, local climate, and behavior of occupants. For the purposes of this research, these dependencies have been grouped accordingly: • Urban context: Urban form scale, configuration and local climate • Building morphology: Building shape and size • Archetype: Building use, construction and systems • Occupant behavior: Scheduling, demand and thermal comfort  1.3.1  Urban context  Urban morphology broadly refers to the study of the physical (or built) fabric of urban form, and the people and processes that shape it. If our focus is on building energy use, we see that energy demand is significantly influenced by the urban context surrounding a building.  In particular,  self-shading, vegetation, and the configuration and scale of surrounding obstructions influences the radiative and energy exchange of a building. Recent research has utilized DEMs (Ratti et al., 2003, 2005; Steemers, 2003, Cheng, 2009; Christen et al., 2011) and vector tracing analysis (Robinson,  1  Energy expended in the delivery (active) of space conditioning, DHW and ventilation. 6  Introduction  2006; Strømann-Andersen and Sattrup, 2011) to estimate the relationship between the urban context and solar access. In a similar fashion, solar guidelines have been developed to guide urban design by prioritizing a building’s access to solar radiation (Littlefair, 1998; Brown and DeKay, 2000; Knowles, 2003; Sarkar, 2009). Lastly, vegetation can both positively and negatively impact energy use, influencing economic (Akbari, 2002), energy (Simpson, 2002) and microclimate conditions (Shashua-bar et al., 2006). Essential to understanding the fluxes of energy exchange between buildings and the atmospheric environment a description of the atmospheric layers whose climates are affected by the surface characteristics of a city. Buildings occupy the urban canopy layer (UCL), which extends from the urban surface to the height of building elements (zH). This layer is driven by local processes and properties generated by the immediate surroundings, which are controlled by canyon1 geometry and surface material. The UCL is nested within the urban boundary layer (UBL), which extends from the top of UCL to a height where surface influences from the urban area become negligible. The UBL is a local and mesoscale phenomena that is governed by the surface characteristics of the urban area below (Oke, 1987). Cities when compared to their rural counterparts have decreased solar irradiance, increased precipitation, and increased temperatures (Taha, 1997; Oke, 1987). The UCL is on average warmer than its rural counterpart — a phenomena known as the urban heat island UHI (Oke, 1982). Factors that are known to contribute to the UHI are: reduced low level winds due to buildings, reduced evaporative cooling due to augmented surface properties, enhanced heat storage in pavement and buildings, altered radiation budgets due to topography and atmospheric pollutants. Furthermore, the radiation budget in a city is confounded by the three dimensional shape of buildings due to the increased surface area added  1  Building walls and elements between buildings. 7  Introduction  by building wall and roof surfaces. This increased surface area reduces the sky view factor (See Section 2.6.2)and promotes radiation trapping. Additionally, the release of heat from building energy consumption, traffic combustion, and human metabolism (anthropogenic heat flux, QF) can drive temperatures up in the urban environment. Increased UCL air temperatures can be unfavourable to human comfort. For instance, the ability to moderate indoor climate has lead to large amounts of energy being expended to maintain optimal comfort temperatures. As a result, many metropolitan areas have large air conditioner loads that deal with even slight rises in UCL air temperature. For example, in Tokyo, Japan Kikegawa and Kondo estimated that a 1K rise in air temperatures resulted in a 6.6% rise in energy use1 (Kondo and Kikegawa, 2003). This leads to an amplification loop where indoor heat is expelled out into the urban environment further driving up the demand for space cooling. That said, reductions in winter space heating demand is expected alongside increased heat waves in many cities and both should be considered in tandem when looking at potential adaptation and mitigation strategies (Davies et al., 2008).  1.3.2  Building morphology  The shape and size (morphology) of a building also have a significant influence on its energy demand. For example, building Energy use intensity2 (EUIV and EUIFA) is strongly influenced by the ratio of building volume to exposed surface area. This ratio is a measure of the building’s ability to retain/expel heat and utilize passive systems — garner solar gains and cross-ventilation potential (Gupta, 1987; Papadopoulos et al., 2002; Pessenlehner and Mahdavi, 2003; Shashua-bar et al., 2006). Furthermore, 1  Sensitivity to daily maximum summer temperatures (>22 K).  2  In this thesis EUI is presented in three metrics: a) per floor area EUIFA, b) per volume EUIV, and c) per gross land area EUILA. Arguably the most appropriate is per volume EUIV , which accounts for variation in floor heights between buildings and is a good proxy of the space to be heated (if underground areas and unconditioned areas are filtered).. 8  Introduction  building orientation, shared wall area and building depth also greatly influence energy demand (Steemers, 2003; Stupka and Kennedy, 2010). A building’s morphological dependence on climatic conditions has been shown to influence the shape of existing building stocks across a wide range of scales and archetypes. Specifically, geometric properties such as building height zH , depth LD, footprint AF, surface area AS and volume V have all been shown to scale allometrically — distorting in order to capture solar radiation gains (Steadman, 2006; Batty et al., 2008; Steadman et al., 2009).  1.3.3  Archetype  A coarse division of landuse into Residential, Commercial and Industrial sectors allows for a general description of energy use in a city. This pattern can be broadly defined for each of the three sectors: Residential has a daily double peak in energy use, corresponding to human activity in the morning and evening, Commercial a single peak in the afternoon, and the Industrial sector is relatively constant throughout the year (Grimmond, 1992; Lee et al., 2009). However, landuse alone is a poor indicator of the spatial heterogeneity of energy use and emissions in a city and does not capture the diversity or intensity of building types and urban form. Cities are in general made up of replicable patterns of development or cases which share similar attributes — an archetype. A building typology can be constructed from these archetypes and based on important indicators, for instance energy use. Classification of appropriate archetypes may involve field based measurements, energy audits or the adaptation of categories inherent in present data sets — such as property assessments or census data. Furthermore, building attributes such as heating source and equipment efficiency evolve over time and should be considered when research topics span different periods of construction or building types. Examples of energy studies that have employed an archetype approach include: March and Martin (1972), Ratti et al. (2003), Steemers (2003), 9  Introduction  Parekh (2005), Steadman (2006), Heiple and Sailor (2008), Salat (2009), Zizzo and Kennedy (2010), and Christen et al. (2011).  1.3.4  Occupant behavior  As energy use efficiency improves in the building-sector, through building code upgrades and building system efficiency gains, the role of the occupant becomes elevated (Santin et al., 2009). However, significant uncertainty as to what impact occupancy has on building energy still exists. This variability has provided a large challenge for energy modeling studies (Clevenger et al., 2006). In the Netherlands, occupancy has been shown to affect energy use by 4.2% alongside differences in building characteristics — which impact energy use by 42% (Guerra Santin et al., 2009). Tenancy has also been shown to have a significant impact on energy use, for instance in Canadian attached dwellings, EUIFA is 37% higher in buildings where the landlord pays for utilities instead of the occupant (Maruejols and Young, 2010). Other cultural and social intricacies impact energy use and only a general comprehension of the important indicators is known (Lutzenhiser, 1993).  1.4  Modeling methodologies  Several methodologies to estimate building-sector energy use exist. Sailor (2011) proposes three main methodologies: inventory (i.e. top-down), direct measurement and building energy modeling (i.e bottom-up). These categories provide a relevant framework for estimating building-sector energy use and emissions and will be discussed below.  1.4.1  Inventory methodology  Annual estimates of energy use and CO2e emissions are required for all municipalities in British Columbia (i.e. Bill 44 - The 2007 BC Greenhouse Gas Reduction Targets Act). This Community energy and emissions inventory  10  Introduction  (CEEI) is aggregated at a large temporal (annual) and spatial (municipal) scale and includes energy and fuel usage for building, transport and waste sectors. Although municipalities are now documenting annual consumption statistics, neighbourhood-scale emissions estimates are rare. Therefore, recent efforts have been directed towards the mapping of inventory data down to smaller scales (e.g. Sailor and Lu, 2004; VandeWeghe et al., 2007; Heiple and Sailor, 2008; Smith et al., 2009; Lee et al., 2009; Hamilton, 2009; Parshall et al., 2010; Raupach et al., 2010). One of the most widely accepted methodologies for down-scaling inventory data was introduced by Sailor and Lu (Sailor and Lu, 2004). In this study, state wide energy use was mapped down to the city-scale by applying seasonal functions for energy use and fitting per capita energy totals to the spatial patterns of population. In recent years, other studies have adopted similar methodologies, utilizing available inventory data, spatial patterns of population and property assessments. Despite the increasingly availability of inventory data, down-scaling alone, typically neglects urban form differences and is limited in its ability to inform context-sensitive policy change.  1.4.2  Direct measurement methodology  Micrometeorological techniques have also been used to directly measure the flux of CO2 (i.e. emissions uptake) in urban areas (Nemitz, 2002; Vogt et al., 2006; Moriwaki and Kanda, 2004; Crawford et al., 2009; Pataki, 2009; Christen et al., 2011). To directly measure CO2 fluxes in a neighbourhood, a micrometeorological tower is set up and the required fluxes are estimated using the eddy-covariance technique (Feigenwinter et al., 2012). This technique determines the covariance between the vertical wind component and CO2 concentration fluctuations from their mean (Baldocchi, 2003). The wind fluctuation is estimated by an ultrasonic anemometer and the CO2 concentration is measured using a infrared gas analyzer. In order to obtain an aerially integrated view of the urban  11  Introduction  surface, the equipment must be setup high enough to avoid sampling problems caused by turbulent wakes from individual surface elements. It is suggested that a height of at least 2-3 times the roughness elements be used (Oke, 2004). A clear advantage of using a direct measurement approach is the increased temporal resolution. The disadvantage however, is the low spatial extent a study can be carried out — a relatively homogenous fetch is required limiting its application in high-density or varied urban terrains. Furthermore, this approach only accounts for local emissions — from fuel burned within the building — and does not account for emissions attributable to imported electricity. That said, many modern cities are made up of extensive uniform areas, such as suburban development, where the eddy-covariance can easily be deployed. Additionally, estimates based on Inventory (Section 1.4.1) and building energy modeling (Section 1.4.3) can be benchmarked against direct flux measurements that account for all sources and sinks of CO2 (Christen et al., 2011).  1.4.3  Building energy modeling methodology  The use of building energy models (BEM) can be traced back to early software developed in the 1960s by the United States Government (Kusuda, 2001). Since then, the number of software tools has grown to well over a hundred and their applications cover various stages of building design and evaluation. Consequently, software capability now includes energy use estimation, passive solar and ventilation potential, integration of renewable energy technologies, carbon emissions, life cycle analysis, and human comfort evaluation. Furthermore, models exist that integrate urban climate modifications (iteratively), greatly improving the representativeness of climate files (Unzeta et al., 2009). There have been several reviews on BEMs in recent years, containing information on general modeling capability (Crawley 2008) and the early design phase (Lam et al., 2004; Attia et al., 2009). 12  Introduction  The spatial pattern of EUILA in a city is widely diverse. This disparity can be addressed using a bottom-up BEM approach, where studies have either: 1) developed a representative BEM (Kikegawa, 2003; Ohashi et al., 2007; Salamanca and Martilli, 2010), or 2) utilized an established simulation software (Heiple and Sailor, 2008; Unzeta et al., 2009; Christen et al., 2011). Both options follow a box-type heat budget calculation that estimates building energy demand with changing climatic conditions. In option one, the BEM is nested within an urban canopy model. The BEM is then driven by local climate data and associated with in-canopy impacts of radiation availability and wind speed. The advantage of this methodology is the direct feedback between the urban context and local climate. However, this method has thus far been limited to one idealized building type and targeted at UHI research (e.g. air-conditioning heat expenditure feedbacks). Alternatively, option two typically involves the development of building archetypes. Once a representative typology has been established, an associated BEM is run and EUI determined. The advantage of this method is the ability to create a wide variety of archetypes and capture their unique responses to changing climate and operational schedules.  However,  characterization of these archetypes and the inclusion of the local urban context remains a challenge.  1.5  Research goals  1.5.1  Research aims  Currently, community-planning decisions surrounding building energy and emissions are based on coarsely defined, statistically derived inventories and mitigation strategies. Neither have been benchmarked against or reconciled with directly measured evidence. As a consequence the degree to which statistical models parallel real experience is unknown.  13  Introduction  This study aims to develop a methodology that integrates both bottomup and top-down modeling approaches to map building-sector energy use and emissions. The novel component of this research is the integration of remotely sensed LiDAR in urban energy mapping applications. The proposed integration includes the use of a BEM and an urban archetypes database to provide context-sensitive energy and emissions estimates at the neighourhood-scale. The questions raised by the thesis are as follows: • Can LiDAR data be used in community energy and emissions modeling? • Are rapid estimates of neighbourhood-scale energy and emissions possible through a scaling approach? • How does development density influence building energy use?  1.5.2  Research objectives  The research follows three objectives aimed at answering the above questions. Figure 1-3 provides a schematic guide to the thesis process. • Develop a methodology to integrate LiDAR data with a BEM using an archetype approach (Section 2.2 - 2.6) • Determine the sensitivity of an archetype’s EUIV to the urban context, building form and population density (Section 3.1) • Quantify, spatialize and evaluate modeled energy and emissions along an urban transect gradient (Section 3.2)  14  Introduction  Figure 1-3 Thesis overview diagram This schematic representation of the thesis process can be read from top-down. First, geospatial integration will be discussed in Section 2, followed by a discussion of simulation strategy in Section 3.1 and lastly spatialized energy and emissions data in Section 3.2.  15  Materials and Methods  2  Materials and Methods  The goal of this research is to provide a methodology to automatically map the spatial pattern of building energy use and subsequent GHG and QFB according to the local context. The development of such an approach involves the synthesis of multiple data sources that are both spatial and aspatial. This section describes the steps necessary to provide up-scaled, context-sensitive energy and emissions estimates.  2.1  Study site  The study area is a linear cross-section of urban land uses (UT), located within the City of Vancouver (UTM 10 Easting: 492601, Northing: 5455253), Canada’s third largest metropolitan region. The transect makes up an area of 740 hectares, encompassing 7.4% of the total area of Vancouver and spans 1 km wide by 8 km long (Figure 2-1). The UT covers a variety of urban form, capturing areas of Vancouver’s central business district, medium density residential, single detached residential, and industrial development. Vancouver is found within the ASHRAE International Climate Zone (CZ) 5C, which is defined as having between 3000 and 4000 heating degree days HDD (Figure 2-1). HDD is a measurement designed to reflect the dependence of building space heating demand on outside temperature: N  HDD = i=1  (Tb − Ti )+  (1) Heating degree days  Where N is the number of days in the month, Tb= the reference temperature (18 ˚C), Ti = the mean daily temperature, and + indicates only positive values are taken into account. Furthermore, CZ5 is the mildest zone in Canada and within this zone, space heating and DHW demand dominate much of the residential energy load. 16  Materials and Methods  Within the UT, several local climate zones (LCZ) are found, as described by Stewart, including ‘Compact highrise’, ‘Compact midrise’, ‘Large lowrise’ and ‘Open lowrise’ (Stewart, 2011). ASHRAE Climate Zones 5 < 4,000 HDD 6 4,000 - 5,000 HDD  North Vancouver  7 5,000 - 7,000 HDD 8 >7,000 HDD  Downtown (DT) Port Moody Mount Pleasant (MP)  Vancouver  Burnaby  Richmond Delta  Sunset (SS) Urban Transect (UT)  0 N  300 km  0  N  6 km  Ladner  0  1500 m  N  Figure 2-1 Study site a) Map of British Columbia including ASHRAE climate zones, b) Vancouver and surrounding municipalities, and c) LiDAR urban transect with subset study areas.  Three neighbourhoods along the UT were selected for detailed comparison, Downtown (DT), Mount Pleasant (MP), and Sunset (SS)(Figure 2.1 and Figure 2.2). The sites are of equal dimensions (500 m x 500 m) and have been chosen to represent a gradient of development type and intensity, characterized by variations in building density, surface cover, and LCZ (Table 2.1). The SS neighborhood is the least dense, consisting of primarily single family detached dwellings and attached commercial development (LCZ Open lowrise). In the MP neighbourhood, there is a larger degree of building diversity, with several attached and stacked residential buildings, including row-houses, low-rise apartments, and several buildings of mixeduse (LCZ Compact lowrise). The DT neighbourhood is the most dense, with heavily built up residential towers and several buildings over 25 stories, many of which include commercial use at grade (LCZ Compact highrise).  17  Materials and Methods  Figure 2-2 Subset overview maps Order from left to right: Downtown (DT), Mount Pleasant (MP), and Sunset (SS); a) Birds eye perspective, Source: BingMaps (http://www.microsoft.com/maps/), b) Aerial context, Source: Vancouver OpenData Catalogue (http://data.vancouver.ca/), and c) Landuse map  18  Materials and Methods Table 2-1 Neighbourhood metrics Frequency denotes total number of conditioned buildings — secondary structures and garages in brackets. All values have been derived in combination with LiDAR.  2.2  Geospatial integration  The use of LiDAR data in this study differs from previous research and provides a valuable resource where urban characterization may prove challenging. In particular, LiDAR provides a means for rapid, objective classification and extraction of parameters in the urban environment. The collection of such data by alternative sources — if available — can be labour intensive, expensive and pitted with administrative and quality control hurdles. The integration of spatial (LiDAR, aerial photography, census dissemination areas, building taxlot outlines) and aspatial data (property assessments, population counts, archetype attributes) is carried out in GIS (ESRI’s ARCGIS Version 10.0 Redlands, CA, USA) and the Interactive Data Language (IDL Version 8.0 Boulder, ITTVis, CO, USA) computing environments (Figure 2-3).  19  Materials and Methods  Figure 2-3 Geospatial integration Spatial integration of LiDAR data starts with products produced in Goodwin et al. (2009), integration of cadastral and classification datasets (Census, Assessments, Archetypes) and finally the extraction/derivation of LiDAR informed metrics.  2.2.1  Spatial data  LiDAR data, at a resolution of 0.7 laser pulses m-2, was collected in March, 2007 by a TRSI Mark II discrete-return sensor mounted on a fixed-wing platform and was flown by Terra Remote Sensing (Sidney, BC, Canada).  20  Materials and Methods  This data, along with high resolution Quickbird satellite imagery (Collected March, 2007, spatial resolution of 2.4 m) and field measurements were combined and processed as outlined in Goodwin et al., 2009 and Tooke et al., 2009. Several products produced by the above authors were made available to this study: 1) a DEM at a 1x1 m resolution where each raster element contained its location in space, and 2) an object classification at a 1x1 m resolution differentiating between building, ground surface and above ground vegetation. The remotely sensed data products outlined above were overlaid with two vector polygon layers: 1) a cadastral layer obtained through the City of Vancouver’s Opendata Catalogue, which when clipped to the UT contained 7941 individual taxlots within the UT (City of Vancouver, 2011), and 2) Census dissemination area (DA) boundaries, which when clipped to the UT contained 498 independent DAs (Statistics Canada, 2007)  2.2.2  Aspatial data  Two additional data sources were joined to the aforementioned spatial data through common identifiers: 1) British Columbia Property Assessment (BCPA), which contained detailed building type descriptions for all buildings and further construction information for ground-oriented residential buildings. The fields used for analysis are presented for groundoriented development in Table 2.2. The BCPA landuse codes (79 found within the UT) were joined to the municipal taxlot polygons and provided the framework for further classification described in Section 2.3 and, 2) 2006 Canadian Census population counts, obtained through the Canadian Census Analyzer (CHASS, 2011). The population counts were joined to DA polygons — which provided the starting point for down-scaling population data (Section 2.4).  21  Materials and Methods Table 2-2 Property assessment statistics Built form averages along the UT for ground-oriented residential properties. Included are total dwelling share (%) of car park type, basement type, and heating system type respectively. * denotes per dwelling averages — opposed to per building averages.  2.2.3  Building threshold and the ‘edge effect’  Once spatial and aspatial data were combined, the object classification raster and BCPA layer were overlaid to extract individual contiguous building elements (Figure 2-4). This step associated a landuse to each contiguous building element regardless of size or landuse (n = 11638). To remove secondary structures with negligible energy demand (e.g. garages and sheds) a building threshold was determined. This threshold was determined by rank order analysis of single detached dwelling (SDD) parcels. Specifically, on each parcel, every building element was ranked by volume. The threshold between primary building elements (rank 1) and secondary building elements (rank 2) was determined as the intersection of their respective frequency tables (Figure 2-5). A threshold of 255 m3 was determined and all non-primary structures below this volume were classified as secondary unoccupied structures. The above threshold, along with BCPA classified vacant structures reduced the total occupied building count from 11638 to 7812 buildings. 22  Materials and Methods  Figure 2-4 Building ‘edge effect’ a) LiDAR derived building footprints, b) Buildings overlaid with land use parcels showing contiguous building elements, and c) Simplification of building edges.  Building Volume: Rank Order Threshold for Secondary Structures 700  600  Threshold = 255 m3  Frequency  500  Rank 2 Structures  400  Rank 1 Structures 300  200  0  0 19  50  -2  00  85  0  -1  18  00  75  0  -1 00  17  -1  55  00  16  -1 00  15  65  0  0 45  0  -1  35  00  14  -1 00  13  12  00  -1  25  0  0  11 5 0-  11 0  -1  05  0 95  00 10  0 85  90  0-  0  0-  0  75  80  0-  0  65  70  0-  0  55  60  0-  45  50  0 35  40  0-  0  0-  0  25  30  0-  15  20  010  0-  50  0  0  100  Building Volume (m3)  Figure 2-5 Single detached dwelling rank order Rank Order (by building volume) plot for primary and secondary building elements found on single detached dwelling Lots within the UT.  Although LiDAR can produce both vector and raster data products from a raw point cloud, this study was limited to processed raster products (Section 2.2.1). Due to the raster format, an ‘edge effect’ is produced along buildings edges that cross raster elements diagonally at 1m steps (Figure 2-4). This is problematic if realistic morphological characteristics,  23  Materials and Methods  such as building perimeter LP or wall area AW are to be extracted (Tooke et al., 2011). Therefore, Bayers polygon simplification algorithm was used, in which a recursive approach smooths edges through regression analysis (Bayer, 2009).  2.3  An archetype approach  An archetype approach can ultimately lead to the assessment of alternative development scenarios and the benchmarking of energy use and carbon emissions inventories. One of the challenges of this work is simplifying the complexity of built form into appropriate archetypes to construct a representative typology. This section will first present an overview of the archetype approach, then discuss the typology used and finally outline what building characteristics are extractable from LiDAR data and those that need to be informed by other sources.  2.3.1  Overview  Cities are assemblages of development patterns, which are in turn made up of replicated parts such as building types (archetypes). These types often share similar characteristics, for instance morphological and energy performance attributes. A building typology (constructed of representative archetypes), which includes these differences, can better inform policy decisions by describing important factors influencing energy use and emissions within a development pattern; systematically synthesizing otherwise complex variations in built form patterns. This section describes one such approach that is informed by LiDAR data. The building typology undertaken here describes 7812 buildings through a series of archetypes and assumes a correlation between building use, physical form and energy performance. Through fieldwork, precedent energy modelling studies (CanmetENERGY, 2011; OEE, 2008) and house audits, 12 building archetypes, along with a mixed-use archetype were  24  Materials and Methods  chosen to represent the UT. The resulting types represent replicable and scalable instances that facilitate a rapid assessment of energy and emissions performances for a large number of buildings (Section 3.1).  2.3.2  Typology structure  The typology used specifically focuses on building attributes relevant to energy consumption and emissions, including: 1) building use; 2) intensity of development, detached, attached and stacked; 3) morphological attributes, such as glazing percentage, floor heights, and insulation values; 4) building system attributes, such as space heating and hot water system type and efficiency; 5) occupancy schedules with associated lighting and appliance loads. These categories form the basis for the typology and were assigned according to the following steps: Step 1: Building Use - The first step classified buildings by their use. This resulted in five categories: Residential, Office, Retail, Industrial and Civic. Of the 7812 buildings, 87% are Residential, a significant percentage of the UT necessitating a sub-classification (non-residential buildings included Civic 1.2%, Retail 5.3%, Office 2.0%, Light Industry 2.3%, and Mixed-use 2.2%). Step 2: Building Form - The second categorization further divided the Residential buildings based on dwelling configuration, resulting in four additional categories: SDD (73.9%), Duplex (1.3%), Multiplex (5.3%), Rowhouse (1.8%), Lowrise/Midrise (3.7%) and Midrise/Highrise dwellings (1.0%). Of the 6796 Residential buildings, the majority are SDD, necessitating an additional categorization. Step 3: Building Vintage - The third categorization further divided SDDs based on year of construction, this resulted in three additional categories: SDD built before 1965 (41.5%), SDD built between 1965-1990 (20.8%), and SDD built post 1990 (11.7%). This categorization assumes a correlation between building age and energy performance and was informed by ecoEnergy retrofit house audits (n=12) and the CMHC Renovating for 25  Materials and Methods  Energy Savings Report (CMHC).  2.3.3  Elements database  Although building volume is a good indicator of the space to be heated and it can be derived from LiDAR, several building attributes important to energy use characterization are not visible to LiDAR and need to be sourced elsewhere (Figure 2-6). To accomplish this, an online database ‘elementsdb’ is used. Elementsdb catalogues comparably illustrated and measured examples of building archetypes — based on the premise that urban form is comprised of replicable elements (or cases) that can be organized by landuse, type and intensity (Elementsdb). Morphological characteristics of cases in this database were matched to the aforementioned typology. Once associated with an archetype, data (from elementsdb) such as window-to-wall ratios and conditioned floor-tovolume ratios — alongside morphological attributes extracted from LiDAR data — were imported into the BEM (Table 2.3). Further attributes were sourced from the Canadian Energy Use Data Handbook (OEE, 2008).  26  Materials and Methods  LiDAR informed inputs Building morphology Urban morphology Population density  TYPOLOGY infomed inputs Building use Building construction Building systems  Vancouver climate file  Figure 2-6 Typology vs LiDAR attributes Schematic diagram of inputs that can be informed or derived from LiDAR data and those that need to be sourced from other data.  Table 2-3 Archetype attributes: Part 1 Built form attributes for selected archetypes. Sources: LiDAR derived metrics, elementsdb, Canadian Mortgage and Housing Corporation (CMHC), house audits  27  Materials and Methods  2.4  Building-scale population down-scaling  In order to accurately estimate DHW demand and develop per capita energy and emissions estimates, a good understanding of population density is needed. In this study, building-scale population estimates are derived through a combination of LiDAR and Census data. Specifically, population counts are derived from the 2006 Canadian Census at the DA scale — a DA typically covers one to several city blocks (CHASS, 2007). As several of the DAs are only partially within the UT a procedure for downscaling population counts was developed. This procedure down-scaled DA population to the building-scale using LiDAR derived residential building volumes as a proxy. This was done through a multiple-linear regression that used only DAs found completely within the UT (n=318) and followed three steps: Step 1 Building density - First, as the volume per person differs between residential building types, buildings were classified according to their density. For example, shared amenity space and underground parking requirements  greatly  differ  across  development  scales.  Therefore,  residential buildings found within the 318 DAs were classified as one of the following: detached, attached, stacked or mixed-use. Step 2 Regression - Next, a multiple-linear regression was preformed on the 318 DAs, using the total volume of detached, attached, stacked, and mixed-use buildings as the independent variables and the total DA population count as the dependent variable.  28  Materials and Methods Table 2-4 Population regression output Regression output for independent variables building volume (Mixed-use, Detached, Attached, Stacked Residential) and dependent variable Census DA population.  Step 3 Census-level population assignment - For DAs found partially within the UT, the regression model was used to predict population in each of the categories accordingly:  PM = αD VD + αA VA + αS VS + αM VM  (2) Population attribution  Where PM= modeled total population, a = regression coefficient, V = total building volume, D,A,S,M = detached, attached, stacked, mixed-use buildings. For DAs contained completely within the UT the regression model was used to partition population into each of their respective categories (detached, attached, stacked, mixed-use) and modeled estimates were then adjusted to actual population counts provided by the Census:  P=  PDA (αD VD + αA VA + αS VS + αM VM ) PM  (3) Final population  Where P = final population, PDA= DA population, PM= Modeled population according to eq. (2), a = regression coefficient, V = total building volume, D,A,S,M  = detached, attached, stacked, mixed-use buildings.  Step 4 Building-level population estimate - Each building element was associated with a DA and the total population found within each category was mapped down to each building based on volume. This approach  29  Materials and Methods  assumes that within each DA all buildings of the same dwelling category (detached, attached, stacked, mixed-use) share the same volume per capita. Lastly, if the DA was found completely within the UT, down-scaled population was adjusted according to actual population. Step 5 Mixed-use buildings - No indication of residential proportion in mixed-use buildings is given by the BCPA. Therefore, in order to establish the relative fraction residential, the average volume per capita found in stacked residential dwellings was used as a proxy to estimate the volume dedicated to residential use in mixed-use buildings:  R=  P · VInh. V  (4) Mixed-use residential proportion  Where R = proportion built volume that is residential, P = Building population, VInh. = Average volume per person in stacked residential development (m3 Inh -1), and V = volume of building (m3).  2.5  Building-scale morphology extraction  One of largest impediments of integrating LiDAR with a BEM has been the geometric description of a building. According to Bazjanac (2001), geometric definition can take up to 80% of the time when preparing a BEM. Nonetheless, a geometric description of the building is necessary to evaluate thermal and energy exchange. Although EnergyPlus (Section 2.7) lacks a dedicated user-interface for manipulating building geometry, this study makes use of a scripting environment (Ruby) in Google Sketchup (Version 8.0 Pro, Mountain View, CA, USA) to parametrically scale building morphology before input into DesignBuilder (Version 2.3, Gloucestershire, UK) — an EnergyPlus extention.  2.5.1  Building morphology defined  When looking at all 7812 buildings in the UT, the large variety of building 30  Materials and Methods  shapes and sizes becomes apparent. Furthermore, building morphology, here defined as volume V, exterior surface area AS, wall area AW, footprint AF, depth LD, height zH, and perimeter LP geometries, has been shown to scale allometrically to gain solar access (Steadman, 2006; Batty et al., 2008; Steadman et al., 2009). This variation is most pronounced in archetypes with wide distributions in size (e.g. Civic, Industrial, Office, Retail, Mixed-use, Lowrise/Midrise and Midrise/Highrise buildings).  2.5.2  Deriving building morphology  V and LD are key attributes that influence energy demand. For example, V is related to a building’s compacity (e.g. characteristic length; LC = V / AS) which influences its ability to retain heat (Gupta, 1987 ; Papadopoulos et al., 2002; Pessenlehner and Mahdavi, 2003; Shashua-bar et al., 2006). LD, on the other hand is related a building’s passive area or its potential to harness solar gains and natural ventilation to offset active energy supply. Given the allometric relationships found between building measures, a scalable morphology was determined according to the following steps: Step 1 Building volume used to predict building footprint - This followed regression analysis between LiDAR extracted building volume V and LiDAR extracted building footprint (both decomposed into principle lengths; V1/3 and A1/2 ) and can be written as follows:  AF = αV 2/3  (5) Building footprint  Where AF = predicted building footprint (m2), a = the regression coefficient , V = LiDAR extracted building volume (m3).  31  Materials and Methods  Step 2: Building depth was estimated based on building volume and wall area following Batty et al. (2009) where:  LD = 2  V AW  (6) Derived depth  And  AW = ZH LP  (7) Building wall area  Where LD = building depth (m), V = building volume (m3), AW = building wall area (m2), ZH = building height (m), LP= building perimeter (m), and LD = building depth (m). Step 3: Building footprint was then used to predict building depth through a second regression, LiDAR extracted building footprint vs calculated building depth (building footprint decomposed into principle length; A1/2 ) for each of the 12 archetypes: 1/2  LD = αAF  (8) Building depth  Where LD = building depth (m), a = the regression coefficient , AF = LiDAR extracted building footprint (m2). Finally, building height was derived as the remaining dimension:  ZH =  V LD LL  (9) Building height  Where ZH = building height (m), V = building volume (m3), LD= building depth (m), and LL = building length (m). Lastly, building height was rounded up or down based on the mean floor-to-floor height of each archetype.  32  Materials and Methods  2.6  Urban context extraction  The amount of energy consumed by a building is also influenced by its surrounding urban form. Specifically, the density of urban areas impact solar radiation access, the canopy layer air temperature and wind flow patterns. These in turn influence a buildings demand for space conditioning and lighting (Oke, 1987; Ratti et al., 2003, 2005; Robinson, 2006). Most BEMs concentrate on the performance of a single building and shading from the surrounding context is often neglected (Purdy and Beausoleil-Morrison, 2001). This is in part due to the challenge of documenting a spatially accurate representation of the adjacent buildings and trees. LiDAR however, provides an opportunity to overcome that challenge through automatic characterization of surrounding urban form.  2.6.1  Urban context defined  For the purposes of this project, urban context refers to the shading and sheltering obstructions surrounding a building of interest and does not differentiate between surrounding buildings, vegetation or infrastructure. Within this framework, two core metrics of potential solar access were selected — sky view factor ySVF (Section 2.6.2) and fractional shading S (Section 2.6.3).  2.6.2  Sky view factor ySVF  Access to solar radiation is largely driven by the spatial pattern of shading obstructions. Likewise, obstructions can influence the longwave radiation exchange and subsequently the magnitude of nighttime cooling. The influence urban form may have on the radiation exchanges can be estimated through ySVF analysis. The ySVF is a measure of the view openness or as outlined by Oke, “the ratio of the amount of sky ‘seen’ from a given point to that potentially available” (Oke, 1987). For instance, a ySVF of 1 describes a completely unobstructed view from a particular surface where  33  Materials and Methods  as a ySVF of 0 describes a completely obstructed view. The ySVF can be calculated a number of ways — for example through building height to width ratios (Oke, 1987), fish eye photographs (Brown et al., 2001) or 3D data (Ratti et al., 2003). In this study, the approach described below is based on a LiDAR derived DEM, where for each building (i.e. building was removed with surrounding context left intact), ySVF was derived at three positions: 1) roof centroid, 2) middle centroid, and 3) ground centroid. These three values were then averaged for each building and used in Section 3.1. For each raster element P’(x’,y’) of the DEM, the horizon obstruction was calculated by determining the element P(x,y) that produces the highest obstruction angle q (Figure 2-7). This was calculated for point clouds P (x,y) within narrow sectors originating at P’(x’,y’):  θ(k) = arctan max  (x − x )2 + (y − y )2 z −z  (10) Obstruction angle  All objects were treated as opaque (including trees) and no holes were expected below the highest return. This procedure was repeated to determine q(k) for k = 16 sectors in different azimuthal directions around the point P’(x’y’) (i.e. 1 sector = 360/k).  Figure 2-7 Obstruction angle Schematic of obstruction angle calculation. Source: van der Laan et al. (2011).  34  Materials and Methods  With known  q(k), the ySVF of a point for the given sector was then  determined as (A - A’) /A, where A is the total surface of a section of the upper hemisphere (thick outline in Figure 2-8). A’ is the surface of the section that is obstructed and the unobstructed section is given as A - A’:  Figure 2-8 Sky view factor calculation Schematic of sky view factor calculation. Source: van der Laan et al. (2011).  The final ySVF is calculated as the sum of all sector sky view areas (blue) divided by the area of the entire upper hemisphere:  Ψsky (x y ) =  K k =1 (A − A K k =1 A  )  (11) Sky view factor  Where the area average of ysky (‘all horizontal surfaces’) was calculated as the average ysky(x’y’) of all pixels located within the region of interest and includes pixels on roofs and trees.  2.6.3  Fractional shading  Although ySVF is a good indicator of radiation exchange, its unit of measure has no directional bias or dependence on latitude and is thus limited in its ability to estimate actual solar radiation potential. Therefore a second  35  Materials and Methods  metric, fractional shading S, a measure of relative solar radiation potential, was defined and derived at the building-scale (Figure 2-9). For each building (i.e. building was removed with surrounding context left intact), S was derived at three positions: 1) roof centroid, 2) middle centroid, and 3) ground centroid. These three values were then averaged for each building and used in Section 3.1. The calculation of fractional shading makes use of a sun hours matrix1 for 360 azimuth angles x 90 elevation angles (Figure 2-9). For each co-occurence (azimuth and elevation) the number of sunshine hours from given direction is provided. This calculation assumes cloudless skies and therefore only describes the potential shading effect from obstructions. The shading factor S is calculated accordingly:  S=  Hs,max − Hs,act Hs,max  (12) Shading factor  Where HS,MAX= hours with direct beam irradiance on a horizontal surface at a given geographical latitude (Vancouver = 4404 hours), HS,ACT= hours with direct beam irradiance to a point in the DEM (average of roof centroid, middle centroid, and ground centroid).  1  A joint probability function for solar azimuth W and solar elevation b 36  Materials and Methods  Solar elevation  90°  0°  a)  0°  Solar azimuth  360°  Sun hours High  Solar elevation  90°  0°  b)  0°  Solar azimuth  360°  Low  Figure 2-9 Azimuth-elevation matrix Sunpath schematic expressed in sunshine hours per year showing: a) unobstructed sunpath intensity, and b) obstructed sunpath intensity — example obstructions outlined in red.  2.7  Modeling building energy use  In order to model the energy performance of a building, one is faced with abstracting the complexities of reality down to a reasonable resolution necessary for simulation input. When carrying out a simulation at the neighbourhood-scale, this abstraction is further complicated due to the variety of building types and reduced amount of data sources to draw upon. Therefore, a suitable BEM for urban scale implementation ideally includes conventional construction libraries, HVAC systems, and operating schedules for multiple building types. Unfortunately, many of the interfaces that facilitate these inputs also have simplified morphological definitions and thus limit the integration of LiDAR data. The BEM selected for this research, EnergyPlus (Version 6.0, US Department of Energy), was selected based on the freedom of morphological inputs, evaluation capability and its extensive documentation.  37  Materials and Methods  EnergyPlus has several important capabilities that make it a suitable tool for the presented research. First, EnergyPlus is structured in a modular format, with a solutions manager that calls sub-modules — such as the ‘shading’ module — that are similar in scope to the morphological variables of interest in this research. Furthermore, this format provides for an easily extendable model and has subsequently garnered ample support from both research and the private-sector. Second, EnergyPlus has the ability to run simulations at a sub-hourly time-step, this along with a text-based weather file (EPW) allows for the integration of directly-measured weather files. The EPW weather file is based on the typical meteorological year (TMY), but the EPW’s comma separated format allows for extended flexibility (Crawley, 2008; Jentsch, 2008). For a list of all the variables included in the EPW file please see (DOE, 2009). Third, simulation results can be output in a text format allowing for post-processing and extended analysis. Last, due to the open source structure of EnergyPlus, several user interfaces have been developed that help facilitate the input of construction and materials libraries, operating schedules, HVAC system type and efficiencies along with computer automated design (CAD) support for morphological inputs. The following sections illustrate several relevant calculations performed by EnergyPlus in relation to building morphology, urban context and population density. For further reference, the reader is directed to the EnergyPlus Engineering Reference (DOE, 2009).  2.7.1  Heat balance  The thermal efficiency of a building — an important indicator of energy demand — is typically evaluated by ‘unit thermal resistance’ or R value of the surfaces enclosing the building (envelope). The R value is the temperature difference (∆T) across an insulator divided by the resultant heat flux through the area, R = ∆T/QA (m2 K/W) where QA is the heat flux  38  Materials and Methods  density per unit area. Therefore aperture1 area, material type, building surface area and volume are all important factors when assessing the energy use in a building. When we have a good understanding of the thermal efficiency of a building’s envelope, we can then estimate the system energy needed to balance the heating or cooling load within a building zone2. −Qsy = Qig + Qcn + Qzn + Qin  (13) System heat balance  Where Qsy= is the system heating/cooling load, Qig= is the convective internal gains, Qcn= is the convective heat flux from internal surfaces, Qzn= is the heat flux from adjoining zones, Qin= is the heat flux due to infiltration. The heat balance of an interior surface is solved through a heat balance equation involving four coupled heat transfer components: 1) net longwave radiation flux between zone surfaces Qlw and from equipment in zone Qlwe, 2) net shortwave radiation flux from solar Qsw and internal lights Qswl, 3) conduction heat flux through the wall Qki, and 4) convective heat flux to zone air Qcn. The heat balance on the interior face of the zone an be written as: Qsw + Qswl + Qlw + Qlwe + Qcn + Qki = 0  (14) Interior heat balance  The heat balance on the outside surface is likewise solved using a coupled method involving four components: 1) net longwave radiation flux from sky, ground and surrounding obstructions Qlw , 2) net shortwave radiation flux from direct, reflected and diffuse sources Qsw , 3) convective heat flux exchange with UCL Qcn, and 4) conduction heat flux into the wall Qko. The  1  The sunlit opening of a building facet (roof or wall)  2  An enclosed volume within a BEM that shares similar scheduling and environmental influences. 39  Materials and Methods  heat balance on the outside face of the zone can be written as: (15) Exterior heat balance  Qsw + Qlw + Qcn − Qko = 0  Both interior and exterior heat balance exchanges are shown in Figure 2-10.  QSW QLW QCN  QKI  QKO  Exterior Heat Balance  QSW+QSWL QCN  QLW +QLWE  Interior Heat Balance  Figure 2-10 Heat balance components EnergyPlus heat balance components for outside zone volume and inside zone volume. See above for term descriptions.  Furthermore, EnergyPlus calculates the vertical variation in air temperature, barometric pressure, and windspeed. The vertical air temperature, is calculated according to the US Standard atmosphere:  TZ = TB + L(HZ − HB )  (16) Air temperature  Where TZ= temperature at height z, TB= air temperature in the UCL, L= environmental lapse rate (-6.5 K km-1), HB= offset (equal to zero for the troposphere), HZ= geopotential altitude.  2.7.2  External radiation  Of the four exterior heat balance components solved in EnergyPlus, two are of particular relevance to this study (shortwave and longwave radiation) due to their relationship to building morphology and urban context.  40  Materials and Methods  For further information on the other two heat balance components (conduction and convection) see DOE, 2009. The external shortwave radiation absorbed by a surface is made up of direct, diffuse and reflected components: qsw = qdr + qdf + qr  (17) Shortwave radiation  The amount absorbed by a surface depends on its location (i.e. latitude and position amongst surrounding obstructions), tilt, material and sky conditions. The direct solar radiation is given as: qdr = aIDR cosθ  ASL A  (18) Direct solar radiation  Where qdr= direct solar radiation, a = the surface solar absorptance, q = the sun’s angle of incidence, A = area of the surface, Asl = the sunlit area of the surface, and Idr = direct solar radiation flux density. The calculation of diffuse shortwave radiation from the sky takes into account the sky anisotropic radiance distribution. The distribution is based on an empirical model and is described in (Perez et al., 1990). In this model, the calculation includes: 1) isotropic distribution that covers the entire sky dome, 2) circumsolar brightening centered at the position of the sun, and 3) a horizon brightening. The external longwave radiation absorbed by a surface is made up of ground, sky and obstruction (trees, buildings) components. The amount of radiation absorbed by a surface depends on surface absorptivity (following Kirchhoff’s law, emissivity = absorptivity), temperature of emitting body (ground, sky or obstruction), and view factors. In EnergyPlus, the following assumptions are made: 1) each surface is a graybody, opaque and emits or reflects diffusely, 2) each surface is at a uniform temperature,  41  Materials and Methods  3) the outward longwave radiative flux is Lambertian, 4) the medium within the enclosure is non-participating. The total longwave radiative flux is the given by the exchange between ground, sky and shading obstructions: qlw = qg + qsk + qo  (19) Longwave 1  Where qlw= total longwave net radiation flux density, qg= ground exchange, qsk= sky exchange, and qo= shading obstruction exchange. This can be further broken down into the following equation: qlw = hr ,g (Tg − Tsf ) + hr ,sk (Tsk − Tsf ) + hr ,o (To − Tsf )  (20) Longwave 2  Where:  hr ,g =  εσFg (Tsf4 − To4 ) Tsf − To  hr ,sk =  4 εσFsk ψSVF (Tsf4 − Tsk ) Tsf − Tsk  hr ,o =  εσFsk (1 − ψSVF )(Tsf4 − To4 ) Tsf − To  (21) Longwave 3  (22) Longwave 4  (23) Longwave 5  Where Tg= outside ground temperature, Tsf= outside surface (of zone) temperature, Tsk= sky temperature, To= shading obstruction temperature, e = surface longwave emissivity, s = Stefan’s constant, ysky= view factor sky, Fg = surface to ground view factor, Fsk = surface to sky view factor. View factors are further defined by: Fg = 0.5(1 − cos φ)  (24) Longwave 6  42  Materials and Methods (25) Longwave 7  Fsk = 0.5(1 + cos φ)  2.7.3  Shading module  The proportion of a surface obstructed from sunlight is an important part of assessing heat gains within a building. EnergyPlus can solve multiple reflections from building components, other buildings and the surrounding ground surface. Furthermore, the shading module in EnergyPlus has the ability to assign transmittance values and schedules to obstructions. The shading algorithm used is based on a coordinate transformation method, where shadow polygons are projected in space and then overlaid with building geometry. Once projected the proportion shaded is determined by calculating the overlap between the shadow polygon and the receiving polygon. This calculation can be performed for multiple overlapping shadow polygons as shown in Figure 2-11.  b  Sunlit area = Aa-(Ab+Ac-Ad) a d c  Figure 2-11 Overlapping shadows Schematic of EnergyPlus sunlit area calculation. A denotes area of polygon.  EnergyPlus can calculate direct and reflected solar radiation striking the building surface. There are three categories of reflecting surfaces in EnergyPlus: 1) shadowing surfaces, 2) exterior building surfaces, and 3) the ground surface. Shadowing surfaces are obstructions, such as surrounding buildings or overhangs attached to the building and have both diffuse  43  Materials and Methods  and beam-to-beam reflectance values. Exterior building surfaces are parts of the building that reflect solar radiation onto another part of the building — such as that found in L-shape or courtyard building morphologies (See Figure 2-12). In this case, walls are assumed to be diffusely reflecting and windows directly reflecting. Lastly, the ground surface reflections account for both beam-to-beam and diffuse reflections although areas in shade do not contribute to beam-to-diffuse or beam-to-beam totals (Figure 2-12).  sky diffuse  beam-diffuse beam-beam  wall window  shaded areas  Figure 2-12 Solar radiation reflections Schematic of how EnergyPlus handles solar radiation reflections. Dashed lines denote diffuse and solid arrows direct (beam).  In order to calculate the amount of direct solar and diffuse solar radiation, EnergyPlus uses a ray tracing methodology. A series of rays (90) are sent outward in a hemispheric pattern after which they hit either an obstruction, the ground surface or are sent out to the sky. The reflection of incident beam or sky solar radiation is then determined and the contribution added to the total radiation balance.  44  Materials and Methods  The total solar gain on an external surface can then be calculated as:  Qso = a Ib cosθ  Ssl + Is + Fss + Ig Fsg S  (26) Total solar radiation  Where Qso= total solar radiation (W m-2), a = the surface absorptivity in the shortwave, Idr = direct solar flux density, q = the sun’s angle of incidence, A = area of the surface, Asl = the sunlit area of the surface, Idf = diffuse solar flux density, Ir = reflected solar flux density, Fg = surface to ground view factor, Fsk = surface to sky view factor  2.7.4  Hot water demand  Although occupant density impacts heating/cooling demand along with ventilation requirements, the largest impact on energy use is the demand for DHW — estimated to be the second largest energy end-use in Canadian households (CBEEDAC, 2005). The amount of DHW used in a zone is based on: 1) occupant density, 2) demand per occupant, and 3) system type and efficiency. In EnergyPlus, the BEM is partitioned into zones according to activity (e.g. Residential, Office, Circulation), in turn each zone is associated with internal gains and control schedules. The internal gains are made up of occupancy, equipment, and lighting and are driven by a corresponding zone schedule. In contrast, controls are driven by set points and occupancies, such as heating/cooling temperature set-points, ventilation requirements, lighting levels, and DHW per occupant. In order to define the necessary occupant density input for EnergyPlus (Inh. m-2 floor area), LiDAR derived Inh. m-3 V was mapped down according to the elementsdb ‘floor area-to-volume’ ratios. Next, an average DHW demand per occupant was assigned according to the Canadian Building Energy End-Use Data and Analysis Centre CBEEDAC (Aguilar et al., 2005). In this report, the total per capita water consumption is estimated to be  45  Materials and Methods  350 Litres Inh. -1 day-1, of which 39.6% (from DeOreo and Mayer, 2000) is estimated to be for DHW use (138.6 Litres Inh.-1 day-1). The DHW demand is assumed to be independent of building type and location along the UT. Lastly, the DHW efficiency is the energy content of the water drawn divided by the energy required to heat and maintain the water and the systems set point temperature: EF =  MCp (Tt − Ti ) Qd  (27) DHW efficiency  Where EF = the energy factor, M = mass of the water used (kg), CP = specific heat of water (J kg-1K-1), Tt = water heater thermostat set-point temperature (˚K), Ti = inlet water temperature (˚K), Qd = daily water heater energy consumption (J).  46  Results and Discussion  3  Results and Discussion  This chapter has been split into two primary sections. The first section, 3.1 Scaling approach, presents archetype characteristics and energy simulation results for three scaling factors: urban context, building morphology, and population density. The second section, 3.2 Upscaled energy and emissions, presents context-sensitive modeling results along the UT.  3.1  Scaling approach  This research uses a combination of bottom-up and top-down modeling to document, classify and model individual building energy use and subsequent CO2e and QFB along the UT. Rather than directly simulate all 7812 buildings in a BEM, a scaling approach was adopted to rapidly adjust EUIV according to local conditions. In particular, an adjustment was performed for each of the twelve archetypes according to their EUIV sensitivity to local: 1) building morphology, 2) urban context, and 3) population density. The sensitivity to each of these three scaling factors was determined through a series of building energy simulations (n = 160), from which all (n = 7812) building’s EUIV was scaled and mapped accordingly (Figure 3-1). The following section will discuss the simulation methodology, scaling sensitivities and present up-scaled EUILA, CO2e, and QFB along the UT.  47  Results and Discussion  Figure 3-1 Simulation strategy Three scaling factors were used to spatialize energy and emissions according to building energy sensitivity to building morphology, urban context and population density.  3.1.1  Set-up  Within the BEM environment, several variables were standardized across each simulation according to the following steps: Step 1: A 3D model was created for each archetype in Designbuilder. DesignBuilder is an extension to EnergyPlus that facilitates attribute assignment and 3D automation. Archetype attributes important to the simulation were assigned and are summarized in Table 3.1. Step 2: A generic array of shading obstructions was modelled surrounding each archetype - this consisted of a three by three array of shading blocks, where the middle block was replaced by the archetype (Figure 3-2). The shading obstructions were then set as type: component block type standard (thermal absorptance = 0.9, thermal resistance = 0.15).  48  Results and Discussion  Figure 3-2 Shading array example Shading obstruction array example in DesignBuilder. Building shown in grey is the Multiplex archetype, purple blocks the shading obstruction, and green the model ground interface.  Step 3: The following simulation variables were set accordingly: 1) A shadow calculation interval of seven days, 2) Full exterior reflections, 3) Six simulation time-steps per hour, 4) Vancouver EPW weather file, 5) Compact HVAC simulation, 6) Convergence warm-up period set to max of 25 days ensuring a correct heat distribution at the beginning of simulation 7) Air temperature control with a heating set-point at 21°C. Step 4: Annual simulations of 365 days were performed according to following three Sections and EUIV was output for lighting, auxiliary (appliance and system loads), DHW, space heating and cooling demand.  49  Results and Discussion Table 3-1 Archetype attributes: Part 2 Built form attributes important to simulation inputs and building scaling calculations.  3.1.2  Influence of the urban context  In order to determine each archetype’s sensitivity to the urban context, five simulations were carried-out with varying ySVF for each archetype — all other model attributes were held constant (i.e. mean values for V and population density). The urban context simulations were conducted as follows: Step 1: The ySVF percentiles (5th, 30th, 50th, 70th, 95th) of all buildings (for each archetype) were extracted from LiDAR data along the UT according to Section 2.6. Step 2: The corresponding shading obstruction height (surrounding array) was determined iteratively for each ySVF and used as an input in the model.  50  Results and Discussion  Step 3: Five annual simulations were conducted for each archetype using arrays of various according to Step 2, resulting in a total of 60 simulations. Figure 3-3 illustrates (for all archetypes) the relationship between ysky and EUIV (MJ m-3 yr-1).The correlation is very strong among all archetypes (R2 range of 0.90 - 0.99; Table 3.2) and all archetypes except Retail display a negative relationship between ysky and EUIV (i.e. increasing a buildings access to solar gains decreases EUIV). This is in part due to the positive receipt of solar energy which offsets space heating and lighting demand. In contrast, much of the energy use in the Retail archetype is expended on space cooling — opposed to space heating — due to the increased heat input from auxiliary equipment, appliances and lighting. Therefore a larger  ysky increases the amount of energy needed for space cooling. In order to determine the best metric to scale urban context effects archetypes were further sorted by fractional shading S (Figure 3-4). S  unlike ysky is a directional metric and as expected the correlation is  improved in eight of the twelve archetypes (R2 range 0.88 - 1.00; Table 3.3), with average RMSE decreased by 38% (S = 0.89 MJ m-3 yr-1; ysky =1.43 MJ m-3 yr-1). This is due to the sun direction introduced in S which is not present in ysky. The greatest UM influence was seen in residential Lowrise/Midrise and Civic archetypes, where EUIV varied by a factor of 1.4 as a function of S. This was in part due to the large range in fractional shading seen in both of these archetypes - Lowrise/Midrise S ranged from S = 0.12 - 0.90 and in Civic S ranged from S = 0.09 - 0.96.  51  Results and Discussion  Lastly, due to the directional nature of S and the general improvement in the correlation with EUIV, S was determined as the more meaningful scaling factor for urban morphometry influences. Accordingly, the association between S and EUIV has been approximated linearly:  EUI = αS S + βS  (28) Shading derived EUI  Where EUI = Energy use intensity EUIV (MJ m-3 yr-1), aS = shading coefficient, S = fractional shading of archetype, and bS = baseline EUIV with a S of zero (no shading).  52  Results and Discussion  Figure 3-3 Energy use intensity EUIV vs. sky view factor Modeled EUIV versus varying sky view factor. Linear fits correspond to the regression statistics presented in Table 3.2.  Table 3-2 Energy use intensity EUIV vs. sky view factor Regression statistics for modeled EUIV versus varying sky view factor.  53  Results and Discussion  Figure 3-4 Energy use intensity EUIV vs. fractional shading Modeled EUIV versus varying fractional shading. Linear fits correspond to the regression statistics presented in Table 3.3.  Table 3-3 Energy use intensity EUIV vs. fractional shading Regression statistics for modeled EUIV versus varying fractional shading.  54  Results and Discussion  3.1.3  Influence of building morphology  As seen from Figure 3-5 and 3.6, building shape grows allometrically as opposed to a cube which grows isometrically (over-plot). For example, in all the archetypes modelled, the footprint of the building grows more rapidly with increasing volume than what we would expect in a cube. Furthermore, building depth grows more slowly with increasing volume than what we see in a cube. Batty et al. (2008), attribute this to building code requirements that facilitate solar access and subsequent building form augmentations. In order to determine each archetype’s sensitivity to building morphology, five simulations were carried-out with varying V for each archetype — all other model attributes were held constant (i.e. ysky =1.0, and Inh. m-2 floor area for residential archetypes). The building morphology simulations were conducted as follows: Step 1: The V percentiles (5th, 30th, 50th, 70th, 95th) of all buildings (for each archetype) were extracted according to Section 2.5. Step 2: The corresponding volumes for each building were then modeled in DesignBuilder according to two morphological categories: 2a) Stacked archetypes (Industrial, Civic, Office, Retail) Due to large variation in building morphology within each archetype category, building footprint AF and depth LD were parametrically modeled alongside V (adjusted to represent the range in building size and shape). First, V was used to predict AF by regression analysis (Figure 3-5), which was in-turn used to predict building depth by regression analysis (Figure 3-6). Regression coefficients and floor-tofloor heights are presented in Table 3-1 and building morphology equations explained in Section 2.5.2. This initial building form was then automatically adjusted according to the archetype’s floor-tofloor height. 55  Results and Discussion  2b) Ground-oriented archetypes (SDDs, Duplex, Multiplex, Rowhouse) Due to less variation in building morphology, smaller number of floors and assumed floor-to-floor height standards, ground-oriented residential archetypes were scaled along building length alone. Specifically, building depth and height were kept constant while length was adjusted alongside V. Although this approach limits the inclusion of allometric relationships, it allows for floor-to-floor standard heights to be included in the modeling framework — an important attribute in ground-oriented archetypes where variation in height amongst buildings is reduced. Step 3: Five annual simulations were conducted for each archetype based on the V extracted in Step 1, resulting in a total of 60 simulations.  56  Results and Discussion  Figure 3-5 Building morphology: footprint AF vs. volume V LiDAR derived building volume and footprint area are plotted together. The dimensions of a scaling cube (i.e. isometric example) are plotted in grey. Fits above the cube plot indicate footprint areas grow more rapidly with volume than one would see in a cube.  57  Results and Discussion  Figure 3-6 Building morphology: depth vs. footprint LiDAR derived building footprint area and depth are plotted together. The dimensions of a scaling cube (i.e. isometric example) are plotted in grey. Fits below the cube plot indicate building depths grow more slowly with volume than one would see in a cube.  58  Results and Discussion  Figure 3-7 illustrates (for all archetypes) the relationship between building volume V (m-3)and EUIV (MJ m-3 yr-1). The correlation is relatively strong among most archetypes (Table 3.4). In order to determine the best metric to scale building morphology effects archetypes were further sorted by characteristic length LC (Figure 3-8). Although LC is derived from V, it takes into account the shape of the building through the consideration (inclusion in calculation) of building surface area. When plotting EUIV and LC even stronger relationships emerge (Table 3-6), with R2 values improved for 8 of the 12 archetypes. Both V and LC plots confirm expectations of increased EUIV in smaller less compact building forms (Mahdavi and Pessenlehner, 2003). This result however, is sensitive to Vancouver’s climate and the envelope performances of individual archetypes. The strongest correlations between V and EUIV are found in residential dwellings where space heating is the dominant load (e.g. Residential Detached 1965-1990, R2 = 0.90; Residential Rowhouse, R2 = 0.90; Residential Lowrise/Midrise, R2 = 0.93). The relationship between space heating and V is further substantiated if we plot the space heating component of EUIV against V (Figure 3-8, Table 3.5), where we find a dramatic improvement in the correlation of residential detached 1990-2009 archetype. Furthermore, some of the unexplained scatter may be further clarified by the exclusion of multiple building orientations in the current modeling scheme. Previous studies have shown both an exponential decay (Mahdavi and Pessenlehner, 2003) and a linear relationship (Papadopoulos et al., 2002) between LC and space heating demand. Mahdavi and Pessenlehner however, modeled a much wider spectrum of LC and thus results presented in Figure 3-9 may only start to outline the true relationship.  59  Results and Discussion  Figure 3-7 Energy use intensity EUIV vs. volume Modeled EUIV versus varying building volume. Linear fits correspond to the regression statistics presented in Table 3.4.  Table 3-4 Energy use intensity EUIV vs. volume Regression statistics for modeled EUIV versus varying building volume.  60  Results and Discussion Residential Detached 1900-1964  Residential Detached 1965-1990 − 1  80  120  110 2  50  1200  − 3  − 3  40 2  3  Volume V (m )  4  2.0× 10  − 1  year )  − 3  60 55 50  − 3  − 1  2  3  Volume V (m )  − 1  year )  − 1  60 55 50 RMSE = 8.9 MJ m 4 0 2.0× 104  -2.0× 10  Space heating (MJ m  − 1  year )  − 3  Space heating (MJ m  65  RMSE = 6.9 MJ m year , r = 0.480 4 4 0 2× 10 4× 10  45  = -0.0 V + 37.7  4  6× 10  30  − 3  − 1  2  year , r = 0.102 4  6.0× 10  35 30 25 20 15  − 3  − 1  2  RMSE = 5.9 MJ m year , r = 0.577 0 2000 4000 6000 3 Volume V (m )  = -0.0 V + 43.2 45  3  Volume V (m )  40 35 30  5  1.0× 10  = -0.0 V + 20.3  20  10  0  1200  40  − 3  − 3  − 1  2  RMSE = 7.4 MJ m year , r = 0.479 4 4 4 4 4 0 1× 10 2× 10 3× 10 4× 10 5× 10 3  Volume V (m )  − 1  2  RMSE = 3.7 MJ m year , r = 0.586 4 3 0 1.0× 10 5.0× 10 3  Volume V (m )  Commmercial Retail  = -0.0 V + 61.4  2  − 3  65  45  − 1  RMSE = 3.5 MJ m year , r = 0.863 200 400 600 800 1000 3 Volume V (m )  Civic  = -0.0 V + 59.2  year )  70  Commercial Office  45  2500  4  1.5× 10  Light Industrial − 1  − 1  2  − 3  − 3  − 1  RMSE = 10.5 MJ m year , r = 0.680 0 500 1000 1500 2000 3 Volume V (m )  Space heating (MJ m  − 1 − 3  50  70  − 3  100 − 3  − 3  40  Residential Midrise-Highrise  60  RMSE = 11.3 MJ m year , r = 0.637 4 4 3 0 1.0× 10 1.5× 10 5.0× 10  Space heating (MJ m  120  Residential Lowrise-Midrise = -0.0 V + 71.4  70  30  50  Residential Rowhouse  140  2000  year )  80  2  Space heating (MJ m  − 3  − 1  year )  90  Space heating (MJ m  Space heating (MJ m  50  − 1  1200  = -0.0 V + 152.4  − 1  year )  − 1  year )  − 3  Space heating (MJ m  60  − 3  2  60  Residential Multiplex  70  RMSE = 4.1 MJ m year , r = 0.856 500 1000 1500 3 Volume V (m )  − 1  RMSE = 3.1 MJ m year , r = 0.889 400 600 800 1000 3 Volume V (m )  Residential Duplex = -0.0 V + 80.1  40  70  = -0.0 V + 80.6  − 3  60  year )  − 1  70  Space heating (MJ m  − 3  RMSE = 5.8 MJ m year , r = 0.630 200 400 600 800 1000 3 Volume V (m )  Space heating (MJ m  Space heating (MJ m  Space heating (MJ m  − 3  − 3  130  Residential Detached 1991-2009  = -0.0 V + 94.0  year )  − 1  year )  − 1  year )  = -0.0 V + 137.1  90  = -0.0 V + 73.4  80 70 60 50 40  − 3  − 1  2  RMSE = 11.9 MJ m year , r = 0.454 4 4 0 2× 10 4× 10 3  4  6× 10  Volume V (m )  Figure 3-8 Energy use intensity EUIV (space heating) vs. volume Modeled EUIV (space heating only) versus varying building volume. Linear fits correspond to the regression statistics presented in Table 3.5.  Table 3-5 Energy use intensity EUIV (space heating) vs. volume Regression statistics for modeled EUIV (space heating only) versus varying building volume.  61  Results and Discussion  Figure 3-9 Energy use intensity EUIV vs. characteristic length LC Modeled EUIV versus varying characteristic length. Linear fits correspond to the regression statistics presented in Table 3.6.  Table 3-6 Energy use intensity EUIV vs. characteristic length LC Regression statistics for modeled EUIV versus varying characteristic length.  62  Results and Discussion  Although LC could be an ideal scaling factor given the strong correlation between LC and EUIV in many of the archetypes, extracting surface area from LiDAR is suspect due to the ‘edge effect’ (See Section 2.2.3). The extraction of V however is more robust and given the relatively adequate correlation found between V and EUIV, V was selected as the most appropriate scaling factor. Accordingly, the association between V and EUIV has been approximated linearly:  EUI = αV V + βV  (29) Volume derived EUI  Where EUI = Energy use intensity EUIV (MJ m-3 yr-1), aV = volumetric coefficient, V = building volume (m3), and bV = base EUIV (MJ m-3 yr-1).  3.1.4  Influence of population density  In order to determine each archetype’s sensitivity to population density and subsequent DHW demand, five simulations were carried-out with varying population densities for each archetype, where all other model attributes were held constant (i.e. mean values for V and ySVF). The population density simulations were conducted accordingly: Step 1: The volumetric population density (Inh. m-3) percentiles (5th, 30th, 50th, 70th, 95th) of all buildings (for each archetype) were extracted according to Section 2.4 and converted into population per conditioned floor area (Inh. m-2). Step 2: The corresponding values for population density and DHW demand per occupant (Section 2.7.4; 139 L Inh.-1) were adjusted in DesignBuilder accordingly. Step 3: An annual simulation for each population density was then conducted for all residential archetypes resulting in a total of 40 simulations.  63  Results and Discussion  Figure 3-10 shows (for all residential archetypes) the relationship between population density P (Inh. m-3)and EUIV (MJ m-3 yr-1). Not surprisingly there is a near perfect correlation between P and EUIV. This is due to the direct relationship between modeled DHW demand and P in the BEM. Beyond DHW, there is a slight offset to space heating demand with increased P, this however is negligible when compared to increased energy demand for DHW. The relative demand for DHW in residential archetypes was on average 28-48% of the total EUIV, the second largest demand of energy use besides space heating. Lastly, socioeconomic and behavior are expected to have further impact on DHW use but are currently excluded from the model.  Figure 3-10 Energy use intensity EUIV vs. population density Modeled EUIV versus varying population density. Linear fits correspond to the regression statistics presented in Table 3.7.  64  Results and Discussion Table 3-7 Energy use intensity EUIV vs. population density Regression statistics for modeled EUIV versus varying population density.  The association between P and EUIV can be approximated linearly:  EUI = αP (P − P) + βP  (30) Population derived EUI  Where EUI = Energy use intensity EUIV (MJ m-3 yr-1), aP = population coefficient, P = population density (Inh. m-3), P = average archetype population density (Inh. m-3), and bP = base EUIV at average archetype population density.  3.1.5  Up-scaling energy and emissions  The intent here was to develop a methodology that could rapidly provide context-sensitive EUILA, CO2e and heat emissions estimates for a large number of buildings based on a relatively small number of actual BEM simulations. Following this goal, each building (assigned to an archetype) was scaled according to (the discussed) urban context, building morphology and population density sensitivities. Accordingly, the following final linearization was proposed:  EUI = αS S + αP (P − P) + αV V + βV  (31) equation for upscaling  Where EUI = Energy use intensity EUIV (MJ m-3 yr-1), aS = shading coefficient, S = fractional shading, aP = population coefficient, P = population density (Inh. m-3), P = average archetype population density (Inh. m-3), aV =  65  Results and Discussion  volumetric coefficient, V = building volume (m-3), bV = base EUIV. Furthermore, CO2e emissions factors were based on the BC Greenhouse Gas Emissions Assessment Guide (2008), which provides emissions factors for common building energy sources in BC: 1) Natural Gas: 0.051 tCO2e/ GJ, 2) Electricity: 0.0000022 tCO2e/kwh 3) Heating Oil: 0.00284 tCO2e/litre , 4)Propane: 0.00154 tCO2e/litre (BCa, 2008). All emissions factors were assumed constant within each archetype and fuel source split were derived from the CEEI (2007) and OEE (2007) for each archetype. Lastly, all heat released from buildings is simply mapped (MJ converted to Watts sensible heat) and assumes consistent heat emission from the V. Once EUIV and subsequent CO2e and heat emissions had been mapped along the UT, flux densities were extracted and presented for the entire UT (Section 3.2). The flux density (i.e per gross land area) was calculated as follows:  FE,C,H = IE,C,H ·  V AGL  (32) equation for calculating flux  Where FE,C,H = Flux density (MJ m-2 yr-1), IE,C,H = Energy use intensity (MJ m-3 yr1  ), CO2e (kg CO2e m-3 yr-1), and Heat (W m-3 yr-1 ), V = building volume (m-3),  and AGL = gross land area. Table 3.7 shows (for all archetypes) the final EUIV along with the corresponding scaling sensitivities and average CO2e and heat emissions intensities used to create the maps in Section 3.2.  66  Results and Discussion Table 3-8 Archetype attributes: Part 3 Average modeled EUI, CO2e emissions, and heat emissions for each archetype. Sensitivities are calculated as the % difference (±) from the average archetype EUI (for all buildings within each archetype category).  67  Results and Discussion  3.2  Upscaled energy and emissions  The following section presents modeled results for the UT. The results are broken down into transect maps with corresponding profile plots and subset neighbourhoods (DT, MP, SS) with corresponding tables. Each subset represents a small (3%) fraction of the total transect area yet is intended to provide insight into three representative neighbourhoods along the transect. The sensitivity to all scaling factors is presented in Table 3.9, which outlines the importance of each within the context of varying urban density. Table 3-9 Neighbourhood scaling sensitivity Modeled sensitivities calculated as the % difference (±) from the average archetype energy intensity (for all buildings within each archetype category).  3.2.1  Modeled energy use  According to the scaling approach used to spatialize energy demand, EUILA is on average 512 MJ m-2 yr-1 along the entire UT, of which 57.8% (295 MJ m-2 yr-1) is consumed in residential buildings and the remaining 42.2% (216 MJ m-2 yr-1) in Commercial, Civic, Institutional and Light Industrial buildings. Energy use in the Central Business District (CBD) is significantly higher than the UT average and at times averages four times larger peaking at 2467 MJ m-2 yr-1 (Figure 3-11). Once out of the CBD and across the False Creek, EUILA drops due to decreased building density besides one significant peak at the commercial arterial street Broadway. Along this arterial there are several large commercial, lowrise and mixed-use buildings and development is especially built up near the Broadway Cambie Street intersection. Further  68  Results and Discussion  south along the UT there are two small rises in EUILA which correspond to increased development along Main Street near King Edward and along Fraser Street near East 33rd Avenue. A useful product of spatializing building energy use is the opportunity to define contextually appropriate energy and CO2e emissions mitigation strategies. For example, the potential for district energy systems (DES). A DES localizes the production of energy for a neighbourhood, most often supplying radiant heat through a network of steam or hot water — although some advanced systems may also provide electricity (combined heat and power CHP) or space cooling. When properly installed, a DES can reduce infrastructure costs, utilize waste heat and integrate multiple clean energy sources. Thermal energy use intensity per land area (TEILA), when considered alongside the temporal variability in demand and energy source can provide a good indication of the suitability of a DES in a neighbourhood. Of the total energy use along the UT, 51% (261 MJ m-2 yr1  ) is thermal energy — here defined as space heating and domestic how  water demand — and the remaining 49% (251 MJ m-2 yr-1) is used in lighting, appliance, system, and space cooling demand. The peak in TEILA (1060 MJ m-2 yr-1, 800 m South) follows the peak in total energy intensity (600 m South) as development transitions from large commercial buildings to residential Highrises along the waterfront. Southwards from DT, the TEILA follows a similar profile to EUILA, with several more delayed (south of EUILA peak) peaks corresponding to residential development and the subsequent increased demand for space heating and DHW. The relationship between residential buildings and thermal energy fraction can also be seen in the Figure 3-12, where the thermal energy fraction stays around 50% south of MP besides one large dip at 4.7 km where several Civic institutions and the Mountain View Cemetery dominate much of the UT.  69  Results and Discussion  Building Energy Use 94.3 Building height (meters)  0.57 Building fraction  0.39 Sky view factor  290 People hectare-1  311.5 Kg CO2e m-2 yr-1  12754 MJ m-2 yr-1  403 W m-2 yr-1  17.9  0.47  0.46  166  31.8  1353  67.8  9.7  0.36  0.50  104  18.2  712  26.3  7.7  0.30  0.55  90  14.4  558  17.3  6.7  0.26  0.57  78  12.4  473  13.8  6.1  0.25  0.59  68  11.7  421  12.2  5.8  0.22  0.60  58  10.3  372  10.8  5.5  0.19  0.63  49  8.8  311  9.3  5.4  0.16  0.67  35  7.3  245  7.7  5.0  0.11  0.70  9  5.1  113  5.5  0.0  0  0.94  0  0  0  0  DT  Downtown (DT)  Cambie Street  MP  SS  Mount Pleasant (MP)  Broadway  Main Street  50  150 250  375  100 300 500  750  Meters  King Edward Avenue  Fraser Street  N  Building energy density (MJ m-2 yr -1)  3000  2500  Building energy density Moving average  2000  Thermal energy density  1500  1000  Sunset (SS)  500  0  0  1  DT  2  3  4  5  MP  6  7  SS  Distance South on transect (km)  Figure 3-11 Modeled building energy use map Modeled building energy use presented at a 50 x 50 m grid cell resolution.  70  Results and Discussion  Building Thermal Energy Use 94.3 Building height (meters)  0.57 Building fraction  0.39 Sky view factor  290 People hectare-1  311.5 Kg CO2e m-2 yr-1  12754 MJ m-2 yr-1  403 W m-2 yr-1  311.5 Kg CO2e m-2 yr-1  4146 MJ m-2 yr-1  17.9  0.47  0.46  166  31.8  1353  67.8  31.8  517  9.7  0.36  0.50  104  18.2  712  26.3  18.2  343  7.7  0.30  0.55  90  14.4  558  17.3  14.4  283  6.7  0.26  0.57  78  12.4  473  13.8  12.4  247  6.1  0.25  0.59  68  11.7  421  12.2  11.7  218  5.8  0.22  0.60  58  10.3  372  10.8  10.3  184  5.5  0.19  0.63  49  8.8  311  9.3  8.8  149  5.4  0.16  0.67  35  7.3  245  7.7  7.3  106  5.0  0.11  0.70  9  5.1  113  5.5  5.1  37  0.0  0  0.94  0  0  0  0  0  0  DT  Downtown (DT)  Cambie Street  MP  SS  Mount Pleasant (MP)  Broadway  Main Street  50  150 250  375  100 300 500  750  Meters  King Edward Avenue  Fraser Street  N  Building thermal energy density  1000  Moving average Thermal energy fraction  800  0.70 0.60  600  0.50 0.40  400  0.30 0.20  200  Thermal energy fraction  Building thermal energy density (MJ m-2 yr -1)  1200  Sunset (SS)  0.10 0  0  1  DT  2  3  4  5  MP  6  7  SS  Distance South on transect (km)  Figure 3-12 Transect raster of thermal energy Modeled building energy use (space heating and domestic hot water) presented at a 50 x 50 m grid cell resolution.  71  Results and Discussion  Building energy density (MJ m-2 yr -1)  3000  2500  Building energy density Moving average  2000  Thermal energy density  1500  1000  500  0  0  1  2  3  4  5  6  7  SS  MP  DT  Distance South on transect (km)  Figure 3-13 Transect plot of energy use Total building energy density along the urban transect. Values plotted are east-west row averages.  Building thermal energy density  1000  Moving average Thermal energy fraction  800  0.70 0.60  600  0.50 0.40  400  0.30 0.20  200  Thermal energy fraction  Building thermal energy density (MJ m-2 yr -1)  1200  0.10 0  0  1  DT  2  3  4  5  MP  6  7  SS  Distance South on transect (km)  Figure 3-14 Plot of thermal and thermal energy Building thermal energy density along the urban transect. Values plotted are east-west row averages.  72  Results and Discussion  Looking at the three subsets we can start to see some of the driving forces of EUILA differences between development densities and composition. For example, along the UT, energy use in the DT subset is the highest at 1615 MJ m-2 yr-1, energy use in MP 438 MJ m-2 yr-1 and 302 MJ m-2 yr-1 in SS. When looking at the DES potential of each of the subsets we can compare average thermal density to existing projects, DT is estimated at 2.29 GWh ha-1 yr-1 , MP 0.78 GWh ha-1 yr-1 and 0.56 GWh ha-1 yr-1. Recent projects in Vancouver have been built with average thermal densities of 0.80 - 2.9 GWh ha-1 yr-1 (Compass, 2011). Furthermore, when looking at the spatial pattern of energy use we see unique urban form patterns emerge across neighbourhoods. For example, the threshold for a successful DES system would be met in the majority of the areas in DT, as opposed to large areas of MP and SS with lower EUILA. That said, the pattern of energy use in MP is built up between Ontario Street and Quebec Street, which is dominated by Lowrise residential development and may provide the density needed due to the proximity of development. Contrasting this, the energy use pattern at SS is more disaggregated and the spatial pattern dictated more by occupant behavior, building construction and system efficiency than by landuse or development density.  73  Results and Discussion  Figure 3-15 Neighbourhood energy density Annual energy use presented in 50 x 50 grid cell format. From left to right: Downtown, Mount Pleasant and Sunset neighbourhoods.  Table 3-10 Building energy use Subset and transect total energy use and average energy flux. The thermal energy fraction is calculated as the sum of space heating and domestic hot water loads.  Although building occupant density and scheduling are accounted for in each of the archetypes, variability in energy use due to socioeconomic factors is not included in the current model. These factors can play a significant role alongside urban context, building morphology and population density and should be considered when investigating energy reduction strategies (Lutzenhiser, 1993; Clevenger et a., 2006; Santin et al., 2009 Maruejols and Young, 2010; Allcott, 2011).  3.2.2  Modeled CO2e emissions  According to the scaling approach used to spatialize building CO2e emissions, buildings on average account for 14.2 kg CO2e m-2 yr-1 along the UT, of which 65% (9.2 kg CO2e m-2 yr-1) is from Residential buildings and the remaining 35% (5.0 kg CO2e m-2 yr-1) from Commercial, Civic, and Light  74  Results and Discussion  Industry. An estimated 12.6 kg CO2e m-2 yr-1 is emitted locally (within the UT) due to the combustion of fossil-fuels for space heating and DHW, of which 96.8% is from natural gas, 1.8% from propane and 1.4% from heating oil. The estimated 1.6 kg CO2e m-2 yr-1 emitted externally (i.e. outside the transect due to electricity consumption inside the transect) is attributed to fossil-fuels used in the production of electricity (BCb, 2008). Lastly, although present, wood-burning is assumed to be an insignificant heating source when compared to more conventional fossil-fuel and electricity sources and has been omitted from the current modeling framework — wood burning is also absent from the 2007 Vancouver CEEI. In order for municipalities to reach emissions reduction targets, a good characterization of CO2e emissions patterns is necessary. Emissions can be aggregated and compared against population density to benchmark reduction targets to similar cities. However, for the purposes of spatially comparing residential development patterns along the transect, per capita emissions in this paper are determined solely from Residential emissions. This approach allows for a direct comparison amongst Residential development and excludes Commercial, Industrial and Civic development. Transportation emissions, a vital component to a comprehensive comparison have been left out due to the BEM focus of this research (e.g. Newman and Kenworthy, 1987, 2006) and urban form variation (e.g. Calthorpe, 2010). The per capita residential emissions profile follows a different path than the total emissions profile, reflecting differences in population density, building type and morphology (Figure 3-17). The average per capita residential emissions along the UT is 1.30 tonnes CO2e capita-1 yr-1, with the lowest (0.85 tonnes capita-1 yr-1 ) found between DT and MP where building type transitions from Residential Highrise to primarily Rowhouse and Lowrise/ Midrise development. This is in part due to the reduced built volume per person in Rowhouse (209 m3 capita-1) and Lowrise (206 m3 capita-1)  75  Results and Discussion  buildings when compared to Highrise buildings (223 m3 capita-1). The highest per capita emissions, which fluctuate between 1.4 and 1.6 tonnes CO2e capita-1 yr-1 are found just after MP, where the primary building type transitions from older SDD and Multiplex dwellings to SDD of all vintage categories. This increase however, is not as steep as one might expect due to the high proportion (55%) of secondary suites found along the UT — in part due to the high demand for rental housing in Vancouver. This results in an average of 4.4 persons per single detached dwelling, much higher than the Canadian average of 2.6 persons per dwelling. Lastly, when building emissions from all landuses (i.e. Residential, Commercial, Civic, Industrial) are added to the emissions total, the per capita emissions rises to 2.03 tonnes CO2e capita-1 yr-1. Although this number agrees extremely well the CEEI (2.02 tonnes CO2e capita-1 yr-1) it should be taken with caution as the UT does not necessarily represent the actual landuse mix of Vancouver nor does it include emissions from on-road transportation or solid waste — which represent a large portion (52%) of the 4.2 tonnes CO2e capita-1 yr-1 estimated for the City of Vancouver (BCa, 2008). Furthermore, municipal totals also discount emissions attributable to agriculture, large industry and deforestation — total emissions (including all sectors) are estimated to be 14.2 tonnes CO2e capita-1 yr-1 in British Columbia (BCc, 2008).  76  Results and Discussion  CO2e Emissions 94.3 Building height (meters)  0.57 Building fraction  0.39 Sky view factor  290 People hectare-1  311.5 Kg CO2e m-2 yr-1  12754 MJ m-2 yr-1  403 W m-2 yr-1  17.9  0.47  0.46  166  31.8  1353  67.8  9.7  0.36  0.50  104  18.2  712  26.3  7.7  0.30  0.55  90  14.4  558  17.3  6.7  0.26  0.57  78  12.4  473  13.8  6.1  0.25  0.59  68  11.7  421  12.2  5.8  0.22  0.60  58  10.3  372  10.8  5.5  0.19  0.63  49  8.8  311  9.3  5.4  0.16  0.67  35  7.3  245  7.7  5.0  0.11  0.70  9  5.1  113  5.5  0.0  0  0.94  0  0  0  0  DT  Downtown (DT)  Cambie Street  MP  SS  Mount Pleasant (MP)  Broadway  150 250  375  100 300 500  750  Main Street  50  Meters  King Edward Avenue  Fraser Street  N  Total emissions density (kg CO2e m-2 yr -1)  70 60 Total emissions Moving average  50  Electricity emissions 40 30  Sunset (SS)  20 10 0  0  1  DT  2  3  4  5  MP  6  7  SS  Distance South on transect (km)  Figure 3-16 Transect raster of CO2e emissions Modeled building CO2e emissions presented at a 50 x 50 m grid cell resolution.  77  Results and Discussion  Per Capita CO2e Emissions S  94.3 Building height (meters)  0.57 Building fraction  0.39 Sky view factor  290 People hectare-1  311.5 Kg CO2e m-2 yr-1  12754 MJ m-2 yr-1  403 W m-2 yr-1  138 m  1.00 ψSVF  1.00  17.9  0.47  0.46  166  31.8  1353  67.8  124  0.90  0.90  1728  9.7  0.36  0.50  104  18.2  712  26.3  110  0.80  0.80  1628  7.7  0.30  0.55  90  14.4  558  17.3  96  0.70  0.70  1525  6.7  0.26  0.57  78  12.4  473  13.8  82  0.60  0.60  1428  6.1  0.25  0.59  68  11.7  421  12.2  68  0.50  0.50  1375  5.8  0.22  0.60  58  10.3  372  10.8  54  0.40  0.40  1305  5.5  0.19  0.63  49  8.8  311  9.3  40  0.30  0.30  1228  5.4  0.16  0.67  35  7.3  245  7.7  26  0.20  0.20  1142  5.0  0.11  0.70  9  5.1  113  5.5  12  0.10  0.10  995  0.0  0  0.94  0  0  0  0  0  0  0  0  4664  Kg Inh.-1 yr-1  DT  Downtown (DT)  Cambie Street  MP  SS  Mount Pleasant (MP)  Broadway  Main Street  50  150 250  375  100 300 500  750  Meters  King Edward Avenue  Fraser Street  N  40 Residential emissions  35  Moving average  30  Per capita moving average  2.5  25 2.0 20 1.5  15  1.0  10  0.5  5 0  0  1  DT  2  3  4  5  MP  6  Tonnes CO2 e per capita  Residential emissions density (kg CO2e m-2 yr -1)  45  Sunset (SS)  7  SS  Distance South on transect (km)  Figure 3-17 Transect raster of CO2e emissions (per capita) Modeled building CO2e emissions (per capita) presented at a 50 x 50 m grid cell resolution.  78  Results and Discussion  Total emissions density (kg CO2e m-2 yr -1)  70 60 Total emissions Moving average  50  Electricity emissions 40 30 20 10 0  0  1  2  3  4  5  6  7  SS  MP  DT  Distance South on transect (km)  Figure 3-18 Transect plot of CO2e emissions Total building CO2e emissions density along the urban transect. Values plotted are east-west row averages.  40 Residential emissions  35  Moving average  30  Per capita moving average  2.5  25 2.0 20 1.5  15  1.0  10  0.5  5 0  0  1  DT  2  3  4  5  MP  6  Tonnes CO2 e per capita  Residential emissions density (kg CO2e m-2 yr -1)  45  7  SS  Distance South on transect (km)  Figure 3-19 Transect plot of CO2e emissions (per capita) Total building CO2e emissions (per capita) density along the urban transect. Per capita totals only include residential building emissions. Values plotted are east-west row averages.  79  Results and Discussion  Looking at the three subsets we can start to see some of the driving forces among emissions differences. For example, on a land area basis, the DT subset CO2e emissions are the highest of the three neighbourhoods (40.2 kg CO2e m-2 yr-1). This is largely driven by built-up Highrise development (49% of built volume), Office (15.8%), Retail (15.8%) and mixed-use (10.2%) development. A good proportion of DT residential Highrises are in the North False Creek neighbourhood, an area which has been completely redeveloped since Expo ‘86. The larger proportion of electric heating found in Highrise development and more compact dwelling density result in the lowest per capita emissions of the three subset areas (1.08 tonnes CO2e capita-1 yr-1). The MP subset on the other hand, is an older neighbourhood with greater building diversity where all archetypes are represented except Highrise and SDD 1965-90. The greater proportion of older dwellings results in a slightly higher per capita emissions (1.25 tonnes CO2e capita-1 yr-1). This estimation however should be taken with caution as the inclusion of energy retrofits has not been considered due to the difficulty in characterizing such renovations — although this is quite possible as many heritage dwellings have been converted in recent years to accommodate multiple tenants. On the contrary, the SS area has a large proportion SDD (92.3% of built volume), 32.6% built before 1965, 33.7% built between 1965 and 1990, and 26% built after 1990, the remaining 7.3% is Civic. The greater proportion of newer SDD and above average population per building result in 1.22 tonnes CO2e capita-1 yr-1, lower than one might expect in typical suburban development in British Columbia. For example, in a recent urban archetypes report, NRCAN reports for a typical SDD in Coquitlam BC, an average of 47-93 m2 floor area per person and 2.2-3.4 tonnes CO2e capita-1 yr-1, compared to the average SDDs modeled in this study, which on average have 36-50 m2 floor area per person and 1.1-1.7 tonnes CO2e capita-1 yr-1.  80  Results and Discussion  Figure 3-20 Neighbourhood CO2e density Annual CO2e emissions presented in a 50 x 50 grid cell format. From left to right: Downtown, Mount Pleasant and Sunset neighbourhoods.  Table 3-11 Building CO2e emissions CO2e emissions and average CO2e flux. The kg CO2e Inh-1 yr-1 is calculated as the sum of all Residential CO2e emissions (Res), including the proportion Residential in Mixed-use buildings, divided by the number of inhabitants within the study area. Values without the prefix Res include all buildings (i.e. Residential, Commercial, Industrial and Civic buildings).  3.2.3  Modeled QFB  The modeled results in this section follow a similar methodology to that presented by Sailor (2011), who outlines an ideal approach for modeling QFB. Sailor recommends a combination of BEM and inventory approaches (See Section 1.4) because “it enables inclusion of time-dependant occupancy, energy use, environmental loads, and HVAC schedules in buildings (Sailor, 2011). This paper differs slightly from the framework provided by Sailor in two respects: 1) the use of LiDAR data to inform the BEM, and 2) the scaling approach used to provide context-sensitive emissions results.  81  Results and Discussion  According to the scaling approach used to spatialize QFB, buildings on average emit 16.2 W m-2 along the UT, of which 59% (9.6 W m-2) is from Residential buildings and the remaining 41% (6.6 W m-2) from Commercial, Civic, and light Industry. The profile of QFB varies dramatically, with profile averages ranging from 2 - 71 W m-2 (Figure 3-22). In comparison, Ichinose reports for Tokyo, QFB values of 200-400 W m-2 and a max cell value of 1540 W m-2 (Ichinose, 1999). The QFB values modeled in this study are more comparable to those estimated by Sailor in Lu (2004) for 6 US cities (5 - 75 W m-2 , including building, transportation, human metabolism emissions). Additionally, QFB estimates from recent research has been summarized in Table 3.11. It should be noted that the research summarized in Table 3.11 includes heat flux densities from all three urban components (buildings, transportation, human metabolism) where as modeled results presented in here only include buildings and human metabolism.  Table 3-12 Anthropogenic heat flux case studies Selected research that has either modeled or measured the anthropogenic heat flux in cities. *Results presented in van der Laan (2011) are the neighbourhood totals for Sunset, Mout Pleasant and Downtown. These values are buildings only as opposed to other research in this table which quantifies the total flux (buildings, transportation, and human metabolism). BEM = building energy model, EBC = energy balance closure.  82  Results and Discussion  The advantage of the results presented in Figure 3-21 is the ability to aggregate QFB emissions up from the building scale to the desired cell resolution necessary for atmospheric modeling input. Furthermore, the integration of LiDAR data allows for detailed patterns of QFB to be spatialized rapidly. The author proposes the use of LiDAR data to improve the downscaling of inventory data. Specifically, V provides a good estimate of heated volume in a city and as outlined in Section 2 can be combined with convention data sets such as population and landuse data. Also, the morphological data extractable from LiDAR allows for the comparison of QFB with urban form attributes (Appendix 6.2).  83  Results and Discussion  Building Heat Emissions (QFB) 94.3 Building height (meters)  0.57 Building fraction  0.39 Sky view factor  290 People hectare-1  311.5 Kg CO2e m-2 yr-1  12754 MJ m-2 yr-1  403 W m-2 yr-1  17.9  0.47  0.46  166  31.8  1353  67.8  9.7  0.36  0.50  104  18.2  712  26.3  7.7  0.30  0.55  90  14.4  558  17.3  6.7  0.26  0.57  78  12.4  473  13.8  6.1  0.25  0.59  68  11.7  421  12.2  5.8  0.22  0.60  58  10.3  372  10.8  5.5  0.19  0.63  49  8.8  311  9.3  5.4  0.16  0.67  35  7.3  245  7.7  5.0  0.11  0.70  9  5.1  113  5.5  0.0  0  0.94  0  0  0  0  DT  Downtown (DT)  Cambie Street  MP  SS  Mount Pleasant (MP)  Broadway  150 250  375  100 300 500  750  Main Street  50  Meters  King Edward Avenue  Fraser Street  N  80  Building heat emissions, QFB (W m-2)  70 Building heat emissions  60  Moving average 50 40 30  Sunset (SS)  20 10 0  0  1  DT  2  3  4  5  MP  6  7  SS  Distance South on transect (km)  Figure 3-21 Transect raster of QFB emissions Modeled building QFB emissions presented at a 50 x 50 m grid cell resolution.  84  Results and Discussion  80  Building heat emissions, QFB (W m-2)  70 Building heat emissions  60  Moving average 50 40 30 20 10 0  0  1  DT  2  3  4  5  MP  6  7  SS  Distance South on transect (km)  Figure 3-22 Transect plot of QFB emissions Total building QFB emissions density along the urban transect. Values plotted are east-west row averages.  Figure 3-23 Neighbourhood QFB density Annual QFB emissions presented in a 50 x 50 grid cell format. From left to right: Downtown, Mount Pleasant and Sunset neighbourhoods.  85  Results and Discussion Table 3-13 Anthropogenic heat emissions QFB for neighbourhoods, transect and Residential sector. The Residential total W and W m-2 is calculated as the sum of all residential buildings including the residential proportion of mixeduse buildings. Values without the prefix Res include all buildings (i.e. Residential, Commercial, Industrial and Civic buildings).  3.2.4  Comparing modeled and measured  In Table 3.11 modelled energy and CO2e emissions are compared to the 2007 CEEI for the City of Vancouver. The CEEI report provides a highlevel inventory of energy consumption and CO2e emissions to help inform local governments develop targets and policies in accordance with the Local Government Green Communities Statues Amendment Act Bill 27, 2008 (BCa, 2008). From this comparison we can see the 50923 residents of the UT represent 8.1% of Vancouver’s total population and live on approximately 7.4% (783 ha.) of Vancouver’s total land area. This results in an average population density of 68.8 Inh. ha-1, which is slightly higher than the population density of the municipality of Vancouver (62.9 Inh. ha-1). This close agreement in population density highlight that although the UT traverses a wide range in population density it may present a representative snapshot of Vancouver’s development pattern. When we compare energy and emissions, we can start to see the relative performance of the model. The total energy use modeled in the UT represents 10.4% (4121176 GJ yr-1) of total building energy consumption in Vancouver. In contrast, the modeled CO2e emissions represent 8.2% (102008 tonnes CO2e yr-1) of the building-sector emissions. These modeled results are encouraging when considered alongside total population. For instance, per capita emissions compare extremely well (Transect 2.00  86  Results and Discussion  tonnes CO2e yr-1 , Vancouver 1.98 tonnes CO2e yr-1). That said, beyond a simple comparison of relative magnitude, model performance can only truly be evaluated with either: 1) a better urban surface characterization in future CEEI reports, or 2) extending the LiDAR extent to include the entire municipality. This modeling approach has also been evaluated against directly measured carbon emissions over two years (Christen et al, 2011). In this study directly measured carbon dioxide fluxes were compared to modeled transportation, buildings, human metabolism, and vegetation components. The comparison produced encouraging results, where total modeled emissions (7.46 kg C m-2 yr-1) agreed within 11% of directly measured carbon emissions (6.71 kg C m-2 yr-1). Further research is warranted to separate carbon fluxes into their respective components (buildings, transportation, human metabolism and vegetation) to properly assess the building submodel. Nonetheless this research presents a unique study that has been benchmarked against both inventory data and directly measured data. Table 3-14 Energy and emissions comparison Total modeled energy and CO2e emissions are presented alongside Vancouver Community Energy and Emissions (2007). Results in grey are compared as proportion (Urban transect contribution to Vancouver’s total). The Per capita comparison (kg CO2e Inh-1) contains all building-sector emissions.  3.2.5  Scaling sensitivity  When averaged over the 7813 buildings, annual energy use sensitivity to local population density was 3.2%, building morphology 2.8% and urban context 2.8% — a significant proportion of total energy use. This research has shown that the sensitivities to these three factors varies widely across  87  Results and Discussion  development density and type, with the highest density neighbourhood (DT) influenced largely by building morphology 7.8%, the medium density neighbourhood (MP) by urban context 4.4% and the least dense neighbourhood (SS) by population density 6.1%. Furthermore, the DT study site was significantly influenced by all three scaling factors and when added together the sensitivity was 18.3%, much larger than MP (10.8%) and SS (10.0%).  88  Conclusion  4  Conclusion  This research has shown the successful integration of LiDAR data in the mapping and quantification of neighbourhood to communityscale building energy and GHG emissions estimates. Furthermore, the scaling-approach used in this research, has been shown to be a viable methodology, that links LiDAR derived metrics with context-sensitve results. In particular, a small number of simulations (n = 160) have been used to up-scale energy and emissions estimates to a large number of buildings (n = 7812). This was made possible through a stepped scaling of modeled archetype sensitivity to local population density, building morphology, and urban context. Furthermore, the scaling of these three factors was applied to a small number of archetypes (n = 12) and compared against directly measured emissions and community-scale inventories. This research has shown that by integrating LiDAR derived morphology with building energy simulation, a wide variety of energy metrics can be output for analysis. For example, the thermal energy density for the UT has been shown to be approximately 261 MJ m-2 yr-1 or 51% or total energy density (512 MJ m-2 yr-1). Such energy density estimates presented (along with the developed methodology) have the potential to improve community energy planning (e.g. DES suitability) and energy demand forecasting for utilities. Likewise, a transect of anthropogenic heat flux from buildings QFB has been presented. Specifically, heat emissions from buildings averaged 16.2 W m-2 yr-1 of which 59% (9.6 W m-2 yr-1 ) was attributable to residential buildings. The heat emissions estimates presented (along with the developed methodology) have the potential to improve urban climate research, specifically improved atmospheric model input, benchmark surface energy balance models and improve urban weather prediction. Furthermore, the integration of LiDAR data in the mapping of QFB provides great promise in  89  Conclusion  improving the spatial resolution of building-sector and human metabolism emissions.  4.1 The  Practical significance  outlined  research  supports  the  development  of  increasingly  sophisticated projections of future environmental conditions, sustainability scenarios, back casting methods (what reduction targets look like), and visioning techniques (spatializing energy and emissions data). The methodology developed represents an innovative spatialization of energy and emissions in an urban environment in the hopes of contributing to a swift response to BC emissions reduction goals by providing benchmarked and locally relevant data needed to inform adaptation and mitigation strategies. Through the development of a refined method for characterizing building-sector CO2 emissions, policy makers will be able to target areas with the most dramatic change potential.  4.2  Recommendations for future work  Without considering effective development densities necessary to promote walkable  neighbourhoods  and  alternative  transportation  choices,  emissions reduction targets are largely unachievable. Thus, a successful comparison between development density and CO2e emissions must also consider the transportation implications if valuable relationships are to be derived. Although this study may start to illustrate important differences amongst a variety of building archetypes and development densities, actual building energy use is complicated by several externalities, notably building occupant behavior. Furthermore, many building types can be built to a passivhaus1 standard if the proper design decisions are considered. Although transportation and socioeconomic impacts on energy use are  1  A rigorous, voluntary standard for energy efficiency in a building (e.g. Parker, 2009) 90  Conclusion  beyond the scope of this study, these topics may warrant future integration into the modeling methodology. In this case, a general scaling approach may be appropriate for an easily extendable framework. For instance, the integration of socioeconomic factors and transportation mode-split could be integrated as modules alongside the three already investigated sensitivities (population density, building morphology, urban context). Furthermore, energy use sensitivity to landuse diversity, proximity and connectivity could be integrated into an adapted archetype approach. It is important to note that although EnergyPlus is a comprehensive energy modeling software, it represents a simplification of the real world and therefore results presented should not be considered all-encompassing. In particular, the building stock has been broken down into 12 archetypes in an attempt to model energy use and determine particular sensitivities amongst a variety of building types. It should aslo be noted that no computational fluid dynamic (CFD) analyses was conducted in the modeling environment and only a simple sensitivity to wind sheltering effects has been included. Considerations for future research: • Model structure: 1) Create a stand-alone program or add-on to a software package that extends the scaling approach to include transportation and socioeconomic impacts. 2) Consider automatic integration of LiDAR derived morphology attributes into the BEM (e.g. GBXML schema). 3) Extend urban context calculations by estimating solar load falling on building envelope. 4) Consider integration of directly measured climate data to suit micro-climatic conditions (i.e manipulate weather file to include directly measured data or couple BEM with an UCL model, for example, Unzeta et al., 2009). • Model capability: 1) Consider explicit effects from vegetation  91  Conclusion  (e.g. scheduling vegetation components). 2) Consider the integration of local energy potential by overlaying LiDAR derived solar potential (currently derived without sky conditions) and wind potential (by including an urban CFD analysis). 3) Extend population density estimates to include non-residential estimates (e.g. a transportation model based on commercial area, residential population and archetype occupation schedules). • Model implementation: 1) Upscale methodology to an entire City, where CEEI data is available to further benchmark model performance. 2) Benchmark against metered data from all building utilities at the neighbourhood-scale (i.e. natural gas, heating oil, propane, electricity use, and DES).  92  Conclusion  References Aguilar, C., and White, D. (2005). Domestic water heating and water heater energy consumption in Canada. Canadian Building Energy End-Use Data and Analsis Centre. Akbari, H. (2002). Shade trees reduce building energy use and CO2 emissions from power plants. Environmental Pollution. Aleklett, K., Hook, M., Jakobsson, K., Lardelli, M., Snowden, S., and Soderbergh, B. (2010). The Peak of the Oil Age - Analyzing the world oil production Reference Scenario in World Energy Outlook 2008. Energy Policy, 38(3), pp. 1398–1414. Allcott, H. (2011). Social norms and energy conservation. Journal of Public Economics, 95(9-10), pp. 1082–1095. Asawa, T., Yamamura, S., & Hoyano, A. (2010). Prediction of sensible heat flux from buildings and urban spaces using a detailed geometry model of a substantial urban area - introduction of a prediction model of anthropogenic heat into an urban heat balance simulation model, in: The seventh International Conference on Urban Climate, pp. 1–4. Attia, S., Beltrán, L., De Herde, A., and Hensen, J. (2009). “Architect friendly”: Acomparison of ten different building performance simulation tools, in: Eleventh International IBPSA Conference. Glasgow: Eleventh International IBPSA Conference. Baldocchi, D. (2003). Assessing the eddy covariance technique for evaluating the carbon balance of ecosystems. Global Change Biology. Batty, M., Carvalho, R., Hudson-Smith, A., Milton, R., Smith, D., and Steadman, P. (2008). Scaling and allometry in the building geometries of Greater London. The European Physical Journal B, 63(3), pp. 303–314. Bayer, T. (2009). Automated Building Simplification Using a Recursive Approach. Cartography in Central and Eastern Europe. Bazjanac, V. (2001). Acquisition of building geometry in the simulation of energy performance. Building Simulation. BC (2008a). Ministry of the Environment. Community Energy and Emissions Inventory: Reports. British Columbia Provincial Government. Victoria, BC; http:// www.env.gov.bc.ca/cas/mitigation/ceei/reports.html BC (2008b). Ministry of the Environment. British Columbia Climate Action Plan. Victoria, BC; http://www.livesmartbc.ca/government/plan.html BC (2008c). Community Energy Association and the Ministry of Community Services, 2008. Greenhouse Gas Emission Assessment Guide: For British Columbia Local Governments; http://www.localmotion.gov.bc.ca/docs/ghg_assessment_ guidebook_feb_2008.pdf Brown, G., and DeKay, M. (2000). Sun, Wind and Light: Architectural Design Strategies, 2nd Edition. 2nd ed. Wiley.  93  Conclusion Calthorpe, P (2010). Urbanism in the light of climate change. Island Press. Washington, DC. CanmetENERGY (2009). Urban Archetypes Project. Natural Resources Canada; http://canmetenergy-canmetenergie.nrcan-rncan.gc.ca/eng/buildings_ communities/communities/urban_archetypes_project.html CHASS (2007). Canadian Census Analyser. University of Toronto; http://dc.chass. utoronto.ca/census/ Cheng, V., Steemers, K., Montavon, M., and Compagnon, R. (2006). Urban form, density and solar potential, in: The 23rd Conference on Passive and Low Energy Architecture. City of Vancouver (2011). Open Data Catalogue Beta V2. Vancouver, BC, Canada; http://data.vancouver.ca/ Christen, A., & Vogt, R. (2004). Energy and radiation balance of a Central European city. International Journal of Climatology, 24, pp. 1395–1421. Christen, A., Coops, N. C., Crawford, B. R., Kellett, R., Liss, K. N., Olchovski, I., Tooke, T.R., van der Laan, M. (2011). Validation of modeled carbon-dioxide emissions from an urban neighborhood with direct eddy-covariance measurements. Atmospheric Environment, 45(33), pp. 6057–6069. Clevenger, C. (2006). The Impact of the Building Occupant on Energy Modeling Simulations. Decision Making in Civil and Building. CMHC. Renovating for Energy Savings — Case Studies. Canada Mortgage and Housing Corporation; http://www.cmhc-schl.gc.ca/en/co/renoho/reensa/ index.cfm Compass, 2011. City of Revelstoke: District Energy Expansion Pre-feasibility Study; http://www.cityofrevelstoke.com/DocumentView.aspx?DID=180 Coppock, J. (1991). The history of GIS. Geographical information systems Crawford, B., Grimmond, C. S. B., and Christen, A. (2011). Five years of carbon dioxide fluxes measurements in a highly vegetated suburban area. Atmospheric Environment, 45(4), pp. 896–905. Crawley, D., Hand, J., Kummert, M., and Griffith, B. (2008). Contrasting the capabilities of building energy performance simulation programs. Building and Environment, 43(4), pp. 661–673. Davies, M., Steadman, P., and Oreszczyn, T. (2008). Strategies for the modification of the urban climate and the consequent impact on building energy use. Energy Policy, 36(12), pp. 4548–4551. de Almeida, P., and Silva, P. D. (2009). The peak of oil production-Timings and market recognition. Energy Policy, 37(4), pp. 1267–1276. DeOreo, W.B., and Mayer P.W.(2000), “The End Uses of Hot Water in Single Family Homes from Flow Trace Analysis”, Aquacraft Inc. Report, undated.  94  Conclusion DOE, 2009. EnergyPlus Manual, Engineering Reference. United States Department of Energy Elementsdb. School of Architecture and Landscape Architecture, University of British Columbia; http://elementsdb.sala.ubc.ca/ Fan, H., & Sailor, D. J. (2005). Modeling the impacts of anthropogenic heating on the urban climate of Philadelphia: a comparison of implementations in two PBL schemes. Atmospheric Environment, 39(1), pp. 73–84. Feigenwinter, C., Vogt, R., Christen, A. (2012): ‘Eddy covariance measurements over urban areas’ (Chapter 16) in Aubinet M., Vesala T., Papale D. [Eds.] ‘Eddy Covariance - A Practical Guide to Measurement and Data Analysis’, Springer, ISBN 978-94-007-2350-4, 270 p Florides, G. (2002). Measures used to lower building energy consumption and their cost effectiveness. Applied Energy, 73(3-4), pp. 299–328. Goodwin, N.R., Tooke, T.R., Christen, A., and Voogt, J.A. (2009). Characterizing urban surface cover and structure with airborne lidar technology. Can. J. Remote Sensing, 35(3), pp. 297–309. Grimmond, C. (1992). The suburban energy balance: methodological considerations and results for a mid-latitude west coast city under winter and spring conditions. International Journal of Climatology, 12(5). Guerra Santin, O., Itard, L., and Visscher, H. (2009). The effect of occupancy and building characteristics on energy use for space and water heating in Dutch residential stock. Energy and Buildings, 41(11), pp. 1223–1232. Gupta, V. (1987). Thermal efficiency of building clusters: An index for non airconditioned buildings in hot climates. Energy and urban built form. Hamilton, I., Davies, M., Steadman, P., Stone, A., Ridley, I., and Evans, S. (2009). The significance of the anthropogenic heat emissions of London’s buildings: A comparison against captured shortwave solar radiation. Building and Environment, 44(4), pp. 807–817. Harrison, R., McGoldrick, B., and Williams, C. (1984). Artificial heat release from Greater London, 1971–1976. Atmospheric Environment, 18(11), pp. 2291–2304. Heiple, S., and Sailor, D.J. (2008). Using building energy simulation and geospatial modeling techniques to determine high resolution building sector energy consumption profiles. Energy and Buildings, 40(8), pp. 1426–1436. Hegger, M., Fuchs M., Stark T., Zeumer M. (2008). Energy Manual: Sustainable Architecture, 1st Edition. Brickhauser Architecture Ichinose, T., Shimodozono, K., and Hanaki, K. (1999). Impact of anthropogenic heat on urban climate in Tokyo. Atmospheric Environment, 33(24-25), pp. 3897–3909. IEA (2011a). Key World Energy Statistics. International Energy Agency, OECD/ IEA,Paris. IEA (2011b). CO2 Emissions from Fuel Combustion. International Energy Agency, OECD/IEA,Paris. 95  Conclusion IPCC, Climate Change (2007). Mitigation. Contribution of Working Group III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Press Cambridge, United Kingdom, New York, NY, USA Kikegawa, Y., Genchi, Y., Yoshikado, H., and Kondo, H. (2003). Development of a numerical simulation system toward comprehensive assessments of urban warming countermeasures including their impacts upon the urban buildings’ energy-demands. Applied Energy, 76(4), pp. 449–466. Klysik, K. (1996). Spatial and seasonal distribution of anthropogenic heat emissions in Lodz, Poland. Atmospheric Environment, 30, pp. 3397–3404. Knowles, R. (2003). The solar envelope: its meaning for energy and buildings. Energy and Buildings, 35(1), pp. 15–25. Kondo, H., and Kikegawa, Y. (2003). Temperature variation in the urban canopy with anthropogenic energy use. Pure And Applied Geophysics, 160(1-2), pp. 317–324. Kusuda, T. (1999). Early History and Future Prospects of Building System Simulation. IBPSA, pp. 1–13. Lam K., Huang Y., Zhai C. (2004) Energy Modeling Tools Assessment For Early Design Phase. Center for Building Performance and Diagnostics School of Architecture Carnegie Mellon University. Pittsburgh, PA Lee, S., Song, C., Baik, J., and Park, S. (2009). Estimation of anthropogenic heat emission in the Gyeong-In region of Korea. Theoretical and Applied Climatology, 96(3), pp. 291–303. Lindberg, F., and Grimmond, C.S.B. (2011). The influence of vegetation and building morphology on shadow patterns and mean radiant temperatures in urban areas: model development and evaluation. Theoretical and Applied Climatology. Littlefair, P. (1998). Passive Solar Design. Renewable and Sustainable Energy Reviews, pp. 1–24. Lutzenhiser, L. (1993). Social and behavioral aspects of energy use. Annual Review of Energy and the Environment. Mahdavi, A., and Pessenlehner, W. (2003). Building morphology, transparence, and energy performance, in: Eighth International IBPSA Conference. Eindhoven. Martin, L., and March, L. (1972), Urban Space and Structures, Cambridge University Press, UK. Maruejols, L., and Young, D. (2010). Energy use in Canadian Multifamily Dwellings. Canadian Building Energy End-Use Data Analysis Centre. http://www.cbeedac. com/publications/index.html Moriwaki, R., and Kanda, M. (2004). Seasonal and diurnal fluxes of radiation, heat, water vapor, and carbon dioxide over a suburban area. Journal of Applied Meteorology and Climatology, 43(11), pp. 1700–1710.  96  Conclusion Neidhart, H. (2004). Identifying building types and building clusters using 3D-laser scanning and GIS-data. Geo-Imagery Bridging Continents. Nemitz, E., Hargreaves, K., and McDonald, A. (2002). Micrometeorological measurements of the urban heat budget and CO2 emissions on a city …. Environmental Science and Technology. Newman, P., and Kenworthy, J. (1989). Gasoline Consumption and Cities: A Comparison of U.S. Cities with a Global Survey. Journal of the American Planning Association, Vol. 55, No. 1, pp. 24–37. Newman, P and Kenworthy, J (1999). Costs of Automobile Dependence: Global Survey of Cities. Transportation Research Record 1670, pp. 17–26. Newman, P and Kenworthy, J (2006). Urban Design to Reduce Automobile Dependence. Opolis: An International Journal of Suburban and Metropolitan Studies, Vol. 2, No. 1, pp. 35–52 OEE (2008). Energy Use Data Handbook Tables. Office of Energy Efficiency, Natural Resources Canada; http://oee.nrcan.gc.ca/corporate/statistics/neud/dpa/ handbook_tables.cfm Ohashi, Y., Genchi, Y., Kondo, H., Kikegawa, Y., Yoshikado, H., and Hirano, Y. (2007). Influence of air-conditioning waste heat on air temperature in Tokyo during summer: Numerical experiments using an urban canopy model coupled with a building energy model. Journal of Applied Meteorology and Climatology, 46(1), pp. 66–81. Oke, T.R. (1982). The energetic basis of the urban heat island. Quarterly Journal of the Royal Meteorological Society, 108(455), pp. 1–24. Oke, T.R. (1987). Boundary Layer Climates, 2nd Edition. Methuen, London. Oke, T.R. (2004), Urban Observations, Instruments and Methods of Observation Programme, IOM Report, World Meteorol. Organiz., Geneva Papadopoulos, A. (2002). Feasibility of energy saving renovation measures in urban buildings The impact of energy prices and the acceptable pay back time criterion. Energy and Buildings, 34(5), pp. 455–466. Parekh, A. (2005). Development of archetypes of building characteristics libraries for simplified energy use evaluation of houses, in: Ninth Internation IBPSA Conference. Parker, D. S. (2009). Very low energy homes in the United States: Perspectives on performance from measured data. Energy and Buildings, 41(5), pp. 512–520. Parshall, L., Gurney, K., Hammer, S. A., Mendoza, D., Zhou, Y., and Geethakumar, S. (2010). Modeling energy consumption and CO2 emissions at the urban scale: Methodological challenges and insights from the United States. Energy Policy, 38(9), pp. 4765–4782. Pataki, D., Emmi, P., Forster, C., Mills, J., Pardyjak, E., Peterson, T., et al. (2009). An integrated approach to improving fossil fuel emissions scenarios with urban ecosystem studies. Ecological Complexity, 6(1), pp. 1–14.  97  Conclusion Perez R., Ineichen P., Seals R., Michalsky J,. Stewart R., Modeling daylight availability and irradiance components from direct and global irradiance, Solar Energy, Volume 44, Issue 5, 1990, Pages 271-289 Pigeon, G., Legain, D., Durand, P., & Masson, V. (2007). Anthropogenic heat release in an old European agglomeration (Toulouse, France). International Journal of Climatology, 27(14), pp. 1969–1981. Purdy, J., Beausoleil-Morrison. (2001). The significant factors in modelling residential buildings, in: Seventh International IBPSA Conference. Quah, A., & Roth, M. (2009). Spatial and temporal distribution of anthropogenic heat emissions in Singapore, in: The seventh International Conference on Urban Climate. Yokohama. Ratti, C., Di Sabatino, S., Britter, R., Brown, M., Caton, F., and Burian, S. (2002). Analysis of 3-D urban databases with respect to pollution dispersion for a number of European and American cities. Water, Air, and Soil Pollution: Focus, 2(5), pp. 459–469. Ratti, C., Baker, N., and Steemers, K. (2005). Energy consumption and urban texture. Energy and Buildings, 37(7), pp. 762–776. Ratti, C., Di Sabatino, S., and Britter, R. (2006). Urban texture analysis with image processing techniques: winds and dispersion. Theoretical and Applied Climatology, 84(1), pp. 77–90. Ratti, C., Raydan, D., and Steemers, K. (2003). Building form and environmental performance: archetypes, analysis and an arid climate. Energy and Buildings, 35(1), pp. 49–59. Raupach, M. R., Rayner, P. J., and Paget, M. (2010). Regional variations in spatial structure of nightlights, population density and fossil-fuel CO2 emissions. Energy Policy, 38(9), pp. 4756–4764. Robinson, D. (2006). Urban morphology and indicators of radiation availability. Solar Energy, 80(12), pp. 1643–1648. Sailor, D. J. (2011). A review of methods for estimating anthropogenic heat and moisture emissions in the urban environment (M. Roth, R. Emmanuel, T. Ichinose, and J. Salmond, Eds.). International Journal of Climatology, 31(2), pp. 189–199. Sailor, D. J., and Lu, L. (2004). A top-down methodology for developing diurnal and seasonal anthropogenic heating profiles for urban areas. Atmospheric Environment, 38(17), pp. 2737–2748. Salamanca, F., Krpo, A., Martilli, A., and Clappier, A. (2010). A new building energy model coupled with an urban canopy parameterization for urban climate simulations—part I. formulation, verification, and sensitivity analysis of the model. Theoretical and Applied Climatology, 99(3-4), pp. 331–344. Salat, S. (2009). Energy loads, CO 2 emissions and building stocks: morphologies, typologies, energy systems and behaviour. Building Research And Information, 37(5), pp. 598–609.  98  Conclusion Sarkar, A. (2009). Low energy urban block: Morphology and planning guidelines, in: 45th ISOCARP Congress. Satterthwaite, D. (2008). Cities’ contribution to global warming: notes on the allocation of greenhouse gas emissions. Environment and Urbanization, 20(2), p. 539. Shashua-Bar, L., Hoffman, M., and Tzamir, Y. (2006). Integrated thermal effects of generic built forms and vegetation on the UCL microclimate. Building and Environment, 41(3), pp. 343–354. Simpson, J. (2002). Improved estimates of tree-shade effects on residential energy use. Energy and Buildings, 34(10), pp. 1067–1076. Smith, C., Lindley, S., and Levermore, G. (2009). Estimating spatial and temporal patterns of urban anthropogenic heat fluxes for UK cities: the case of Manchester. Theoretical and Applied Climatology, 98(1-2), pp. 19–35. Statistics Canada (2007). Census 2006 Population and Dwelling Counts Release No. 1. UN Population Division (2011). World Population Prospects; The 2010 Revision. United Nations, New York; http://esa.un.org/unpd/wpp/index.htm Steadman, P. (2006). Allometry and built form: revisiting Ranko Bon’s work with the Harvard Philomorphs. Construction Management and Economics, 24(7), pp. 755–765. Steadman, P., Evans, S., and Batty, Michael. (2009). Wall area, volume and plan depth in the building stock. Building Research And Information, 37(5), pp. 455– 467. Steemers, K. (2003). Energy and the city: density, buildings and transport. Energy and Buildings, 35(1), pp. 3–14. Stewart, I. (2011). Redefining the urban heat island. PhD disseration. University of British Columbia. Strømann-Andersen, J., and Sattrup, P. A. (2011). Urban Density and Building Energy Use: The impact of height/width ratios of urban canyons. Energy and Buildings. Taha, H. (1997). Urban climates and heat islands: albedo, evapotranspiration, and anthropogenic heat. Energy and Buildings, 25, pp. 99–103. Tooke, T.R. (2011). Classification of residential building architectural typologies using LiDAR. Urban Remote Sensing Event (JURSE), 2011 Joint, pp. 221–224. Tooke, T.R., Coops, N., Goodwin, N. R., and Voogt, J. A. (2009). Extracting urban vegetation characteristics using spectral mixture analysis and decision tree classifications. Remote Sensing of Environment, 113(2), pp. 398–407. Torrance, K. E., and Shum, J. S. W. (1976). Time-varying energy consumption as a factor in urban climate. Atmospheric Environment, 10(4), pp. 329–337. Unzeta, B., and Norford, L. (2009). An Urban Weather Generator Coupling Building Simulations with a Physically Based Urban Model, in: The seventh International Conference on Urban Climate.  99  Conclusion van der Laan M., Tooke R., Christen A., Coops N., Heyman E., Olchovski I. (2011): ‘Statistics on the built infrastructure at the Vancouver EPiCC experimental sites’. EPiCC Technical Report No. 4, 30pp VandeWeghe, J. R., and Kennedy, C. A. (2007). A spatial analysis of residential greenhouse gas emissions in the Toronto Census Metropolitan Area. Journal of Industrial Ecology, 11(2), pp. 133–144. Vogt, R., Christen, Andreas, Rotach, M., Roth, M., and Satyanarayana, A. (2006). Temporal dynamics of CO 2 fluxes and profiles over a Central European city. Theoretical and Applied Climatology, 84(1), pp. 117–126. Yu, B, Liu, H, and Wu, J. (2009). Investigating impacts of urban morphology on spatio-temporal variations of solar radiation with airborne LIDAR data and a solar flux model: a case study of downtown Houston. International journal of remote sensing, 30(17-18), pp. 4359–4385. Yu, Bailang, Liu, Hongxing, Wu, Jianping, Hu, Y., and Zhang, L. (2010). Automated derivation of urban building density information using airborne LiDAR data and object-based method. Landscape and Urban Planning, 98, pp. 210–219. Zizzo, R., and Kennedy, C. (2010). Designing an optimal urban community mix for an aquifer thermal energy storage system, in: ACEEE Summer Study on Energy Efficiency in Buildings.  100  Conclusion  Appendices A.1 Maps This section contains additional maps used in thesis analysis. Several maps (Figure A-4 - 6.7) are derived from LiDAR using image processing techniques and others are products of integrated datasets such as Census, BCPA and LiDAR (Figure A-2, 6.3. 6.8).  101  Conclusion  Aerial Photography  DT  Downtown (DT)  MP  SS  Mount Pleasant (MP)  50  150 250  375  100 300 500  750  Meters  N  Sunset (SS)  Figure A-1 Transect aerial map Aerial image of transect. Source: Vancouver opendata 2009 orthophoto  102  Conclusion  Figure Ground  DT  Downtown (DT)  MP  SS  Mount Pleasant (MP)  50  150 250  375  100 300 500  750  Meters  N  0.50 Plan area fraction Moving average  Plan area fraction  0.40  0.30  0.20  Sunset (SS)  0.10  0.00  0  1  DT  2  3  4  5  MP  6  7  SS  Distance South on transect (km)  Figure A-2 Transect figure ground Building footprints and corresponding plan area plot.  103  Conclusion  Building Landuse Residential detached Residential attached Residential stacked Commercial Civic Light Industrial Mixed-use Vacant / Secondary DT  Downtown (DT)  MP  SS  Mount Pleasant (MP)  50  150 250  375  100 300 500  750  Meters  N  Sunset (SS)  Figure A-3 Transect land use map Land use map including archetypes (i.e Residential attached: Duplex, Multiplex, Rowhouse; Residential stacked: Lowrise/Midrise, Midrise/Highrise; Commercial: Retail, Commercial).  104  Conclusion  LiDAR visualization  DT  Downtown (DT)  MP  SS  Mount Pleasant (MP)  50  150 250  375  100 300 500  750  Meters  N  Sunset (SS)  Figure A-4 Transect obstruction angle map Obstruction view factor (horizon) — one of 24 used in skyview/fractional shading calculations.  105  Conclusion  Digital Elevation Model (DEM) S  94.3 Building height (meters)  0.57 Building fraction  0.39 Sky view factor  290 People hectare-1  311.5 Kg CO2e m-2 yr-1  12754 MJ m-2 yr-1  403 W m-2 yr-1  138 m  1.00 ψSVF  1.00  17.9  0.47  0.46  166  31.8  1353  67.8  124  0.90  0.90  1728  9.7  0.36  0.50  104  18.2  712  26.3  110  0.80  0.80  1628  7.7  0.30  0.55  90  14.4  558  17.3  96  0.70  0.70  1525  6.7  0.26  0.57  78  12.4  473  13.8  82  0.60  0.60  1428  6.1  0.25  0.59  68  11.7  421  12.2  68  0.50  0.50  1375  5.8  0.22  0.60  58  10.3  372  10.8  54  0.40  0.40  1305  5.5  0.19  0.63  49  8.8  311  9.3  40  0.30  0.30  1228  5.4  0.16  0.67  35  7.3  245  7.7  26  0.20  0.20  1142  5.0  0.11  0.70  9  5.1  113  5.5  12  0.10  0.10  995  0.0  0  0.94  0  0  0  0  0  0  0  0  DT  Downtown (DT)  MP  SS  Mount Pleasant (MP)  50  150 250  375  100 300 500  750  Meters  N  Mean building height, ZH (Meters)  30 Mean building height 25  Moving average  20  Sunset (SS)  15 10 5 0  0  1  DT  2  3  MP  4  5  7 km South on transect  6  SS  Figure A-5 Transect digital elevation model 2.5D urban surface with corresponding plot of building heights along the transect.  106  4664  Kg Inh.-1 yr-1  Conclusion  Sky View Factor S  94.3 Building height (meters)  0.57 Building fraction  0.39 Sky view factor  290 People hectare-1  311.5 Kg CO2e m-2 yr-1  12754 MJ m-2 yr-1  403 W m-2 yr-1  138 m  1.00 ψSVF  1.00  17.9  0.47  0.46  166  31.8  1353  67.8  124  0.90  0.90  4664 1728  9.7  0.36  0.50  104  18.2  712  26.3  110  0.80  0.80  1628  7.7  0.30  0.55  90  14.4  558  17.3  96  0.70  0.70  1525  6.7  0.26  0.57  78  12.4  473  13.8  82  0.60  0.60  1428  6.1  0.25  0.59  68  11.7  421  12.2  68  0.50  0.50  1375  5.8  0.22  0.60  58  10.3  372  10.8  54  0.40  0.40  1305  5.5  0.19  0.63  49  8.8  311  9.3  40  0.30  0.30  1228  5.4  0.16  0.67  35  7.3  245  7.7  26  0.20  0.20  1142  5.0  0.11  0.70  9  5.1  113  5.5  12  0.10  0.10  995  0.0  0  0.94  0  0  0  0  0  0  0  0  DT  Downtown (DT)  MP  SS  Mount Pleasant (MP)  50  150 250  375  100 300 500  750  Meters  Sky view factor  N  0.90  Sky view factor  0.80  Moving average  0.70  Sunset (SS)  0.60 0.50 0.40 0.30  0  1  DT  2  3  4  5  MP  6  7  SS  Distance South on transect (km)  Figure A-6 Transect raster of sky view factor Modeled sky view factor with corresponding transect averages.  107  Kg Inh.-1 yr-1  Conclusion  Shading factor S  94.3 Building height (meters)  0.57 Building fraction  0.39 Sky view factor  290 People hectare-1  311.5 Kg CO2e m-2 yr-1  12754 MJ m-2 yr-1  403 W m-2 yr-1  138 m  1.00 ψSVF  1.00  17.9  0.47  0.46  166  31.8  1353  67.8  124  0.90  0.90  4664 1728  9.7  0.36  0.50  104  18.2  712  26.3  110  0.80  0.80  1628  7.7  0.30  0.55  90  14.4  558  17.3  96  0.70  0.70  1525  6.7  0.26  0.57  78  12.4  473  13.8  82  0.60  0.60  1428  6.1  0.25  0.59  68  11.7  421  12.2  68  0.50  0.50  1375  5.8  0.22  0.60  58  10.3  372  10.8  54  0.40  0.40  1305  5.5  0.19  0.63  49  8.8  311  9.3  40  0.30  0.30  1228  5.4  0.16  0.67  35  7.3  245  7.7  26  0.20  0.20  1142  5.0  0.11  0.70  9  5.1  113  5.5  12  0.10  0.10  995  0.0  0  0.94  0  0  0  0  0  0  0  0  Kg Inh.-1 yr-1  DT  Downtown (DT)  MP  SS  Mount Pleasant (MP)  50  150 250  375  100 300 500  750  Meters  N  0.8 0.7 0.6  Shading factor S  0.5 0.4 0.3  Sunset (SS)  0.2 0.1 0  0 0  1  DT  2  3  4  5  MP  6  7  SS  Distance South on transect (km)  Figure A-7 Transect raster of fractional shading Transect raster of fractional shading values along the transect.  108  Conclusion  Population Density 94.3 Building height (meters)  0.57 Building fraction  0.39 Sky view factor  290 People hectare-1  311.5 Kg CO2e m-2 yr-1  12754 MJ m-2 yr-1  403 W m-2 y  17.9  0.47  0.46  166  31.8  1353  67.8  9.7  0.36  0.50  104  18.2  712  26.3  7.7  0.30  0.55  90  14.4  558  17.3  6.7  0.26  0.57  78  12.4  473  13.8  6.1  0.25  0.59  68  11.7  421  12.2  5.8  0.22  0.60  58  10.3  372  10.8  5.5  0.19  0.63  49  8.8  311  9.3  5.4  0.16  0.67  35  7.3  245  7.7  5.0  0.11  0.70  9  5.1  113  5.5  0.0  0  0.94  0  0  0  0  DT  Downtown (DT)  MP  SS  Mount Pleasant (MP)  50  150 250  375  100 300 500  750  Meters  N  Population density (persons hectare-1)  450 400 350  People per hectare Moving average  300 250 200 150  Sunset (SS)  100 50 0  0  1  DT  2  3  4  5  MP  6  7  SS  Distance South on transect (km)  Figure A-8 Transect raster of population density Downscaled population presented at a 50 x 50 m grid cell resolution with corresponding transect plot.  109  Conclusion  Morphometric QFB plots This section contains additional plots of anthropogenic heat flux density due to buildings QFB. The QFB data used in this analysis is spatially mapped in Figure 3-21 and each point represents one of the 50 x 50 m grid cells.  − 2  Building Heat Emissions Q FB (W m )  200  150  100  50  0  10  20 3 Volume V (m )  30  40  Figure A-9 Anthropogenic heat flux density vs. Building volume Raster plots of all cells at a 50 x 50 m grid cell resolution.  110  Conclusion  − 2  Building Heat Emissions Q FB (W m )  200  150  100  50  0  20  40 60 Building Height zH (m)  80  Figure A-10 Anthropogenic heat flux density vs. Building height Raster plots of all cells at a 50 x 50 m grid cell resolution.  − 2  Building Heat Emissions Q FB (W m )  200  150  100  50  0.0  0.2  0.4 0.6 Sky View Factor Ψ sky  0.8  Figure A-11 Anthropogenic heat flux density vs. Sky view factor Raster plots of all cells at a 50 x 50 m grid cell resolution.  111  Conclusion  − 2  Building Heat Emissions Q FB (W m )  200  150  100  50  0.01 0.02 0.03 0.04 − 2 Population density P (Inh. m )  0.05  Figure A-12 Anthropogenic heat flux density vs. Population density Raster plots of all cells at a 50 x 50 m grid cell resolution. All buildings included in anthropogenic heat flux density calculation.  − 2  Building Heat Emissions Q FB (W m )  200  150  100  50  0.0  0.2  0.4 0.6 Fractional shading S  0.8  1.0  Figure A-13 Anthropogenic heat flux density vs. Fractional shading Raster plots of all cells at a 50 x 50 m grid cell resolution.  112  Conclusion  − 2  Building Heat Emissions Q FB (W m )  200  150  100  50  0.0  0.2  0.4 0.6 Plan Area λb  0.8  1.0  Figure A-14 Anthropogenic heat flux density vs. Plan area Raster plots of all cells at a 50 x 50 m grid cell resolution.  113  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0072500/manifest

Comment

Related Items