Environmental Research Letters

LETTER • OPEN ACCESS

Rooftop solar photovoltaic potential in cities: howscalable are assessment approaches?To cite this article: Sergio Castellanos et al 2017 Environ. Res. Lett. 12 125005

 

View the article online for updates and enhancements.

Related contentUsing GIS-based methods and lidar datato estimate rooftop solar technicalpotential in US citiesRobert Margolis, Pieter Gagnon, JenniferMelius et al.

-

Evidence and future scenarios of a low-carbon energy transition in CentralAmerica: a case study in NicaraguaDiego Ponce de Leon Barido, JosiahJohnston, Maria V Moncada et al.

-

Was it worthwhile? Where have thebenefits of rooftop solar photovoltaicgeneration exceeded the cost?Parth Vaishnav, Nathaniel Horner andInês L Azevedo

-

This content was downloaded from IP address 73.162.99.248 on 23/12/2017 at 06:48

OPEN ACCESS

RECEIVED

16 March 2017

REVISED

29 May 2017

ACCEPTED FOR PUBLICATION

9 June 2017

PUBLISHED

11 December 2017

Original content fromthis work may be usedunder the terms of theCreative CommonsAttribution 3.0 licence.

Any further distributionof this work mustmaintain attribution tothe author(s) and thetitle of the work, journalcitation and DOI.

Environ. Res. Lett. 12 (2017) 125005 https://doi.org/10.1088/1748-9326/aa7857

LETTER

Rooftop solar photovoltaic potential in cities: how scalable areassessment approaches?

Sergio Castellanos1,2,3,5 , Deborah A Sunter1,2 and Daniel M Kammen1,2,4

1 Energy and Resources Group, University of California, Berkeley, Berkeley, CA 94720, United States of America2 Renewable and Appropriate Energy Laboratory, University of California, Berkeley, Berkeley, CA 94720, United States of America3 Berkeley Energy and Climate Institute, University of California, Berkeley, Berkeley, CA 94720, United States of America4 Goldman School of Public Policy, University of California, Berkeley, Berkeley, CA 94720, United States of America5 Author to whom any correspondence should be addressed.

E-mail: sergioc@berkeley.edu

Keywords: rooftop solar photovoltaic systems, solar photovoltaics, potential estimation, urban solar planning, built environment

AbstractDistributed photovoltaics (PV) have played a critical role in the deployment of solar energy,currently making up roughly half of the global PV installed capacity. However, there remainssignificant unused economically beneficial potential. Estimates of the total technical potential forrooftop PV systems in the United States calculate a generation comparable to approximately 40%of the 2016 total national electric-sector sales. To best take advantage of the rooftop PV potential,effective analytic tools that support deployment strategies and aggressive local, state, and nationalpolicies to reduce the soft cost of solar energy are vital. A key step is the low-cost automation ofdata analysis and business case presentation for structure-integrated solar energy. In this paper,the scalability and resolution of various methods to assess the urban rooftop PV potential arecompared, concluding with suggestions for future work in bridging methodologies to better assistpolicy makers.

1. Introduction

In response to the dramatic cost reductions in solarenergy and energy storage, the ease of buildingintegration, and increasing climate change risks,mitigation strategies involving renewable energydeployment have recently gained substantial traction.One low-carbon technology that has seen exponentialgrowth is solar photovoltaics (PV). PV deployment hasgrown by a factor of 40 in the last 10 years, and nowcomprises close to 300 GWof global installed capacity(Kurtz et al 2017), with growth projections pointingtowards approximately 430 GWby 2020, as reported bythe International Energy Agency (IEA) (2015).

Distributed PV has historically dominated thesolar industry, as seen in figure 1. However, thereremains a tremendous untapped potential for furtherdeployment. In fact, the total technical generationpotential for rooftop PV systems in the United Statesalone is estimated to be almost 40% of the of 2016 totalnational electric-sector sales (Gagnon et al 2016).

Some of the barriers that have hindered thedevelopment of distributed PV are, among others

© 2017 IOP Publishing Ltd

(Margolis and Zuboy 2006, Strupeit and Palm 2016),the lackof awarenessbyfinalusers andstakeholders, highlevels of riskaversion, systemperformanceconcerns, andlack of suitable rooftop space for installations (Schwartzet al 2017),with the greatest hindrance perhaps being thecombination of these barriers. Considerable researchefforts have pursued understanding, addressing, andsolving the multiple adoption barriers.

In the case of locating suitable rooftops, particularattention has been focused on urban areas for their highdensity of rooftops. As the world becomes more urban,with an expected influx of 2.5 billion people into urbanareas by 2050 (Department of Economic and SocialAffairs Population Division 2014), an increase in builtinfrastructure to support this influx is expected,rendering urban areas as critical venues for distributedPV deployment. Furthermore, Kammen and Sunter(2016) suggest that city-integrated PV could have thepotential to satisfy the energy needs of most cities ifcurrent advanced laboratory-tested PV technologies arecommercialized and become cost-competitive.

Past growth in distributed PV in urban areas hasbeen highly policy-dependent, and future growth may

[GW]

120

100

80

60

40

20

0

2000

2002

2004

2006

2008

2010

2012

2014

2016

Year

Cum

ulat

ive

Glo

bal I

nsta

lled

Cap

acity

DecentralizedCentralizedOff-grid

Figure 1. Evolution of cumulative global installed capacity (GW) from 2009–2015. The red line (squares) represents the cumulativecapacity for decentralized grid-connected PV installations. The blue line (circles) represents centralized grid-connected PV systems.The yellow line (triangles) represents off-grid PV installations which, while small in capacity, are expected to continue growing(Lighting Global and Bloomberg New Energy Finance 2016) (International Energy Agency 2000, 2016).

Environ. Res. Lett. 12 (2017) 125005

be as well (Schwartz et al 2017). Identifying andaccurately predicting PV potential, and communicat-ing this potential to often risk-adverse nontechnicalstakeholders is a challenge. Having accurate, accessi-ble, and easily understood tools to assess distributedPV potential estimates is, therefore, an expectedcomponent for appropriate policy development.

Thegoalof this letter is toaddress thequestions:whatis the reported scalability between rooftop PVassessmentmethods? and what is the expected deviation betweenreported methods of different spatial resolution? from apolicymaker’sperspective,andviaacomparativeanalysisoffewcities. Inthiscontribution,wefocussolelyonurbanareas.We provide a framework to categorizemethods toassess rooftop PV potential in cities, and evaluate theresults amongst them.Wefirstdevelopour frameworkbycomparing a selection of reported assessment methods,and the tradeoffs between the amountof individual citiesanalyzed and their spatial resolution. Next, we compareresults on PV rooftop potential from different method-ologies to determine their variations on selected cities.Lastly, we conclude with suggestions for future work inbridging methodologies to better assist policy makers intheir rooftop PVassessment efforts.

2. Methods

An initial assessment of review materials is performedusing Google Scholar with keywords such as ‘rooftopsolar’, ‘PV rooftop assessment’, ‘GIS PV rooftop’, ‘PVrooftop potential’, and permutations. Obtained resultsinclude peer-reviewed academic studies, conferenceproceedings, and professional reports from specializedagencies. Results are ordered by relevance and thenbroadened by analyzing the forward citations made

2

until the time of this publication. A narrowing downof the selected articles is then manually implementedby their relevance and application to cities, excludingcountry-wide, or region-wide aggregated results. Thismanual procedure may unintentionally omit somestudies; therefore, this letter is an overview of aselection of methods and not an exhaustive review ofall rooftop PV assessment research.

Reported PV potential estimates from literatureare captured and categorized based on the spatialresolution of the techniques and reported results, andthe cities covered in their contributions.

A methodological starting-point utilized by theIEA Energy Technology Perspectives report (2016) isto simply parameterize the rooftop PV potential by thepopulation density and solar insolation. This approachestimates the total PVrooftop potential per city by firstcalculating the rooftop area per capita, Acapita, usingthe population density, r, as stated in equation (1).The constants in this equation (awith a value of 172.3,and b with a value of 0.352) were found by the IEAafter performing a linear regression on 1600 cities witha correlation coefficient of 44% (IEA 2016)

Acapita ¼ a·r�b ð1ÞMultiplying equation (1) by the total city population,P, gives the suitable roof area per city, Acity, as shown inequation (2),

Acity ¼ Acapita·P: ð2ÞThe total electricity generation potential EPV,IEA is thencalculated, as shown in equation (3),

EPV;IEA ¼ Acity·H solar;city·h·PR·f orientation ð3ÞwhereHsolar,city is the solar insolation (kWh−1m−2 yr−1),h is the rooftop PV system efficiency, PR is the

Environ. Res. Lett. 12 (2017) 125005

performance ratio (assumed to be 75%, as indicated in(IEA 2016)), and forientation is the orientation factor(assumed to be 1 in aggregate, as indicated in (IEA2016)).

While equation (3) is very general and can beapplied to any city to estimate the rooftop PVpotential,many researchers have studied city-specific solutions. Aselection of these city-specific solutions have beengathered both from literature and from online sourcessuch as www.mapdwell.com (Mapdwell 2017) andGoogle Project Sunroof (Google 2017). To compare thecity-specific solutions to the corresponding IEAsolution, values for P, r, and h, used in equation (3),are consistent with those indicated in city-specificresearch methods. When these parameters are notspecified, thePV systemefficiency is assumed tobe 15%and the population and population density are foundusing city-specific statistics (Italian National Instituteof Statistics 2011, Korean Statistical InformationService 2010,Demographia 2016). The solar insolation,Hsolar,city, for each city is acquired from NASA’s SurfaceMeteorology and Solar Energy data (2014). Themethods are then compared in terms of their totalelectricity generation potential from rooftop PV andtheir percent difference from the IEA method, wherethe percent difference, D, is calculated as follows:

D ¼ EPV;IEA � EPV;city�specific

EPV;IEA·100% ð4Þ

3. Results and discussion

A simplified schematic of selected publicationsreporting rooftop PV assessment, somewhat similarin style to that reported by Mainzer et al (2014), isshown in figure 2. The schematic focuses solely onindividual cities and the spatial resolution of themethodology used for the results presented in eachwork. A total of 24 publications are incorporated. Thehierarchical methodology by Bergamasco and Asinari(2011a), in which the potential is categorized intophysical, geographical, theoretical, and energy exploi-tation, are accounted for at different levels of detail toguide in this classification. The number of citiesanalyzed by a specific research method is representedin the x-axis: from a fraction (e.g. district, or cityregion) to multiple cities, and three broad categories inthe y-axis described as ‘low’, ‘medium’, and ‘high’categorize the spatial resolution of the techniques andresults. A darker shade of green in the gradient-shadedbackground represents the optimal area to be locatedin the technique resolution–city coverage space: highlyresolved rooftop PV potential technique that isapplicable to many cities.

Low-level spatial resolution techniques and resultsare considered as those that rely mainly on aggregatedstatistical data that is assumed to be homogeneousthroughout the city analyzed. An example is IEA’s

3

approach described in the Energy TechnologyPerspectives report (2016) which aggregates statisticaldata frommore than 1500 cities. Similar approaches inthis category (Lehmann and Peter 2003, Kurdgelashviliet al 2016) correspond to regions where multipleGerman cities (extrapolated to Europe), and USAcities (with substantial state-level information),respectively, are evaluated on the basis of aggregateddata. Similarly, other works fall in this level (Wigintonet al 2010, Schallenberg-Rodríguez 2013), and othersbegin to blend with medium-level classification results(Nguyen and Pearce 2013, Karteris et al 2013).

Medium-level category is herein defined asapproaches that combine aggregated statistical datawith spatially-resolveddata acquired through geograph-ical informationsystems(GIS)and light-detection-and-ranging (LiDAR) approaches. As an example, Singh andBanerjee (2015)mixhigh-granularity landuse statisticaldata, GISmaps to calculate building footprint area, andcouple these findings with PV system performancesimulations to estimate rooftop PV potential inMumbai, India. Cole et al (2016) perform a rooftopassessment methodology in Plymouth, UK, (which isextrapolated afterwards) by extracting 3D urbanfeatures from medium resolution LiDAR data, andcombine with statistical scale-up methods fromindividual roofs within a segment of the city to applyto theentireUK. Similarly, by combiningGISdata that isused in solar shading calculation routines, the rooftopPV potential in Osaka, Japan is assessed by pairing thatinformation with surveyed data from building use andnumber of buildings on different categories by Take-bayashi et al (2015).

High-level category mostly comprise studies thatutilize advanced methods for rooftop digitization,insolation calculations, and accounting for aspects andshading of buildings. As an example, Bergamasco andAsinari (2011b) incorporate geographical and cadas-tral data in GIS, combine with computationalalgorithms for roof shading, topology, and surfaceoccupied in roofs across Turin, Italy. Hong et al (2016)utilize GIS maps for insolation calculations andbuilding suitability assessments, and calculate buildingshadows for the technical, physical, and geographicalrooftop PV potential assessment in the Gangnamdistrict in Seoul, Korea. Another example of a studythat falls in this category is that from Hofierka andKanuk (2009) who develop a GIS-based 3D model ofBardejov, Slovakia, and incorporate digital orthomapsand elevation models to study the city’s PV potential.Jakubiec and Reinhart (2013) utilized a suite of GISdata, LiDAR measurements and daylight simulations(Daysim engine) to accurately predict and validate arooftop PV output both within selected buildingswithin Cambridge, Massachusetts, and then the cityitself. This work ultimately lead to the development ofwww.mapdwell.com (Mapdwell 2017). In a similarvein, Google has developed Project Sunroof whichuses GIS data, 3D modeling derived from aerial

Fraction 1 > 10 > 100 > 1000

Low

Med

ium

Hig

h

Cities covered

Spat

ial R

esol

utio

n of

Tec

hniq

ues

and

Res

ults

for

Roo

ftop

PV P

oten

tial A

sses

smen

t Hong 2

016

Kodys

h 201

3

Santos

2014

Adams 2

016

Palmer

2016

Jaku

biec 2

013

Hofierk

a 200

9 Mapdw

ell

Najem 20

17

Ko 201

5Vardim

on 20

11

Bergam

asco

2011

(b)

Bergam

asco

2011

(b)

Bergam

asco

2011

(a)

Google

Sun

roof

IEA 20

16

Singh 2

015

Take

baya

shi 2

015

Karteri

s 201

3

Nguye

n 201

3

Wigi

nton 2

010

Schall

enbe

rg-Rod

rigue

z 201

3

Kurdge

lashv

ili 20

16

Lehm

ann 2

003

Figure 2. Spatial resolution of techniques and results for total rooftop PV potential across different cities. Increasing color opacitydenotes the location of the most desirable assessment of rooftop PV potential corresponding to a large number of cities covered at ahigh spatial resolution.

Environ. Res. Lett. 12 (2017) 125005

imagery, and shading calculations to predict PVenergygeneration potential at a rooftop level across hundredsof cities in the United States (Google 2017). Thesemethods can be more computationally intensive(Arnette 2013).

It is important to note that a breadth of literaturecan be justified as being placed in transition regions infigure 2. An example is the case of Najem (2017) whodoes not develop high-resolution techniques but ratherutilizes them in combination with additional data (e.g.road network topology), and provides a generalizedapproach to calculate PV rooftop potential.

A selection of studies are excluded from figure 2chiefly due to their broad regional, province, ornational scope (Izquierdo et al 2008, Lopez et al 2012,Ordóñez et al 2010), or their focus on only a segmentof building types (Gagnon et al 2016, Kurdgelashviliet al 2016).

By inspecting the spread of the literature datapoints across figure 2, methodologies that can coverthousands of cities at medium or high resolution arenoticeably lacking.

To determine the deviation between highlygeneric, widely-applicable rooftop PV assessmentsand computationally intensive, highly resolvedtechniques, we compare the highest resolution leveltechniques in figure 1 with the IEA’s methodologydefined in the previous section. The IEA EnergyTechnology Perspectives report (2016) aggregates themost statistical data of any of the methods, andtherefore, has been chosen as the baseline low-levelapproach to determine rooftop PV potential. Figure 3(a) shows the annual PV electricity potential for 3

4

select cities or fractions of them: Gangnam district ofSeoul, Korea (Hong et al 2016), Bardejov, Slovakia(Hofierka and Kanuk 2009), and San Francisco, USA(Mapdwell 2017). Figure 3(b) shows the percentdifference, as defined in equation (4), for highly-resolved techniques that were applied to multiplecities. Ko et al (2015) considers seven cities in Taiwan;Mapdwell considers ten cities, eight in the USA andtwo in Chile (Mapdwell 2017); Bergamasco andAsinari (2011b) consider 134 municipalities in Turin,Italy; Google Project Sunroof considers over 40 000urban census tracts in USA (Google 2017). As can beseen in figure 3, there are large discrepancies in theestimated rooftop PV potential when comparinghigh-level and low-level approaches, irrespective ofthe number of cities covered. This suggests thatexisting generic PV rooftop assessments may be tooinaccurate to be widely used for tailored policydesigns.

4. Conclusion

Despite the attractiveness of employing a method thatcould assess rooftop PV potential across thousands ofcities, current approaches tend to vary widely whencompared with more in-depth approaches over thesame geographies. The results presented aim toquantify the variation between different methodolo-gies with varying spatial resolutions across multiplecities. The rooftop PV estimates found usingthe generic IEA method varied significantly fromthe highly spatially resolved techniques with an

Seoul, Korea (Gangnam district) San Francisco, USA

Bardejov, Slovakia Boston, USA

1000

500

0

4000

2000

0

20

1500

1000

500

0

10

0

Hong 2016

Hofierka 2016

IEA 2016

IEA 2016

IEA 2016Mapdwell

IEA 2016Mapdwell

Δ 115.7%

Δ 16.2%Δ 84.2%

Δ 207.1%

GW

h/ye

ar

(a) (b)

[%]

300

200

100

-100

Ko 2015Bergamasco 2011MapdwellGoogle Sunroof

Ko 2015n=7

Bergamasco 2011n=134

Mapdwelln=10

Google Sunroofn>150

Roo

ftop

PV

ele

ctric

ity g

ener

atio

n po

tent

ial:

Per

cent

err

or d

evia

tion

vs. I

EA

2016

Figure 3. (a) Rooftop PV electricity generation potential (GWh yr−1) comparison between selected high spatial resolutionmethodologies that focus on individual cities (Hong et al (2016), Hofierka and Kanuk 2009, and Mapdwell (2017): San Francisco,USA and Boston, USA) and the low spatial resolution methodology from IEA (2016). (b) Percent error deviation between fourdifferent high resolution and low resolutionmethodologies involving multiple reported cities: Ko et al (2015), Bergamasco and Asinari(2011b), Mapdwell (2017), and Google Project Sunroof vs. IEA (2016). Box plots show the 25th, 50th, and 75th percentiles of the data,with cap ends showing the 5th and 95th percentiles to exclude outliers. Mean values are represented as a square, and a dashed line isshown at the 0% error deviation compared to the IEA 2016 methodology. Plotted values are in units of percent, ranging between�78% and 315%.

Environ. Res. Lett. 12 (2017) 125005

average absolute percent difference of 110%. It wasdifficult to compare the highly spatially resolvedtechniques against each other, as they considereddifferent geographic areas. Furthermore, lack ofvalidation across models or against existing rooftopinstallation performances tend to increase the uncer-tainty in the assessments of rooftop PV potential.

Policy makers are often faced with a difficultdecision. They either need to rely on generic rooftopPV assessments with potentially low accuracy or toinvest in high resolution research in their geographicarea of interest. For many decision makers riskaversion would prevent the use of the former andresources may not be available for the latter, whichtend to require expensive data collection with difficultcalibration and large computational resources. Furtherresearch is needed to (i) validate the high resolution,geostatistical approaches, (ii) apply high resolutiontechniques to more cities, and (iii) extract informationfrom the high resolution models to build a moreaccurate and robust generic rooftop PV assessmenttool applicable to most, if not all, cities.

The authors acknowledge that the body ofliterature herein presented might have left room formore publications to be included in figure 2, andcompared in figures 3(a) and (b). Nevertheless, ouraim is to provide a framework to categorize methodsto assess rooftop PV potential at the city level andelucidate variations observed amongst high- and low-level approaches.

Acknowledgments

We thank Joel Conkling for providing valuable dataand useful discussions. This researchwas supported by

5

the Berkeley Energy and Climate Institute—InstitutoTecnológico de Estudios Superiores de Monterrey(BECI–ITESM) Energy Fellowship (to SC) and by theEnergy Efficiency and Renewable Energy PostdoctoralResearch Award from the US Department of Energy(to DAS).

ORCID iDs

Sergio Castellanos https://orcid.org/0000-0003-3935-6701

References

Arnette A N 2013 Integrating rooftop solar into a multi-sourceenergy planning optimization model Appl. Energy 111456–67

Bergamasco L and Asinari P 2011a Scalable methodology for thephotovoltaic solar energy potential assessment based onavailable roof surface area: application to Piedmont Region(Italy) Sol. Energy 85 1041–55

Bergamasco L and Asinari P 2011b Scalable methodology for thephotovoltaic solar energy potential assessment based onavailable roof surface area: further improvements by ortho-image analysis and application to Turin (Italy) Sol. Energy85 2741–56

Cole I, Palmer D, Gottschalg R and Betts T 2016 Assessment ofpotential for photovoltaic roof installations by extraction ofroof tilt from light detection and ranging data andaggregation to census geography IET Renew. Power Gener.10 467–73

Demographia 2016 Demographia World Urban AreasDepartment of Economic and Social Affairs, Population Division

UN 2014 World Urbanization Prospects: the 2014Revision, CD-ROM Edition

Gagnon P, Margolis R, Melius J, Phillips C and Elmore R 2016Rooftop solar photovoltaic technical potential in theUnited States: a detailed assessment NREL Tech. Rep.NREL/TP-6A20-65298 (www.nrel.gov/docs/fy16osti/65298.pdf)

Environ. Res. Lett. 12 (2017) 125005

Google 2017 Project Sunroof Data Explorer (www.google.com/get/sunroof/data-explorer/) (Accessed: 13 June 2017)

Hofierka J and Kanuk J 2009 Assessment of photovoltaicpotential in urban areas using open-source solar radiationtools Renew. Energy 34 2206–14

Hong T, Lee M, Koo C, Jeong K and Kim J 2016 Developmentof a method for estimating the rooftop solar photovoltaic(PV) potential by analyzing the available rooftop areausing Hillshade analysis Appl. Energy 194 320–32

International Energy Agency (IEA) 2016 Energy technologyperspectives 2016: towards sustainable urban energy systemsAnnex H (www.iea.org/etp/etp2016/annexes/)

International Energy Agency (IEA) 2015 Renewable EnergyMedium-Term Market Report

International Energy Agency (IEA) 2016 Trends in PhotovoltaicApplications—Survey Report of Selected IEA Countriesbetween 1992 and 2015

International Energy Agency (IEA) 2000 Trends in PhotovoltaicApplications in Selected IEA Countries between 1992 and1999

Italian National Institute of Statistics 2011 I. Stat Database:Population and Housing Census

Izquierdo S, Rodrigues M and Fueyo N 2008 A method forestimating the geographical distribution of the availableroof surface area for large-scale photovoltaic energy-potential evaluations Sol. Energy 82 929–39

Jakubiec J A and Reinhart C F 2013 A method for predictingcity-wide electricity gains from photovoltaic panels basedon LiDAR and GIS data combined with hourly Daysimsimulations Sol. Energy 93 127–43

Kammen D M and Sunter D A 2016 City-integrated renewableenergy for urban sustainability Science 352 922–8

Karteris M, Slini T and Papadopoulos A M 2013 Urban solarenergy potential in Greece: a statistical calculation modelof suitable built roof areas for photovoltaics Energy Build.62 459–68

Ko L, Wang J-C, Chen C-Y and Tsai H-Y 2015 Evaluation of thedevelopment potential of rooftop solar photovoltaic inTaiwan Renew. Energy 76 582–95

Korean Statistical Information Service 2010 Population andHousehold Census

Kurdgelashvili L, Li J, Shih C-H and Attia B 2016 Estimatingtechnical potential for rooftop photovoltaics in California,Arizona and New Jersey Renew. Energy 95 286–302

Kurtz S, Haegel N, Sinton R and Margolis R 2017 A new era forsolar Nat. Photon. 11 3–5

Lehmann H and Peter S 2003 Assessment of roof & façadepotentials for solar use in Europe Report (Aachen: Institutefor sustainable Solutions and Innovations) (http://isusi.de/downloads/roofs.pdf)

6

Lighting Global and Bloomberg New Energy Finance 2016 Off-grid solar market trends 2016 Report (www.lightingglobal.org/wp-content/uploads/2016/03/20160301_OffGridSolarTrendsReport.pdf)

Lopez A, Roberts B, Heimiller D, Blair N and Porro G 2012 USRenewable Energy Technical Potentials: A GIS-BasedAnalysis (Golden, CO: National Renewable EnergyLaboratory (NREL))

Mainzer K, Fath K, McKenna R, Stengel J, Fichtner W andSchultmann F 2014 A high-resolution determination of thetechnical potential for residential-roof-mountedphotovoltaic systems in Germany Sol. Energy 105 715–31

Mapdwell 2017 Mapdwell Solar (www.mapdwell.com) (Accessed:13 June 2017)

Margolis R and Zuboy J 2006 Nontechnical Barriers to SolarEnergy Use: Review of Recent Literature (Golden, CO:National Renewable Energy Laboratory (NREL)) (www.osti.gov/scitech/biblio/893639)

Najem S 2017 Solar potential scaling and the urban roadnetwork topology Phys. Rev. E 95 012323

NASA 2014 Surface Meteorology and Solar Energy Dataset 6.0Nguyen H T and Pearce J M 2013 Automated quantification of

solar photovoltaic potential in cities Int. Rev. Spat. Plan.Sustain. Dev. 1 49–60

Ordóñez J, Jadraque E, Alegre J and Martínez G 2010 Analysis ofthe photovoltaic solar energy capacity of residentialrooftops in Andalusia (Spain) Renew. Sustain. Energy Rev.14 2122–30

Schallenberg-Rodríguez J 2013 Photovoltaic techno-economicalpotential on roofs in regions and islands: the case of theCanary Islands. Methodological review and methodologyproposal Renew. Sustain. Energy Rev. 20 219–39

Schwartz L et al 2017 Electricity end uses, energy efficiency, anddistributed energy resources baseline Lawrence BerkeleyNational Laboratory Report LBNL-1006983 (https://emp.lbl.gov/publications/electricity-end-uses-energy)

Singh R and Banerjee R 2015 Estimation of rooftop solarphotovoltaic potential of a city Sol. Energy 115 589–602

Strupeit L and Palm A 2016 Overcoming barriers to renewableenergy diffusion: business models for customer-sited solarphotovoltaics in Japan, Germany and the United StatesJ. Clean. Prod. 123 124–36

Takebayashi H, Ishii E, Moriyama M, Sakaki A, Nakajima S andUeda H 2015 Study to examine the potential for solarenergy utilization based on the relationship between urbanmorphology and solar radiation gain on building rooftopsand wall surfaces Sol. Energy 119 362–9

Wiginton L K, Nguyen H T and Pearce J M 2010 Quantifyingrooftop solar photovoltaic potential for regional renewableenergy policy Comput. Environ. Urban Syst. 34 345–57