Dynamic building stock modelling

(1)

Delft University of Technology

Dynamic building stock modelling

Application to 11 European countries to support the energy efficiency and retrofit

ambitions of the EU

Holck Sandberg, N.; Sartori, I.; Heidrich, O.; Dawson, R.; Dascalaki, E.; Dimitriou, S.; Vimm-r, T.; Filippidou,

Faidra; Stegnar, G.; Sijanec Zavrl, M.

DOI

10.1016/j.enbuild.2016.05.100

Publication date

2016

Document Version

Final published version

Published in

Energy and Buildings

Citation (APA)

Holck Sandberg, N., Sartori, I., Heidrich, O., Dawson, R., Dascalaki, E., Dimitriou, S., Vimm-r, T., Filippidou,

F., Stegnar, G., Sijanec Zavrl, M., & Brattebø, H. (2016). Dynamic building stock modelling: Application to

11 European countries to support the energy efficiency and retrofit ambitions of the EU. Energy and

Buildings, 132, 26-38. https://doi.org/10.1016/j.enbuild.2016.05.100

Important note

To cite this publication, please use the final published version (if applicable).

Please check the document version above.

Copyright

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons. Takedown policy

Please contact us and provide details if you believe this document breaches copyrights. We will remove access to the work immediately and investigate your claim.

This work is downloaded from Delft University of Technology.

(2)

EnergyandBuildings132(2016)26–38

ContentslistsavailableatScienceDirect

Energy

and

Buildings

jo u r n al h om ep age :w w w . e l s e v i e r . c o m / l o c a t e / e n b u i l d

Dynamic

building

stock

modelling:

Application

to

11 European

countries

to

support

the

energy

efﬁciency

and

retroﬁt

ambitions

of

the

EU

Nina

Holck

Sandberg

a,∗

_,

_Igor

_Sartori

b

_,

_Oliver

_Heidrich

c

_,

_Richard

_Dawson

c

_,

Elena

Dascalaki

d

_,

_Stella

_Dimitriou

e

_,

_Tomáˇs

_Vimmr

f

_,

_Faidra

_Filippidou

g

_,

_Gaˇsper

_Stegnar

h

_,

Marjana ˇSijanec

Zavrl

h

_,

_Helge

_Brattebø

a

a_Industrial_Ecology_Program_and_Department_of_Energy_and_Process_Engineering,_Norwegian_University_of_Science_and_Technology_(NTNU),₇₄₉₁

Trondheim,Norway

b_SINTEF,_Department_of_Building_and_{Infrastructure,}_P.O._Box₁₂₄_Blindern,₀₃₁₄_Oslo,_Norway

c_School_of_Civil_Engineering_and_Geosciences_Cassie_Building,_Newcastle_University_Newcastle_upon_Tyne,_NE1_7RU,_UK

d_Group_Energy_Conservation_(GREC),_Institute_for_{Environmental}_Research_&_Sustainable_Development_(IERSD),_National_Observatory_of_Athens_(NOA),_I.

Metaxa&Vas.Pavlou,GR-15236PaleaPenteli,Greece

e_Cyprus_University_of_Technology,₃₀_Archbishop_Kyprianou_Str.,₃₀₃₆_Lemesos,_Cyprus

f_STÚ-K,_a.s.,_Saveljevova_18,₁₄₇₀₀_Praha,_Czech_Republic

g_Delft_University_of_Technology,_Faculty_of_Architecture_and_the_Built_Environment,_OTB_–_Research_for_the_Built_Environment,_P.O._Box_5043,₂₆₀₀_GA_Delft,

TheNetherlands

h_Building_and_Civil_Engineering_Institute_ZRMK,_Dimiˇceva_12,₁₀₀₀_Ljubljana,_Slovenia

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received1November2015

Receivedinrevisedform15April2016

Accepted31May2016

Availableonline6June2016

Keywords: Dynamicmodelling Comparativeanalysis Dwellingstock Housing Renovation Energyefﬁciency Europe

a

b

s

t

r

a

c

t

Adynamicbuildingstockmodelisappliedtosimulatethedevelopmentofdwellingstocksin11 Euro-peancountries,overhalfofallEuropeandwellings,between1900and2050.Themodelusestimeseries ofpopulationandnumberofpersonsperdwelling,aswellasdemolitionandrenovationprobability functionsthathavebeenderivedforeachcountry.Themodelperformswellatsimulatingthelong-term changesindwellingstockcompositionandexpectedannualrenovationactivities.Despitedifferencesin datacollectionandreporting,themodelledfuturetrendsforconstruction,demolitionandrenovation activitiesleadtosimilarpatternsemerginginallcountries.Themodelestimatesfuturerenovation activ-ityduetothestock’sneedformaintenanceasaresultofageing.Thesimulationsshowonlyminorfuture increasesintherenovationratesacrossall11countriestobetween0.6–1.6%,fallingshortofthe2.5–3.0% renovationratesthatareassumedinmanydecarbonisationscenarios.Despitethis,78%ofalldwellings couldbenefitfromenergyefficiencymeasuresby2050,eitherastheyareconstructed(31%)orundergo deeprenovation(47%).However,asnomorethanonedeeprenovationcycleislikelyonthistimeframe, itiscrucialtoinstallthemostenergyefficientmeasuresavailableattheseopportunities.

1. Introduction

Deliveringenergyefﬁciencyimprovementsinthebuildingstock iscentraltopublishedcityandnationalplanstoachievecarbon

∗ Correspondingauthor.

E-mailaddresses:nina.h.sandberg@ntnu.no

(N.H.Sandberg),igor.sartori@sintef.no(I.Sartori),oliver.heidrich@newcastle.ac.uk

(O.Heidrich),richard.dawson@newcastle.ac.uk(R.Dawson),

edask@noa.gr(E.Dascalaki),demetrioustella@yahoo.com

(S.Dimitriou),t.vimmr@stu-k.cz(T.Vimmr),F.Filippidou@tudelft.nl(F.Filippidou),

gasper.stegnar@gi-zrmk.si(G.Stegnar),marjana.sijanec@gi-zrmk.si

(M. ˇSijanecZavrl),helge.brattebo@ntnu.no(H.Brattebø).

reductiontargetsacrossEurope[1,2].Accordingtoarecent EU-JRCreport[3],energyrenovationisinstrumentalforreachingthe EU2020goalsi.e.reduceGHGby20%,have20%ofenergyfrom renewablesandincreaseinenergyefﬁciencyby20%.Thiscallsfor acommonEUrenovationplanwitharegionalapproach prioritiz-inglessdevelopedregions.IntheEU,feasibilitystudies,national roadmapsandactionplansforenergysavingsinbuildingstocks commonlyassumeasigniﬁcantincreaseintherenovationratesin ordertoobtainfutureenergysavings,butthelikelinessofreaching theseincreasedratesisrarelyevaluatedordiscussed[4–9].

Understandingandinﬂuencingtheexistingandfuturedwelling stockisofvitalimportanceastherearesigniﬁcantlock-inrisks associatedwiththelonglifespansofbuildingsandinfrastructures http://dx.doi.org/10.1016/j.enbuild.2016.05.100

(3)

[10,11].Consequently,themajorityoftoday’sdwellingswillstill existin2050andbeyond[12].Ifstringentregulationsarenot intro-duceduniversallyandhigh-standardenergyretroﬁtsareassured whenbuildingsarerenovated,energyuseandcorrespondingGHG emissionscouldbe‘locked-in’formanydecadestocome.This lock-inisestimatedtoleadtoa33%increaseinglobalenergyusefor buildingsby2050insteadofadecreaseof46%ifchangesaremade [13].

Dwellingstockswereconstructedovervariousperiods(cohorts) andsegmentsofthestocktobeprioritizedforrenovationshould beidentiﬁed [14].Housingstocksareexposedtorefurbishment activitiesduringtheageingprocess,andrenovationinthecoming decadestoalargeextentdependsontheagecompositionofthe stockandthepreviousrenovationactivity.

ThroughoutEurope,nationalapproachesforthemonitoringof thebuildingstockhaveevolvedseparately[15].Informationabout theprogressoftheenergyperformancerenovationisrequiredto tracktheprogressofpolicyimplementation. Betterinformation anddataareneededtohelpdeveloproadmapsinordertoachieve moreenergyefﬁcientbuildings[4].Toaddresstheshortcomings andchallengesidentiﬁedthereisaneedforanewmethodology thatcanbeusedforconsistentandscalableanalysisofbuilding stockacrossmultiplecountries.

Energyanalysesofdwellingstocksaredeﬁnedbyastockmodel andanenergymodel.Thestockmodeldescribesthedevelopment ofthestockinterms ofsize,composition andrenovationstate, whereastheenergymodelincludesaverageenergyintensitiesof thevarioussegmentsofthestock,andassumedsavingsobtained whendwellingsarerenovated.Standardlineardwellingstock mod-elscommonlyassumeﬁxedratesforconstruction,demolitionand renovation activities[16–18] whereas inreality theserates are dynamic,bothintheshortandlongterm,anddependupon exter-naldriversaswellasthetypeandagecompositionofthebuilding stock.Thenatureofhousingsupplyandtheimpactsofdemands andhousingsupplyiselastic,butanincreaseindemandinthelong termisexpectedasaresultwhenpopulationincreases[19].

Intheliterature,therearevariousmodelsandtoolstoassess energyconsumptionindwellingstocks.Kavgicetal.[20] differen-tiatedbetweentop-downandbottom-upapproachinstock-level energyconsumptionmodelling.Theyhighlightedtheimportance oftransparencyandquantificationofinherentuncertaintieswithin anystockmodel.Arangeofbottom-upmodelsareusedfor mate-rial,energy or carbonanalysesof dwelling stocks,e.g.[21–25]. Meijieretal.[24]alsoidentifiedseriousgapsinthemonitoring ofthephysicalresidentialstock,notingthatnoneofthecountries monitoredtherenovationeffectsonthehousingstock.In mate-rial,energyor carbonanalysesof buildingstocksthere isoften alackofdataonthemodels’inputsandoutputs,aswellasthe algorithmsusedthat makethereproductionoftheresults diffi-cult[20,24].Developingscenariosoffuturedwellingstockenergy demandscanunearthsuchdiscrepancies,uncertainties,andareas ofimprovementsaswellashighlightingtheneedofmorerobust datacollection[26].Thereisaneedtoquantifyandanalyzethe robustnessofkeydatafromretrofittingratestototalstockandits associatedassumptionsinordertounderstandtheinfluencesofthe long-termtransformationofthedwellingstock[27].

Thereisalackofdwellingstockmodelsthatdescribethe devel-opmentofthestocksin agood way,and itbecomes clearthat thereisanurgentneedtogetamoredetailedunderstandingof thelong-termdynamicofthedwellingstockstobeableto evalu-atethefutureenergyreductionpotential.Thiswillleadtoadeeper understandingofthedynamicsthatdrivetheactivitiesinthe sys-temand shouldbeapreconditionforamoreconsistentwayto addressevolutionsoftheexistingandfuturebuildingstockandits energydemand.Thisshouldalsosupportthepreviousrequestfor thedeﬁnitiononfuturepracticeforthedwellingstock[28].

Adynamicdwellingstockmodelhasbeendevelopedthrougha rangeof publicationsandisusedtostudythelong-term devel-opment in dwelling stock size and composition with various applications[29–41].Thecoreofthemodelisthepopulation’sneed toresideandthemaininputparametersarethedriversinthe sys-tem,thepopulationandthenumberofpersonsperdwelling.The construction,demolitionandrenovationactivityinthesystemare outputsfromthemodel,aimingatdescribingthedynamicsofthe stockresultingfromthechangingdemandandageingofthestock. Aseparatepaperexploredthesensitivityinmodelresultsand conclusionstochangesininputparameters.ForthecaseofNorway, theyconcludedthatthemostsensitiveinputparameters popu-lationandlifetimeofdwellingsarealsotheinputparametersof highest uncertainty.However, evenwhen changingthese input parameterstoextremeandunrealisticvalues,themainconclusions regardingfuturerenovationratesremainedunchanged.Themodel resultsandconclusionsforthecasestudyofNorwaywererobust tochangesintheinputparameters.Renovationratesatlevels nec-essarytoachieve polity targetsinenergy and emissionsavings seemedunrealistictobeachievedwhenmodellingthe“natural” needforrenovation[38].

Thedynamicdwellingstockmodelisgeneral,thoughthe ver-sionfocussingonrenovationsofarhasbeenappliedonlytoNorway [31,37,40].Inthepresentpaper,weapplythesamemethodto ana-lyzetheconstruction,demolitionandrenovationsactivitiesinthe dwellingstocksof11Europeancountriesusingconsistent deﬁni-tionsanddataforallcountries,toevaluatehowthemodelﬁtsto othercountriesandifgeneralconclusionscanbemadeacrossa rangeofEuropeancountries.Themodelandalgorithmpresented inSartorietal.[40]areappliedtoCyprus,CzechRepublic,France, Germany,GreatBritain,Greece,Hungary,theNetherlands,Norway, SerbiaandSlovenia.Foreachcountrythehousingstockismodelled withrespecttoitspastevolution andwithprojections towards 2050.

Thekeyresearchquestionstobeaddressedare:

• How well doesthe model represent the long-term historical developmentinthedwellingstocksofthegivenEuropean coun-tries?

• Whatarethedifferencesbetweenthecountriesindata availabil-ity,feasibilityofthemodel,qualityoftheresultsandnational conclusionstobedrawnfromtheresults?

• Whatgeneraltrendsareobservedandwhatgeneralconclusions canbemadefromthecomparativeanalysisoftheresultsfrom thedifferentcountries?

• Whatisthepotentialforfuturedevelopmentinenergy-related renovationofthedwellingstocksinthegivenEuropeancountries andtowhatextentaretheseﬁndingsinlinewiththe recommen-dationsfromSahebetal.[3]ortheassumedrenovationratesin traditionalscenariomodels,nationalroadmapsandactionplans [4,6]?

2. Methods

2.1. Overviewofthedynamicdwellingstockmodel

Afulldescriptionofthemodel,includingtheunderlying equa-tionsandjustiﬁcation,isprovidedbySartorietal.[40].Thekey principlesandmodelstepsarenowsummarizedtoenable under-standingoftheanalysispresentedinthispaper,whilethemodel equationsarepresentedinAppendixA.

The dynamic dwelling stock model describes thelong-term developmentofthesizeandagecompositionofthedwellingstock inacountryorregion.Theconceptualframeworkofthemodelis presentedinFig.1.Thecoremodeldriveristhepopulation’sneed

(4)

28 N.H.Sandbergetal./EnergyandBuildings132(2016)26–38

Fig.1.Conceptualframeworkofthemodel.

toreside.AnnualdemandfordwellingsS_Disestimatedasthe popu-lationPdividedbytheaveragenumberofpersonsperdwellingPD.

ThetotalnumberofdemolisheddwellingsD_demisestimatedeach yearbyapplyingademolitionprobabilityfunctiononthe construc-tionfromallpreviousyears.Adeﬁnedshareoftheconstruction fromeachyearisassumednevertobedemolishedtopreservethe nationalbuildingheritage.Massbalanceprinciplesareusedto esti-matetheconstructionactivityDnewinyeari,thatisequaltothesum

ofnewdwellingsthatareneededinordertoreplacethedemolished onesandmeetthechangeindemandfromyeari–1toyeari.

Finally,renovationactivityDreninyeariisestimatedbyapplying

arenovationprobabilityfunctiontotheconstructionfromall previ-ousyears.Themodelallowsforcyclicrepetitionsoftherenovation probabilityfunction,describedbytheaveragetimebetween reno-vationsofacertaindwelling,R_C.Thecyclicrenovationprobability functionislinkedtothelifetimeprobabilityfunction,preventinga dwellingtobedemolishedshortlyafterbeingrenovated.The def-initionoftherenovationactivityiscase-speciﬁcandtherelated renovationcycleshoulddescribetheaveragetimebetween reno-vationsofthedeﬁnedtype.

ThemodelresultsaretheyearlydemandfordwellingsS_D, con-structionactivityDnew, demolitionactivityDdem and renovation

activityD_ren,asshowninFig.1.Modelresultsaresegmentedin cohorts(constructionperiodsc)tovisualizethestockcomposition ofdifferentcohortsaswellastounderstandtheextenttowhich differentcohortsareexposedtodemolitionorrenovationactivity.

2.2. Inputdataandassumptions

Theinputdatausedinthisstudyaresummarizedinthissection. Detailedinformationaboutthedatasources,assumptionsanddata processingforeachcountryispresentedinAppendixB.

Insomeofthecountriesincludedinthestudy,thegeographical areabelongingtothecountryhaschangedovertime.Whenthisis thecase,inputdatareferringtothecurrentterritoryofthe coun-tryarecollectedorestimatedbyrescalingdatafromothertime periodsorfromageographicalareanotcompletelycorresponding tothecurrentterritoryofthecountry.Historicaldataonnumber ofpersonsperdwellingissometimesonlyavailableforpartsof theterritoryoralargerarea,andthisisthenassumedtobe rep-resentativeforthegeographicalareacurrentlybelongingtothe

country.Whenrelevant,thisisexplainedindetailforeachcountry inAppendixB.

Short-termvariationsintheinputtimeseriespopulationPand personsperdwellingP_Dresultinﬂuctuationsinthemodelresults. Thiscangivetheimpressionthatthemodelcapturesshort-term processes.Toavoidthisnoiseintheresults,non-linearregressionis appliedtotherawdatatomakesmoothinputcurvesforpopulation PandpersonsperdwellingP_D.

2.2.1. Population

Populationstatistics,projectionsandsomeadditional assump-tionsareusedtocreatethesmoothinputcurvesforthepopulation. Thedetailsofthedataavailabilityanddataprocessingfor each countryarepresentedinAppendixB.Theavailabilityof popula-tiondataisgoodinallcountries,mainlysourcedfromcensusdata andpopulationprojectionsfromnationalstatisticsoffices.Insome countries,however,liketheCzechRepublicandSerbia,itisdifficult tofitasmoothcurvetotherawdataduetoperiodsofrapidchanges inthepopulation.IntheCzechRepublicthestrongincreaseinthe populationintheearly1900sandthesubsequentdecreaseduring andafterWorldWarIIarenotwellreflectedinthesmoothedcurve. InthecaseofSerbiathesmoothedcurvediffersnotablyfromthe rawdataintherecentpastandthefuture.

ThesmoothpopulationcurvesarepresentedinFig.2.The non-linearregressionfunctionwiththebestﬁt waschosenfor each country.Incountrieswithasteadilyincreasingpopulation,a Sig-moidalregression is used.AGaussian regression is usedinthe countrieswherethepopulationhasincreasedfollowedbyaperiod of(expectedfuture)decrease.TheR2_value_of_the_non-linear

regres-sionwaslargerthan0.95forallcountries.Modelprojectionsare mostsensitivetotherecentpastratesofchange.Forallcountries exceptCyprus,CzechRepublicandSerbiathesmoothregression curveshowsaverygoodﬁtwiththerawdataforthiscriticalperiod. Forthesethreeexceptions,thepoorerﬁtisconsideredwhen ana-lyzingtheresults.

2.2.2. Personsperdwelling

Historicalpersonsperdwellingdataareavailablefromcensuses sinceabout1900in mostofthecountries, inGreat Britainand Norwaysince1800andinFrance,GermanyandSerbiasinceabout 1950.FuturedevelopmentinPDisbasedonassumptionsand

(5)

Fig.2.Developmentinpopulation(P)inallcountries,countriesaregroupedintothreegraphsaccordingtopopulationsize.

available,additionalassumptionsareincludedintheregressionto obtainasmoothinputcurveforthewholeperiod1800–2050.This isexplainedindetailforeachcountryinAppendixB.

The smooth PD curves have the same shape for all

coun-tries,althoughwithdifferentstartingvalues.Sigmoidalnon-linear regressionisused.Theshapeofthetypicalcurveisexempliﬁedfor thecaseofGreatBritaininFig.3togetherwiththerawdataused tomakethesmoothcurve.Further,theextremecasesofSerbiaand CyprusarealsopresentedinFig.3.Serbiahasahighstartingvalue of6.7personsperdwellingin1800.Cyprusisthecountrywiththe loweststartingvalueof4.1andtheatypicaldevelopmentwitha constantlevelofpersonsperdwellingsfrom1800to1965followed byarapiddecreasetoalevelof1.9in2011.Allothercountrieshave asmoothcurvewithashapemoresimilartothecurvespresented forGreatBritainandSerbia,startingatavalueof4.6–5.5and end-ingatapproximately2personsperdwellingin2050,asshownin AppendixB.

2.2.3. Dwellinglifetimeandrenovationparameters

The lifetime probability function is assumed to follow a Weibulldistributiondeﬁnedbytheparametersaveragelifetimeper dwellingandtheinitialperiodafterconstructionwherethe prob-abilityofdemolitioniszero.ThisisexplainedindetailinSandberg etal.[37]andSartorietal.[40],andisconsistentwith recommen-dationsbySereda[42].Thecountry-speciﬁcprobabilityfunction parametersaredescribedinAppendixB.

Thedeﬁnitionoftherenovationactivityinthemodelis case-speciﬁc.An“energysavingmodernization”suchaschangingthe heatingsystemorinstallingaphotovoltaiccanbethoughtashaving a20-yearcycleorevenshorter.Inthisstudythough,weexplorethe dynamicsofdeeperrenovationsthathavethepotentialfor

includ-ingenergy-efﬁciencymeasuresthatleadtomuchlargerreductions inenergydemand.Thesemeasuresarecostlyandunlikelytobe implementedifadwellingisnotgoingthroughamajorrenovation inanycase,perhapsduetoits“natural”ageingprocess,achangeof ownership,ortheneedformaintenanceandupgrading.We esti-matethetotalrenovationactivityresultingfromtheageingprocess ofthedwellingstockineachcountryinvolved.Formostcountries, thisrelatestodeeprenovationoffacades,commonlyestimatedto occurincyclesof40or50years.OnlyinthecaseofGreeceare sin-glemeasureswitharenovationcycleof30yearsareassumedto havealargercontributiontotheenergysavingsthandeep renova-tionoffacades.Therenovationprobabilityfunctionisassumedto followaNormaldistributionforallcountries,asexplainedindetail inSandbergetal.[37]andSartorietal.[40].

Thelifetimeandrenovationparametervaluesforallcountries are listed in Table 1. The column “Construction period”refers towhichsegmentsofthestocktheassumptionsareappliedto. In principle,theparametervalues candifferbetweendwellings constructedin differentyears. However,due tolimited empiri-caldata,thesamevaluesareassumedforalldwellingsregardless ofconstructionyearforallcountriesexceptHungary.Inthecase ofHungary,theinitialperiod withoutdemolitionisassumed to decreaseforfutureconstruction.Furtherdescriptions,references andexplanationsoftheparametervalueschosenforeachcountry aregiveninAppendixB.

2.2.4. Cohortdeﬁnition

Someofthemodelresultswillbesegmentedintocohorts.For easiercomparisonoftheresults,thecohortsaredeﬁnedequally forallcountries,aslistedinTable2.Cohort0representsthe ini-tialstockatthestart ofthemodellinginyear1800,Cohort1is

(6)

Fig.3.Evolutioninthenumberofpersonsperdwelling(PD)inGreatBritain,CyprusandSerbia.

Table1

Lifetimeandrenovationparametervaluesforallcountries.

Country Construction period

AverageLifetime (years)

Initialperiodwithout demolition(years) Sharenever demolished Renovationcycle (years) Cyprus All 125 40 5% 40

CzechRepublic All 120 60 5% 40

France All 125 40 5% 40

Germany All 125 40 5% 40

GreatBritain All 175 40 10% 40

Greece All 70 40 5% 30

Hungary –2015 125 50 5% 40

2016– 125 40 5% 40

TheNetherlands All 120 40 3% 38

Norway All 125 40 5% 40

Serbia All 100 50 5% 50

Slovenia All 120 40 8% 40

Table2 Cohortdeﬁnition.

Cohortnumber Startyear Endyear

0 – 1800

1 1801 1945

2 1946 1980

3 1981 2015

4 2016 2050

theconstructionfrom1801totheendofWorldWarII.Although therearelargedifferencesbetweenthedwellingsconstructedin theearly1800sandinthe1930s,itisassumedthatfuture renova-tiontechnologies,includingthoseforenergy-savingpurposes,will besimilar.Theshareofthecurrentstockconstructedbefore1945 isalsolimited(lessthan25%inmostofthecountriesconsidered here).Cohort2,3and4representperiodsof35yearswhereCohort 2isthepost-warconstructionfrom1946to1980,Cohort3isthe mostrecentconstructionfrom1980to2015andCohort4isfuture constructionfrom2016to2050.

3. Resultsanddiscussion

Country-speciﬁcdetailedresultsaredescribedinAppendixB.In thefollowing,theresultsarepresentedforallcountries,ormaking referencetooneorasubsetofcountries.

3.1. Dwellingstocksizeandcomposition

The observed historical developmentof dwelling stock size, measuredintermsofthenumberofdwellings,isusedasaninputto themodelthroughtheparameterpersonsperdwellingPD.

Never-theless,themodelresultsarecomparedwiththestatisticstoensure thatthesmootheningoftheinputcurveshasnotresultedin signif-icantdifferencesbetweenthemodelresultsandthestatistics.For all11countriesinthisstudy,thereisagoodﬁt,asshowninTable3 andtherelateddiscussion.

InFig.4themodelleddwellingstocksize,andcompositionof cohorts,ispresentedforallcountriesfortheyears1980,2015and 2050.Thecurrentstocksizevariessigniﬁcantly,from0.4million dwellingsinCyprusto39milliondwellingsinGermany. There-fore,allresultsarenormalizedagainstthesizeofthestockinthe respectivecountryin2015.Thetotalnumberofmillioninhabited dwellingsineachcountryin2015isshownnexttothe2015bar.

Fig.4showsthatthesizeofthedwellingstockhasincreasedin allcountriesfrom1980to2015,althoughatdifferentrates.Inall countriesexceptinCyprus,thenumberofdwellingsin1980was 62-84%ofthecurrentstocksize.Cyprushasexperiencedalarge recentgrowthinthedwellingstockandthesimulated1980stock wasonly35%ofthecurrentstocksize.Accordingtothe country-speciﬁccomparisonswithstatisticspresentedinAppendixB,there wasa largeincrease in theconstruction activity inthe

(7)

Repub-Table3

Stockcompositioncomparedwithstatistics.

Cohort0–1(–1945) Cohort2(1946–1980) Cohort3(1981–*) Unknowncohort Total

*Yearofcomparison %share %share %share %share %share

Cyprus 2011 Statistical 2.0 23.1 74.8 – 100

Modelled 11.6 19.8 68.6 – 100

CzechRepublic 2011 Statistical 22.0 43.0 32.6 2.3 100

Modelled 35.2 32.7 32.1 – 100

France 1999 Statistical 32.9 45.7 21.3 – 100

Modelled 32.7 38.6 28.7 – 100

Germany 2009 Statistical 24.5 43.5 32.0 – 100

Modelled 31.5 35.5 32.9 – 100

GreatBritain 2013 Statistical 36.9 39.8 23.3 – 100

Modelled 30.0 32.1 37.9 – 100

Greece 2011 Statistical 5.7 49.3 45.0 – 100

Modelled 8.3 34.3 57.4 – 100

Hungary 2011 Statistical 29.2 40.0 30.8 – 100

Modelled 30.6 34.6 34.7 – 100

TheNetherlands 2012 Statistical 19.5 40.9 39.7 – 100

Modelled 18.2 37.2 44.6 – 100 Norway 2011 Statistical 16.6 42.7 36.4 4.4 100 Modelled 22.4 33.5 44.1 – 100 Serbia 2011 Statistical 11.9 52.6 35.5 – 100 Modelled 15.7 40.9 43.4 – 100 Slovenia 2013 Statistical 21.3 45.0 33.7 – 100 Modelled 27.6 31.3 41.1 – 100

licofCyprusaftertheTurkishinvasionin1974.Thisincreaseis delayedbyabout10yearsinthesimulation,andtherealnumber ofdwellingsin1980wasthereforesomewhathigherthaninthe valueshowninFig.4.

Thedwellingstocksareexpectedtokeepgrowinginall coun-triesexceptthosewithdecreasingfuturepopulationprojections: Germany,HungaryandSerbia.Theexpectedincreaseinthestock sizeintheothercountriesis12–25%,exceptinCyprusandNorway wheretheexpectedincreaseisabout35%.

Adecreaseinthelengthofacertaincolorinthebarsshownin Fig.4fromoneobservationyeartothenext,indicatesthesimulated demolitionofdwellingsfromthecorrespondingcohortbetween thetwoobservationyears.Anewcolorshowsthenewconstruction inthegivenperiod,representedasanewcohort.

InTable3,thesimulateddwellingstocksizeandcomposition iscomparedwiththemostrecentofﬁcialnationalstatisticsonthe dwellingstockcompositionforeachcountry.Acommonpattern isobservedinmanyofthecountries:thestockiscomposedofa smallshareofdwellingsconstructedbefore1945andthemodel resultscomparewellwiththesestatistics.Formostcountries,this shareis6–25%ofthestock.ExceptionsareGreatBritain,Hungary andFrancewithsharesofthestockbeingconstructedbefore1945 of30–37%.Fromthelargersharesofthestockconstructedafter 1946,thepost-warconstructionboomincohort2iscommonly notfullyexplainedbythemodel,andtheconstructionfromthe mostrecentdecadesiscommonlysomewhatoverestimated.These discrepanciesareexpectedtoleadtocorrespondingdistortionsin themodelresultsonrenovationactivityintherelevantcohorts,but theresultingtotalrenovationactivityofthestockisnotexpected tobesigniﬁcantlyaffectedbythis.

3.2. Construction,demolitionandrenovationactivitybycountry The simulated construction and demolition activity is com-paredwithstatisticsforeachcountryinAppendixB.Construction statisticsareavailableinmostcountriessinceabout1950–1980. For mostcountries, thelong-termlevel of constructionactivity is broadlycomparable withreported statistics. Acommon pat-ternishoweverobservedinmanycountrieswhereconstruction statisticsareavailable:themodeltendstounderestimatethe post-warconstructionboombetween1950/60–1980/90,andthereafter overestimatestheconstructionactivityinthemostrecentdecades.

Thisisalsoinlinewiththecomparisonofthesimulatedcurrent stock composition with statistics from Table3.The short- and medium-termvariationsinconstructionactivityareexplainedby factorsnotincludedinthemodel,e.g.widerdriverssuchas eco-nomic,climateandunemployment.

Demolitionstatisticsarehardlyavailable,exceptforintheCzech Republicsince1955andinsomefewothercountriessinceabout year2000.Thelong-termdemolitionactivityseemstobeatthe rightlevelfortheCzechRepublic.However,themodelresultsare generallyslightlyhigherthanthereportedvaluesfor1956–1990, whichcanbeexplainedbythelargenumberofdwellingsdestroyed duringWorldWarIIandthereforenotdemolishedatalaterdate, whentheywouldhavereachedtheirendoflife.Theconsequences ofWorldWarIIarenotcapturedinthemodel.

Resultsforexpectedannualrenovationactivity(R_i)inyear2015, 2030and2050arepresentedforallcountriesinFig.5.Theresults arenormalizedagainstthe2015totalforeachcountry.Fig.5shows thedevelopmentintotalrenovationactivityanditsdistribution todwellingsconstructedinthedifferentcohorts.Inallcountries exceptCyprus,themodelforecaststheyearlynumberofdwellings renovatedtoincreaseby4–18%by2030andby4–41%by2050, comparedwiththe2015rate.Futurerenovationactivitywill,to alargerextent,takeplaceindwellingsconstructedafter1980,as thesedwellingsreachanagewithincreasedneedformaintenance. Theneedforrenovationofolderdwellingswilldecreaseduetothe demolitionofdwellingsinthesecohorts.

Cyprus is in a special situationdue to therapid recent and expectedfuturegrowthinthedwellingstocksize.Thiswilllead toanincreaseintherequiredfuturerenovationactivityto main-tainthenewbuildingstock.By2030,theestimatedannualnumber ofdwellingsneedingrenovation(D_ren,2030)isexpectedtobe70% higherthanthecurrentnumberofdwellingsrenovated(Dren,2015).

Theincreaseinrenovationwillmainlytakeplaceindwellings con-structedintheperiod1980–2015.By2050therewillalsobesome needforrenovationofdwellingsconstructedafter2016,andthe totalrenovationactivityisexpectedtobe85%largerthanin2015. 3.3. Interplayofconstruction,demolitionandrenovation

The simulated developmentin annual rates of construction, demolition and renovation for France, Hungary, Great Britain and Germany is presented in Fig. 6. France and Great Britain

(8)

Fig.4. Normalizeddwellingstockcompositioninallcountriesin1980,2015and2050,relativetothe2015totalstocksize.Thenumbersnexttothe2015barsarethenumber

ofmilliondwellingsintotalinthe2015stock.

demonstrateadevelopmentpatternthatistypicalofmostofthe countrieswithincreasingfuturepopulation,whereasHungaryand Germanydemonstratetypicaldevelopmentpatternincountries withexpecteddecreasingpopulation.Allratesaredeﬁnedasthe numberofdwellingsexposedtotheactivitydividedbythetotal numberofdwellingsinthestockinthesameyear.

MostcountriesinthestudyfollowthesamepatternasFrance andGreatBritain:Thesimulatedconstructionratewashighfrom 1900toabout1950–1980.Thereaftertherehasbeenadecreasein theconstructionrate,andthedecreaseisexpectedtocontinuein

thefuture.Thisisduetoslowerpopulationgrowthandsaturationof thenumberofpersonsperdwelling.Further,thesimulatedannual demolitionrateinmostcountrieshasbeenratherstableat0.3–0.7% inthepastandisexpectedtoremainatthesamelevelorincrease slightlytoabout1.0%by2050.

Similarly,inmostcountriesthesimulatedannualrenovation ratehasbeenstableat1–1.5%andisexpectedtoremainatthesame levelsorincreaseslightlytowards2050.Oursimulationssuggest thatrenovationwill,inmostcountries,bethedominantactivityin

(9)

Fig.5.Segmentedresultsforexpectedannualrenovationactivity(Dren,i)forallcountriesin2015,2030and2050,relativetothenumberofdwellingsrenovatedin2015

(Dren,2015).Thenumbersnexttothe2015barsarethetotalnumberofthousanddwellingsrenovatedin2015.

thesysteminthefuture,intermsofnumberofdwellingsexposed totheactivity.

Insomeofthecountries,thesizeofthepopulationhasleveledoff andisexpectedtodecreaseinthefuture.Theconstruction, demoli-tionandrenovationratesfollowthepatternshownforHungaryand GermanyinFig.6,wheretheannualconstructionratehasstrongly decreasedandisexpectedtofalltoabout0.4%by2050.The sim-ulatedannualdemolitionandrenovationratesincreaseinthese countriesinthefuture,asthestocksizeisdecreasing,andthe

con-structionrateisexpectedtobelowerthanboththedemolitionand therenovationratestowards2050.

Thesimulatedconstruction,demolitionandrenovationratesfor allcountriesintheyears2015,2030and2050arelistedinTable4. Minorvariationsovertimeareobservedintheconstructionandthe demolitionrates.Thesimulatedrenovationrates,alsopresented inFig.7,arestablethrough timeinallthecountriesandnever exceeds1.6%inanycountry.Thesedescribethefutureexpected renovationactivityneededformaintenanceoftheexistingstock andmaybeusedtoestimatetheopportunitiestoreadilyintroduce

(10)

Fig.6. Developmentinannualratesofconstruction,demolitionandrenovationinFrance,HungaryandSerbia.

Table4

Construction,demolitionandrenovationratesinallcountriesin2015,2030and2050.Allratesrelatedtothestocksizeinthesameyear.

Constructionrate Demolitionrate Renovationrate

2015 2030 2050 2015 2030 2050 2015 2030 2050 Cyprus 1.6% 1.2% 1.2% 0.2% 0.3% 0.5% 0.9% 1.3% 1.3% CzechRepublic 1.0% 1.0% 0.9% 0.6% 0.6% 0.7% 1.3% 1.3% 1.3% France 1.4% 1.2% 1.0% 0.5% 0.5% 0.6% 1.3% 1.3% 1.4% Germany 0.8% 0.6% 0.4% 0.6% 0.7% 0.8% 1.3% 1.4% 1.5% GreatBritain 1.1% 0.9% 1.0% 0.3% 0.4% 0.4% 1.6% 1.6% 1.6% Greece 2.0% 1.7% 1.4% 0.9% 1.0% 1.2% 1.1% 1.2% 1.2% Hungary 0.9% 0.5% 0.3% 0.5% 0.6% 0.8% 1.4% 1.5% 1.5% TheNetherlands 1.2% 1.0% 0.9% 0.5% 0.6% 0.7% 1.3% 1.4% 1.4% Norway 1.5% 1.4% 1.2% 0.5% 0.5% 0.6% 1.2% 1.2% 1.3% Serbia 1.1% 0.7% 0.5% 0.6% 0.8% 1.0% 0.6% 0.6% 0.6% Slovenia 1.3% 1.2% 1.0% 0.6% 0.6% 0.6% 1.3% 1.3% 1.3%

energy-efficiencymeasureswhendwellingsaregoingthrougha deeprenovation.Thisisaninterestingfindingregardingtheenergy savingpotentialofthedwellingstock.Scenarioanalysesandroad mapsforenergysavingscommonlyassumerapidincreasesofthe renovationratetolevelsof2.5–3%[4–9].Ouranalysisshowsthat therenovationratesresultingfromthedwellingstocks’ownership turnover,orneedformaintenance,willbefarbelowtheselevels inallthecountriesincludedinthisstudy.Althoughtherenovation processcouldprobablybeacceleratedbyappropriateincentiveor investmentschemes[43],toachieveadoublingoftherenovation ratetomeetnationaltargetsfor reductionof energy consump-tionandCO2emissionswillbedifficult.Fundingschemesshould

beusedtoensurethatwhendwellingsarerenovated,high-level energy-efﬁciencymeasuresareintroducedtoavoidlock-ineffects. Short-andmediumtermvariationsintheratesarenotreﬂected inthemodel.InthecaseofGreece,thisisclearinthecaseofthe

cur-rentconstructionrate.DuetotheongoingturbulenceintheGreek economy,theconstructionindustryhascometoahalt,and accord-ingtothelateststatisticsavailable[44,45]theconstructionratein 2014was0.15%.Suchshortandmediumtermvariationscannotbe explainedbythismodel.

Somestudies define the construction, demolition and reno-vationratesasthenumberofdwellingsexposedtotheactivity comparedtothestockcompositioninafixedyear[4,37].Afixed ratethenmeansaconstantnumberofdwellingsexposedtothe activityeachyear.Incountrieswithagrowingstock,futurerates relatedtothestocksizeinafixedyearwillbehigherthanifthe ratesarerelated tothechangingstocksize.In countrieswitha decreasingstock,thecorrespondingratesarelower.Ifrelatingthe presentedrenovationactivitytothe2015dwellingstocksize,the resulting2050renovationratewillbeintherange1.4–2.0%forall countriesexceptSerbiawith0.7%.Soevenwiththisdefinition,the

(11)

Fig.7. Simulatedrenovationrateinallcountries1900–2050.

renovationrateisnotlikelytoreachthelevelof2.5–3%by2050 andinanycasenotinthenearfuture.

3.4. Applicabilityandreliabilityofthemodelanditsresults

Inthisstudy,arangeofinputdataandassumptionsareapplied fortheanalysisofthelong-termdynamicsofthedwellingstocksin 11Europeancountries.Thepresentationofthedataand assump-tionsinSection2.2andinAppendixBrevealedlargedifferences betweenthecountries.Thisismainlyduetothelargedifferences inhistoricaldevelopment,currentsituationandexpectedfuture developmentinthecountries’population,andprobablyalsodue tovariationsinthefactorsincludedinthepopulationprojections andestimationoflifetimeofdwellings.

Dataavailabilityvariesbetweenthecountries.Somecountries donothavedataavailablebefore1900,requiringadditional esti-mationsteps.Thisisparticularlythecaseforcountrieswherethe nationalterritorychangedovertimepriorto1900.Laterchanges inborder, e.g. afterWWIIarebetterdocumented andeasierto address.Theuncertaintycausedbytheestimationofinputdatathis farbackintimehaslittleimpactonthemodelresultsintheperiod ofhighestinterest;therecentpastandthefuturedevelopment.

Thefuturedevelopmentofdwellingstocksisstronglyrelatedto theexpectedchangeinthepopulationineachcountry.The popu-lationprojectionsaretakenfromthestatisticaloffices,butarestill uncertain.Moreover,expectedlifetime,fertilityratesand migra-tioncanbederivedusingdifferentapproachesandassumptionsin thedifferentcountries.However,theseprojectionsprovidethebest availableinformationonfuturepopulationgrowthinthecountries studiedhere.Migrationistreateddifferentlyinthepopulation pro-jectionsforthedifferentcountries,andthismayhavesignificant effectssince thefuture developmentinthepopulationof Euro-peancountrieswillsubstantiallydependonmigrationpolicies.The futuredwellingstocksizeandconstructionrateishighlysensitive tothefuturesizeofthepopulation[38].Further,thedemolition rateand hence the construction neededto replacedemolished dwellingsaresensitivetotheassumedlifetimeofdwellings. Fur-ther,socialfactors,policiesandchanginguserbehaviormightalso influencetheturnoverratesofthedwellingstock.

Thefuturerenovationtowards2050willmainlytakeplacein dwellingsinthecurrentexistingstock.Thefuturerenovationrate isthereforelesssensitivetofuturedevelopmentofthemost uncer-taininputparametersthanthefutureconstructionanddemolition rates.

Themostpolicy-relevantoutcomeofthepresentedanalysisis theresultingrenovationrates.Futureconstructionisexpectedto behighlyenergy-efficient,andthereforethefutureenergysavings inthesystemmainlydependontheimprovementsofthe exist-ingstock.Thefutureenergysavingpotentialintheexistingstock shouldbeidentifiedthroughsimulationoflikelyrenovationrates andmodelsshowingpossibleaverageenergy-intensityreductions whenadwellingisrenovated.Thefuturerenovationratesare lit-tledependentonthehighly uncertaininputs,and theresulting renovationratesforall11countriesturnedouttoberemarkably similarandstableovertime,atlevelsof1–1.6%formostcountries, asshowninFig.7.This,togetherwiththerobustnessofsuchresults asemergedfromthesensitivityanalysisinSandbergetal.[38], indi-catethatrapidincreasesofrenovationratestolevelsof2.5–3%will beverydifficulttoobtain.

3.5. Implicationsforenergyefﬁciency

Thedynamicdwellingsstockmodelcanbeusedasabasisfor detailedenergyanalysesofdwellingstocks[41].Althoughnofull energyanalysisiscarriedoutinthepresentstudy,themodelresults indicatetheimportanceofnewconstructionversusrenovationof existingdwellingsinthedwellingstocksofthevariouscountries towards2050.

ThesegmentedresultspresentedinFig.4indicatethatthereare largedifferencesintheimportanceofthedifferentcohortsandthat thereisalargedifferenceinthepotentialfor,andnecessary strat-egytodeliver,futureenergy-demandreductions.Countrieswith agrowingstockhavetoimprovetheaverageenergyefﬁciencyof thestockeventokeepthefuturetotalenergyuseinthestockat aconstantlevel.Incontrast,countrieswithadecreasingstockwill achieveenergysavingseveniftheaverageenergyefﬁciencydoes notimprove.

Eurostat[46]report214millionhouseholds(dwellings)in2013 fortheEU28countries.Thisstudyof124milliondwellingsin10

(12)

Fig.8.Numberofdwellingsconstructedorrenovatedintheperiod2016–2050.Left:aspercentageof2050stocksizeineachcountry.Right:absolutenumberofdwellings.

EUcountriesandNorway,thereforeanalysesoverhalfofEuropean housingstock.Iftheresultsarerepresentativeoftheremaining dwellingstock withinEurope,theexpectedlow rates of future renovationactivityposeafargreaterchallengeforpolicymakers seekingtorapidlymeetdecarbonisationobjectivesthanassumed incurrentstrategies.

Thesimulatedaccumulatednumberofdwellingsconstructed orrenovatedintheperiod2016–2050isshownforallcountries inFig.8(left)asaportionofthe2050stocksizeintherespective countryandFig.8(right)asabsolutenumberofdwellings.Aseach deeprenovationandeachnewconstructionisanopportunityto implementenergy-efﬁciencymeasures,thisindicatesthe poten-tialforenergysavingsineachcountry.“Renovation2016–2050” isthesumofthenumberofdwellingsrenovatedeachyearinthe sameperiod.Thisdoesnotcorrespondcompletelywiththe num-bersofdwellingsgoingthroughrenovationinthisperiod,asafew dwellingswillberenovatedtwice.Withina40-yearcycle, how-ever,onlyasmallproportionwillberenovatedtwiceina35-year period.

Fig.8(left) indicatesthesharesof the2050dwellingstocks thataretargetableforenergy efﬁciencymeasuresintheperiod 2016–2050,eitherastheyareconstructedorexposedtodeep reno-vation.Intotal,thiswillbe70–80%ofthe2050stockinallcountries exceptSerbia.Thismeansthateventhoughthesimulated renova-tionratesarenotexpectedtoincrease,thereisalargepotentialfor energyefﬁciencyofthedwellingstockstowards2050.However, asmostdwellingsarerenovatedonlyonceinthisperiod,itwill benecessarytoensurethatthebestavailableenergymeasuresare includedwhenadwellingundergoesrenovation,andtostimulate large-scaleintroductionoftechnologiessuchasheatpumpsand photovoltaics.

Fig.8(left)furtherdemonstrates howthecountrieswithan expecteddecreasingpopulation(Germany,Hungaryand Serbia) willhaveacorrespondinglylowshareofnewdwellingsin2050. Hence,thelargestpotentialfortotalenergysavingsinthese coun-triesisthroughrenovationandupgradingoftheexistingstock.In contrast,countrieswithalargeexpectedgrowthinthedwelling stock,likeCyprus,GreeceandNorway,willhaveahigh energy-savingpotentialinthedwellingsconstructedinthefuture.

Fig.8(right)shows theabsolutevaluesof newconstruction andrenovationinthe11countries,indicatingalsothetotal2050 stocksize.AsFrance,GermanyandGreatBritaincontain77%ofthe dwellingsconsideredhere,acceleratedstimulusinthesecountries wouldcontributemoretoachievingaEurope-wide decarbonisa-tiontarget.

Moredetailedstudiesontheenergystandardofthedwelling stocksineachofthecountries,andtheirpotentialsforcost-efﬁcient

reductions through introduction of energy-efﬁciency measures, couldbecombinedwiththeresultsfromthedynamicdwelling stockmodeltoidentifythemostcost-effectiveenergyefﬁciency schemeforhousingstocksthroughoutEurope.

4. Conclusions

Adynamicdwellingstockmodelisappliedto11European coun-tries(inalphabeticalorder):Cyprus,theCzechRepublic,France, Germany,GreatBritain,Greece,Hungary,theNetherlands,Norway, Serbia and Slovenia. The simulated long-term development in dwellingstocksizeﬁtswellwiththereportedstatisticsforall coun-tries.Despitethedifferencesindatacollectionandreportingbythe differentcountries,themodelledfuturetrendsforthe construc-tion,demolitionandrenovationactivitiesleadtosimilarpatterns emerginginallcountries.

Thecountriesincluded in thestudyhave differentexpected futuredevelopmentinthesizeofthepopulation.Insome coun-tries,astrongincreaseisexpected,whereasinothercountriesthe populationisexpectedtolevelofforevendecrease.Thishaslarge impactsonthesimulatedfuturedevelopmentofthedwellingstock. Countrieswithlargepopulationgrowthwillneedhigh construc-tionactivity,andforenergysavingmattersitisimportantthatthe newconstructionisenergyefficient.Incountriesexpectingalower populationgrowthordecreasingpopulation,theexistingstockis ofhigherimportanceandenergyefficiencyachievementsaremore influencedthroughrenovationoftheexistingstock.

Overall,thepresentedanalysisshowsthatdespitethe differ-encesbetweenthecountriesincludedinthisstudy,themodelis applicableforallthe11countries.Themodelisabletoreproduce thecurrentstocksizeandcompositionandthelong-termdynamics inthesysteminanacceptableway.Short-andmedium-term vari-ationsinconstructionanddemolitionactivitiesmaybeexplained byfactorsnotincludedinthemodel,e.g.widerdriverssuchas eco-nomic,climateandunemployment.Unfortunately,demolitionand renovationstatistics arerarelyavailable.Betterdataavailability wouldbeusefulformodelcalibration.

The future development in construction and demolition is sensitiveto thepopulationinput and thelifetime of dwellings parameter,whichare highlyuncertain.Still, weclaim thatit is bettertoincludethebestavailableestimatesoftheseimportant parametersinthestudyandidentifytheimplicationsoftheir uncer-tainty,ratherthanusingtraditionalmodelswithﬁxedconstruction, demolitionandrenovationratesthatarebasedonrecenttrends withoutdiscussingtheirrealismandapplicabilityforfuture analy-ses.

(13)

We conclude that the model seems to perform reasonably well at simulating the long-term development of changes in dwellingstockcompositionandexpectedannualrenovation activ-ities.Short-andmediumtermvariationsinconstructionactivity arenotwellcapturedbythemodel,asthesedependondrivers notrepresentedinthemodel.However,short-andmediumterm estimationsarenottheintendedpurposeofthemodel.Thegeneral trendobservedinmostofthecountriesstudied,isthatthemodelled shareofthestockconstructedbefore1945showsagoodﬁtwith thereportedstatistics,whereasthepost-warconstructionactivity isoftenunderestimatedandtheconstructionactivityfromthemost recentdecadesisconsequentlyoverestimated.Althoughtheoverall accuracyappearssatisfying,cautionshouldbeappliedwhen inter-pretingthesegmentedresultsfordifferentcohorts.Additional,or moreaccuratedata,couldimprovethequalityofthesegmented resultsforsomeofthecountries.

Akeymodeloutputistherenovationrate,whichexpressesthe dwellingstock’sneedformaintenanceduetoageing.The simu-lationsshowthatonlyminorincreasesareexpectedinthefuture renovationrate,alwayswithintherangefrom0.6%to1.6%towards 2050.Althoughthereareuncertaintiesintheresults,thistrendof smallincrementalchangesisconsistentacrossall11countriesand theresultsreinforceﬁndingsinSandbergetal.[37,38]that reno-vationratesatlevelsof2.5–3%areunlikelytobeachievedthrough thestock’snaturalrenovationrequirements.Furthermore,itshall benotedthatthesimulatedfuturerenovationratetowards2050 mainlydepends onthecurrent stocksize andcomposition and isnotsigniﬁcantlysensitivetofuturedevelopmentin theinput parametersofthemodel(suchasprojectionsonthepopulation development),asshowninthesensitivityanalysis[38].We there-foreconcludethatthemodelresultsonfutureneedforrenovation arerobustdespitetheuncertaintiesintheinputparameters.

Asfuturerenovationratesareexpectedtoremainclosetothe currentlevel,itishighlyimportanttomakesurethatthebest avail-ableenergy-efﬁciencymeasuresareincludedwhenadwellingis renovated.IfEuropeancountriesaregoingtofollowthe recom-mendationsgivenintheEU-JRCreport[3]callingforacommonEU renovationplanwitharegionalapproachprioritizingless devel-opedregions,fundingandotherincentivesmustbeallocatedso thatenergy-efﬁciencymeasuresareincludedwhendwellingsin thelessdevelopedregionsarerenovated,oreventoacceleratethe renovationprocessinthesecountries.

Acknowledgements

This paper is published as a result of participation in the EPISCOPE research project (Energy Performance Indicator TrackingSchemesfortheContinuousOptimisationof Refurbish-ment Processes in European Housing Stocks), with co-funding from the ‘Intelligent Energy—Europe’ Programme, contract No. IEE/12/695/SI2.644739.

Incollaborationwiththeauthorsofthepaper,thefollowing partnersinthevariouscountrieshaveprovidedtheirexpertisein collectingor conﬁrmingassumptions and data onthe country-speciﬁc information: Felipe Vásquez (Norwegian University of ScienceandTechnology,NTNU)inthecaseoftheCzech Repub-lic and Germany,Santhiah Shanthirabalan(Pouget Consultants) inFrance,JackHulme(BuildingResearchEstablishment-BRE)in Great Britain,TamasCsoknyai(BudapestUniversityof Technol-ogyandEconomics)andJanosFarkasinHungary,MilicaJovanovic Popovic(UniversityofBelgrade)inSerbiaand AndraˇzRakuˇsˇcek (BuildingandCivilEngineeringInstituteZRMK)inSlovenia. Niko-lausDiefenbach(InstitutWohnenundUmwelt-IWU,Germany)has givensomeusefulgeneralcommentstothereportingofthework.

OliverHeidrichandRichardDawsonhavebeenfundedbythe EC7thFPprojectRAMSES:ReconcilingAdaptation,Mitigationand SustainableDevelopmentforCities(Ref308497).

AppendixA. Supplementarydata

Supplementarydataassociatedwiththisarticlecanbefound,in theonlineversion,athttp://dx.doi.org/10.1016/j.enbuild.2016.05. 100.

References

[1]D.Reckien,J.Flacke,R.J.Dawson,O.Heidrich,M.Olazabal,A.Foley,J.J.P. Hamann,H.Orru,M.Salvia,S.DeGregorioHurtado,D.Geneletti,F. Pietrapertosa,ClimatechangeresponseinEurope:what’sthereality? Analysisofadaptationandmitigationplansfrom200urbanareasin11 countries,Clim.Change122(1–2)(2014)331–340.

[2]A.Kona,G.Melica,S.R.Calvete,P.Zancanella,A.Iancu,I.Gabrielaitiene,Y. Saheb,G.Janssens-Manhout,P.Bertoldi,in:P.Bertoldi(Ed.),TheCovenantof MayorsinFiguresandPerformanceIndicators:6-yearAssessment,Joint ResearchCentre—InstituteforEnergyandTransport,Ispra(VA),Italia,2015.

[3]Y.Saheb,K.Bódis,S.Szabó,H.Ossenbrink,S.Panev,EnergyRenovation:the trumpcardforthenewstartforeurope,in:JRCScienceandPolicyReports, EuropeanCommisionJointResearchCentreInstituteforEnergyand Transport,PublicationsOfﬁceoftheEuropeanUnion,Luxembourg,2015.

[4]BuildingsPerformanceInstituteEurope,Europe’sBuildingsUnderthe Microscope.ACountry-by-countryReviewoftheEnergyPerformanceof Buidlings,BuildingsPerformanceInstituteEurope,Brussels,2011.

[5]M.Stadler,L.Kranzl,C.Huber,R.Haas,E.Tsioliaridou,Policystrategiesand pathstopromotesustainableenergysystems–thedynamicInvertsimulation tool,EnergyPolicy35(1)(2007)597–608.

[6]EuropeanParliamentandtheCouncil,EnergyEfﬁciencyDirective,in:E.P.A.T.

Council(Ed.)2012/27/EU2012.

[7]T.Dixon,M.Eames,M.Hunt,S.Lannon,UrbanRetroﬁttingforSustainability: MappingtheTransitionto2050,Routledge,2014.

[8]B.K.Boermans,T.Offerman,M.Schimschar,SchimscharRenovationTracksfor Europeupto2050,Ecofys,Germany,2012.

[9]CentreforClimateChangeandSustainableEnergyPolicye,Employment ImpactsofaLarge-ScaleDeepBuildingEnergyRetroﬁtProgrammeinPoland, EuropeanClimateFoundation,TheNetherlands,2012.

[10]IPCC,ClimateChange,in:O.Edenhofer,R.Pichs-Madruga,Y.Sokona,S. Farahani,K.Seyboth,A.Adler,I.Baum,S.Brunner,P.Eickemeier,B.Kriemann, J.Savolainen,S.Schlömer,C.vonStechow,T.Zwickel,JC,J.Minx(Eds.), MitigationofClimateChange.ContributionofWorkingGroupIIItotheFifth AssessmentReportoftheIntergovernmentalPanelonClimateChange, CambridgeUniversityPress,32AvenueoftheAmericas,NewYork,NY 10013-2473,USA,2014.

[11]R.VillarroelWalker,M.B.Beck,J.W.Hall,R.J.Dawson,O.Heidrich,The energy-water-foodnexus:strategicanalysisoftechnologiesfortransforming theurbanmetabolism,J.Environ.Manage.141(0)(2014)104–115.

[12]D.Harvey,RecentAdvancesinSustainableBuildings:ReviewoftheEnergy andCostPerformanceoftheState-of-the-artBestPracticesfromAroundthe World,AnnualReviewofEnvironmentandResources,2013,pp.281–309.

[13]GEA,GlobalEnergyAssessment—TowardaSustainableFuture,GlobalEnergy Assessment(GEA),CambridgeUniversityPress,Cambridge,UKandNewYork, NY,USAandtheInternationalInstituteforAppliedSystemsAnalysis, Laxenburg,Austria,2012.

[14]L.Filogamo,G.Peri,G.Rizzo,A.Giaccone,Ontheclassiﬁcationoflarge residentialbuildingsstocksbysampletypologiesforenergyplanning purposes,Appl.Energy135(2014)825–835.

[15]T.Boermans,K.Dinges,J.Grözinger,M.Schäfer,A.Förster,I.Hermelink,B.V. Manteuffel,Publicconsultationontheevaluationofdirective2010/31/EU,in: FinalSynthesisReport,EuropeanCommission,Brussels,2015.

[16]A.Uihlein,P.Eder,Policyoptionstowardsanenergyefﬁcientresidential buildingstockintheEU-27,EnergyBuild.42(6)(2010)791–798.

[17]N.Garrido-Soriano,M.Rosas-Casals,A.Ivancic,M.D.Alvarez-delCastillo, Potentialenergysavingsandeconomicimpactofresidentialbuildingsunder nationalandregionalefﬁciencyscenarios.ACatalancasestudy,EnergyBuild. 49(2012)119–125.

[18]TheClimateandPollutionAgency,Klimakur,2020,TiltakOgVirkemidlerforå nåNorskeKlimamålMot2020,TheClimateandPollutionAgency,Oslo,2010.

[19]E.L.Glaeser,J.Gyourko,R.E.Saks,Urbangrowthandhousingsupply,J.Econ. Geogr.6(1)(2006)71–89.

[20]M.Kavgic,A.Mavrogianni,D.Mumovic,A.Summerﬁeld,Z.Stevanovic,M. Djurovic-Petrovic,Areviewofbottom-upbuildingstockmodelsforenergy consumptionintheresidentialsector,Build.Environ.45(7)(2010) 1683–1697.

[21]É.Mata,A.S.Kalagasidis,F.Johnsson,Amodellingstrategyforenergy,carbon, andcostassessmentsofbuildingstocks,EnergyBuild.56(2013)100–108.

[22]D.Wiedenhofer,J.K.Steinberger,N.Eisenmenger,W.Haas,Maintenanceand expansion:modelingmaterialstocksandﬂowsforresidentialbuildingsand transportationnetworksintheEU25,J.Ind.Ecol.19(4)(2015)538–551.

(14)

38 N.H.Sandbergetal./EnergyandBuildings132(2016)26–38 [23]F.Nemry,A.Uihlein,C.M.Colodel,C.Wetzel,A.Braune,B.Wittstock,I.Hasan,

J.Kreißig,N.Gallon,S.Niemeier,Y.Frech,Optionstoreducethe environmentalimpactsofresidentialbuildingsintheEuropean Union-Potentialandcosts,EnergyBuild.42(7)(2010)976–984.

[24]F.Meijer,L.Itard,M.Sunikka-Blank,ComparingEuropeanresidentialbuilding stocks:performance,renovationandpolicyopportunities,Build.Res.Inf.37 (5–6)(2009)533–551.

[25]S.Hrabovszky-Horváth,T.Pálvölgyi,T.Csoknyai,A.Talamon,Generalized residentialbuildingtypologyforurbanclimatechangemitigationand adaptationstrategies:thecaseofHungary,EnergyBuild.(2013)475–485.

[26]L.D.D.Harvey,K.Korytarova,O.Lucon,V.Roshchanka,Constructionofa globaldisaggregateddatasetofbuildingenergyuseandﬂoorareain2010, EnergyBuild.76(2014)488–496.

[27]E.Mata,A.SasicKalagasidis,F.Johnsson,Building-stockaggregationthrough archetypebuildings:france,Germany,SpainandtheUK,Build.Environ.81 (2014)270–282.

[28]N.Kohler,U.Hassler,Thebuildingstockasaresearchobject,Build.Res.Inf.30 (4)(2002)226–236.

[29]D.B.Müller,Stockdynamicsforforecastingmaterialﬂows—casestudyfor housinginTheNetherlands,Ecol.Econ.59(1)(2006)142–156.

[30]H.Bergsdal,H.Brattebø,R.A.Bohne,D.B.Müller,Dynamicmaterialﬂow analysisforNorway’sdwellingstock,Build.Res.Inf.35(5)(2007)557–570.

[31]I.Sartori,H.Bergsdal,D.B.Müller,H.Brattebø,Towardsmodellingof construction,renovationanddemolitionactivities:Norway’sdwellingstock, 1900–2100,Build.Res.Inf.36(5)(2008)412–425.

[32]M.M.Hu,S.Pauliuk,T.Wang,G.Huppes,E.vanderVoet,D.B.Muller,Ironand steelinChineseresidentialbuildings:adynamicanalysis,Resour.Conserv. Recycl.54(9)(2010)591–600.

[33]S.Pauliuk,K.Sjostrand,D.B.Muller,Transformingthenorwegiandwelling stocktoreachthe2degreescelsiusclimatetarget:combiningmaterialﬂow analysisandlifecycleassessmenttechniques,J.Ind.Ecol.17(4)(2013) 542–554.

[34]N.H.Sandberg,H.Bergsdal,H.Brattebø,Historicalenergyanalysisofthe Norwegiandwellingstock,Build.Res.Inf.39(1)(2011)1–15.

[35]N.H.Sandberg,H.Brattebø,Analysisofenergyandcarbonﬂowsinthefuture Norwegiandwellingstock,Build.Res.Inf.40(2)(2012)123–139.

[36]C.Gallardo,N.H.Sandberg,H.Brattebø,Dynamic-MFAexaminationofChilean housingstock:long-termchangesandearthquakedamage,Build.Res.Inf.42 (3)(2014)343–358.

[37]N.H.Sandberg,I.Sartori,H.Brattebø,Usingadynamicsegmentedmodelto examinefuturerenovationactivitiesintheNorwegiandwellingstock,Energy Build.82(2014)287–295.

[38]N.H.Sandberg,I.Sartori,H.Brattebø,Sensitivityanalysisinlong-term dynamicbuildingstockmodeling-Exploringtheimportanceofuncertaintyof inputparametersinNorwegiansegmenteddwellingstockmodel,Energy Build.85(2014)136–144.

[39]F.Vasquez,A.N.Lovik,N.H.Sandberg,D.B.Muller,Dynamictype-cohort-time approachfortheanalysisofenergyreductionsstrategiesinthebuildingstock, EnergyBuild.111(2016)37–55.

[40]I.Sartori,N.H.Sandberg,H.Brattebø,Dynamicbuildingstockmodelling:a

generalalgorithmandexempliﬁcationforNorway,EnergyBuild.132(2016)

13–25,http://dx.doi.org/10.1016/j.enbuild.2016.05.098.

[41]N.H.Sandberg,I.Sartori,M.I.Vestrum,H.Brattebø,Modellingthehistorical

energydemandwithasegmenteddynamicstockmodel:Exploringtheenergy

savingsalreadyachievedinthehousingstockofNorway1960–2015,(2015).

[42]P.J.Sereda,Durabilityofbuildingmaterialsandcomponents,in:Proceedings oftheFirstInternationalConference:ASymposium,Ottawa,Canada,1978.

[43]A.Gouldson,N.Kerr,J.Millward-Hopkins,M.C.Freeman,C.Topi,R.Sullivan, Innovativeﬁnancingmodelsforlowcarbontransitions:exploringthecasefor revolvingfundsfordomesticenergyefﬁciencyprogrammes,EnergyPolicy86 (2015)739–748.

[44]HellenicStatisticalAuthority,HousingCensus2011,2015(accessed15.10.15)

http://www.statistics.gr/portal/page/portal/ESYE/PAGE-themes?p param=A1302&rparam=SOP03&yparam=201400&mytabs=0.

[45]HellenicStatisticalAuthority,BuildingActivityStatistics2014,2015(accessed

15.10.15)http://www.statistics.gr/portal/page/portal/ESYE/PAGE-themes?p

param=A1602&rparam=SAM05&yparam=201100&mytabs=0.

[46]Eurostat,PeopleintheEU:whoareweandhowdowelive?2015(accessed

20.06.16)