Effects of white roofs on urban temperature in a global climate modelK. W. Oleson,1G. B. Bonan,1and J. Feddema2Received 18 December 2009; accepted 4 January 2010; published 3 February 2010.[1] Increasing the albedo of urban surfaces has receivedattention as a strategy to mitigate urban heat islands. Here,the effects of globally installing white roofs are assessedusing an urban canyon model coupled to a global climatemodel. Averaged over all urban areas, the annual mean heatisland decreased by 33%. Urban daily maximum temperaturedecreased by 0.6°C and daily minimum temperature by0.3°C. Spatial variability in the heat island response is causedby changes in absorbed solar radiation and specification ofroof thermal admittance. At high latitudes in winter, theincrease in roof albedo is less effective at reducing the heatisland due to low incoming solar radiation, the high albedo ofsnow intercepted by roofs, and an increase in space heatingthat compensates for reduced solar heating. Global spaceheating increased more than air conditioning decreased,suggesting that end-use energy costs must be considered inevaluating the benefits of white roofs.Citation: Oleson,K. W., G. B. Bonan, and J. Feddema (2010), Effects of white roofson urban temperature in a global climate model, Geophys. Res.Lett., 37, L03701, doi:10.1029/2009GL042194.1. Introduction[2] An analysis by Akbari et al. [2009] indicated thatincreasing the albedo of urban roofs and pavements globallycould produce a negative radiative forcing equivalent to a44 Gt CO2emission offset. This is equivalent to offsettingthe effect of the growth in CO2-equivalent emission rates forthe next 11 years. It was also noted that potential energysavings may be realized due to a reduction in the amount ofenergy consumed by air conditioning to cool buildings. Theeffects of increasing the albedo of cities on near-surfaceclimate, in particular air temperature, were not addressed.[3] Surfaces with higher albedo reflect more solar radia-tion, thereby decreasing surface temperature and heating ofthe surrounding air. Several studies have quantified theability of increases in albedo to mitigate the urban heatisland and decrease cooling energy use. Generally, thesehave been conducted with mesoscale weather models appliedto individual cities using large-scale albedo changes (i.e.,changing the albedo of entire cities). Sailor [1995] showedthat increasing the albedo in Los Angeles decreased peaksummertime temperatures by as much as 1.5°C. Taha et al.[1999] showed that large-scale increases in surface albedofor ten cities in the U.S. reduced the near-surface daytimesummer air temperature by 0.5 to 1.5°C and decreased peakelectricity demand by up to 10%. Synnefa et al. [2008]found that large-scale increases in roof albedo decreased thesummer heat island intensity in Athens, Greece by 1–2°Con average.[4] Here, we examine the impact of an increase in roofalbedo on near-surface urban climates using an urbancanyon model coupled to a global climate mo del. Theprimary purpose of the urban canyon model is to providean estimate of near-surface air temperature for urban areas,which is where most people live. Although the influence ofurban areas on large-scale climate is small, a coupled mod-eling approach allows assessment of the impacts of chang-ing climate on urban populations and exploration of climatechange mitigation options. The purpose of this study is tohighlight issues with heat island mitigation using whiteroofs and to identify what processes must be consideredwhen evaluating the effectiveness of this urban heat islandmitigation method.[5] The canyon model allows us to make changes in roofalbedo only, rather than city-wide changes. In contrast toAkbari et al. [2009], who increased the albedo of roofs andpavement, we modified only roof albedo because roofscomprise about 40% of the global urban horizontal surfacesin the model while the impervious pavement is only about15% of the urban surface (T. Jackson et al., Parameteriza-tion of urban characteristics for global climate modeling,submitted to Annals of the Association of American Geog-raphers, 2009). Thus, changing roof albedo should have thelargest impact on near-surface urban climate. The urbanmodel also estimates large-scale space heating and airconditioning (HAC) fluxes by controlling internal buildingtemperatures within specified comfort levels. The effect ofincreased albedo on these fluxes is also quantified.2. Data and Methods[6] The urban canyon model CLMU [Oleson et al., 2008a,2008b] is coupled to the Community Land Model version 3.5(CLM3.5) [Oleson et al., 2008c] and the CommunityAtmosphere Model version 3.5 (CAM3.5) [Neale et al.,2008], which are the land and atmospheric components ofthe Community Climate System Model (CCSM) [Collins etal., 2006]. The canyon system consists of roofs, walls, andcanyon floo r. Walls are divided into shaded and sunlitcomponents. The canyon floor is divided i nto pervio us(greenspace) and impervious (pavement) fractions. Theurban components are arranged in an ‘‘urban canyon’’configuration [Oke, 1987] in which the canyon geometryis described by building height and street width. Theboundary conditions for heat transfer within roofs and wallsare determined by an interior building temperature heldbetween prescribed minimum and maximum temperatures,thus explicitly resolving HAC fluxes. Sources of waste heatfrom inefficiencies in the HAC systems are incorporated asterms in the canyon energy budget. The heat and moistureGEOPHYSICAL RESEARCH LETTERS, VOL. 37, L03701, doi:10.1029/2009GL042194, 2010ClickHereforFullArticle1Climate and Global Dynamics Division, National Center for Atmo-spheric Research, Boulder, Colorado, USA.2Department of Geography, University of Kansas, Lawrence, Kansas,USA.Copyright 2010 by the American Geophysical Union.0094-8276/10/2009GL042194$05.00L03701 1of7fluxes from each surface (including the roof) interact witheach other through a bulk air mass that represents air in theurban canopy layer. The urban model produces turbulent,momentum, and radiative fluxes which are area-averagedwith fluxes from non-urban surfaces (e.g., vegetation andlakes) to supply grid cell averaged fluxes to the atmosphericmodel. The version of CLMU used here is the same as thatused by Oleson et al. [2008a] but with improvements to thepervious greenspace hydrology and a revised treatment ofHAC and waste heat fluxes that results in a more stablenumerical solution (see