Applied Energy 193 (2017) 276–286

Contents lists available at ScienceDirect

Applied Energy

journal homepage: www.elsevier .com/ locate/apenergy

Estimating natural ventilation potential for high-rise buildings

considering boundary layer meteorology

http://dx.doi.org/10.1016/j.apenergy.2017.02.0410306-2619/� 2017 Elsevier Ltd. All rights reserved.

⇑ Corresponding authors at: Center for Green Buildings and Cities, HarvardUniversity, Cambridge, MA 02138, USA.

E-mail addresses: ztong@gsd.harvard.edu, zht940@g.harvard.edu (Z. Tong),ychen@gsd.harvard.edu (Y. Chen).

Zheming Tong a,b,⇑, Yujiao Chen a,b, Ali Malkawi a,baCenter for Green Buildings and Cities, Harvard University, Cambridge, MA 02138, USAbGraduate School of Design, Harvard University, Cambridge, MA 02138, USA

h i g h l i g h t s

� The vertical profiles of NV potential are estimated for high-rise buildings.� Atmospheric stability has a great impact on NV potential.� An in-house ABL meteorology model is developed.� The ABL meteorology model can be coupled with building-resolved CFD simulation.� Los Angeles displays the greatest NV potential with about 7000 NV hours per year.

a r t i c l e i n f o

Article history:Received 23 December 2016Received in revised form 11 February 2017Accepted 13 February 2017

Keywords:Natural ventilationHigh-rise buildingMixed-mode ventilationMeteorologyAtmospheric boundary layer (ABL)NV hour

a b s t r a c t

The design of energy conservative buildings that incorporates natural ventilation (NV) strategy hasbecome increasingly popular around the world. Natural ventilation is a key solution for reducing energyconsumption of buildings and for maintaining a healthy indoor environment. However, the adoption ofnatural ventilation in high-rise buildings is less common. As rapid population growth and urbanizationtake place in cities, it is important to explore the substantial energy saving potential of high rises by uti-lizing natural ventilation. In this study, we have provided the early effort to estimate quantitatively thevertical profiles of NV potential for high rises at major cities from six climate zones in the U.S. (i.e., Miami,Houston, Los Angeles, New York City, Chicago, and Minneapolis), using an in-house boundary layer mete-orology model. The diurnal cycle of atmospheric boundary layer (ABL) and local climate characteristicsare found to have a great effect on the vertical structure of NV potential. In general, negative vertical gra-dients of NV hours are observed for all cities except Miami where the vertical distribution is nearly uni-form. For example, the annual NV hour decreases from 7258 at ground level to 4866 at 300 m above theground in Los Angeles. Our analysis shows that outdoor temperature is a key meteorological parameterthat determines vertical profiles of NV hours in New York City, Los Angeles, Chicago, and Minneapolis. Incontrast, humidity plays a greater role in cities like Miami and Houston where the outdoor temperature isoften favorable for using natural ventilation except in the summer. Among studied cities, Los Angeles pro-vides the ideal climate (warm and dry) for utilizing natural ventilation, displaying the greatest NV poten-tial (7258 NV hours or 83% time of the year at ground level), followed by New York City with 3360 NVhours. The remainder of the four studied cities display comparable numbers of NV hours of approxi-mately 2500 at ground level. The methodology and findings from this study are intended to assist archi-tects and policy makers in quantifying the potential energy savings of natural ventilation, and illustratingthe importance of considering the vertical variations of elevated thermal environment in high-rise build-ings across different climate zones in the U.S.

� 2017 Elsevier Ltd. All rights reserved.

1. Introduction

Rapid population growth and urbanization have led to increas-ing energy demand in many cities of the world. Given the large vol-ume and occupancy of the fast-growing number of high-risebuildings, the energy consumption of these buildings is enormous

Table 1Definition of eight climate zones in the U.S.

Zonenumber

Zone name Thermal criteria (SI Units)

1A and 1B Very hot – humid (1A) Dry(1B)

5000 < CDD10 �C

2A and 2B Hot-humid (2A) dry (2B) 3500 < CDD10 �C � 50003A and 3B Warm – humid (3A) dry (3B) 2500 < CDD10 �C < 35003C Warm – marine (3C) CDD10 �C � 2500 AND

HDD18 �C � 20004A and 4B Mixed-humid (4A) dry (4B) CDD10 �C � 2500 AND

HDD18 �C � 30004C Mixed – marine (4C) 2000 < HDD18 �C � 30005A, 5B, and

5CCool-humid (5A) dry (5B)marine (5C)

3000 < HDD18 �C � 4000

6A and 6B Cold – humid (6A) dry (6B) 4000 < HDD18 �C � 50007 Very cold 5000 < HDD18 �C � 70008 Subarctic 7000 < HDD18 �C

Z. Tong et al. / Applied Energy 193 (2017) 276–286 277

in which the heating, ventilation and air conditioning (HVAC) sys-tems account for roughly 33% of the overall energy consumptions[1]. Many advanced technologies have been developed to reducebuilding energy consumption [2–20]. Among them, natural venti-lation (NV) that supplies and removes air to and from an indoorspace using natural forces of wind and buoyancy shows greatpotential to reduce energy consumption, and to reduce the costof the HVAC system in high-rise buildings [10,21–23].

Although most high-rise buildings are fully sealed, the numberof successful examples that incorporates natural ventilation princi-ple is growing globally. For instance, the Commerzbank Tower isperhaps the most well-known high rise that utilizes mixed-modeventilation (i.e., using natural ventilation for periods when theexternal weather conditions are allowed, but mechanical ventila-tion takes over when external weather conditions are not suitable).It was built in 1997 in Frankfurt (temperate climate), with a floorarea of about 70,000 m2 providing working space to 2500 employ-ees. The tower features a double-skin façade with automated open-ing control, central atrium (every 12 stories) and 4-story high skygardens. The building operates on mixed-mode ventilation, inwhich natural ventilation is utilized approximately 80% of the year.Another well-known example that features natural ventilation is30 St. Mary Axe in London. It was completed in 2004 with a grossarea of 64,470 m2. The building was designed with double-skinfaçades and stepping atria that tempers air before being dis-tributed to office. It is designed to rely on natural ventilation forabout 40% of the year.

There are many advantages of utilizing natural ventilation inhigh-rise buildings. For instance, high rises are less influenced bythe surroundings as opposed to low-rise buildings, which resultsin sufficient driving force from ambient wind to achieve desiredair change rate (ACH) for natural ventilation [24]. Additionally, incomparison to low-rise buildings, buoyancy-driven natural venti-lation can be utilized more effectively due to vertical pressure vari-ation in high rises through atria, sky gardens, solar chimneys, etc. Anumber of studies have looked at various aspects with regard tothe design of naturally ventilated high-rise buildings. Zhou et al.[25] employed CFD simulation to optimize natural ventilation forhigh-rise residential buildings by adjusting the building orienta-tion, building spacing, window opening positions, etc. Pasquay[26] conducted a series of onsite measurements of buildings withdouble skin façades that are commonly seen on high-rise buildingsto regulate incoming wind speed, air temperature, and outsidenoise. The results from his study confirmed some advantages ofdouble-skin façade in noise control, shading device integrationand night cooling, but also raised some concerns about overheatingin the summer. Niu [27] proposed using window vents as an alter-native for natural ventilation in high-rise buildings. His analysisshows that window vents can provide constant air flow by self-regulating the opening degree in response to pressure differences.Prajongsan and Sharples [28] demonstrated that a ventilation shaftis an effective way to enhance single-sided natural ventilation intall residential buildings in Bangkok, in which the shaft canincrease the pressure difference between windows and the shaft’sexhaust at roof level. Similar stack systems have been evaluatedthrough experiments by Priyadarsini et al. [29]. They found thatthe fan-assisted active stack system substantially increased theair change rate in the room with limited access to externalwindows.

Existing studies mostly focused on specific design strategies fornaturally ventilated high rises. The effect of boundary layer mete-orology on natural ventilation potential has never been investi-gated. Atmospheric boundary layer (ABL) is the lowest portion ofthe troposphere that directly interacts with the built environmenton the ground surface, and responds to surface forcing with a timescale of about an hour or less. The influence of surface friction,

heating and evaporation is transmitted to the entire ABL throughthe mechanism of turbulent mixing. The height of the ABL overthe land surface varies depending on the rate of heating or coolingof the surface, wind speed, surface roughness, etc. High-rise build-ings that either fully rely on natural ventilation or operate onmixed-mode closely interact with the ABL. Therefore, to quantifythe NV potential of high rises, it is critical to understand the verti-cal structure of ABL, because many meteorological parameters(e.g., wind speed, temperature, and humidity) vary considerablywith altitude and time.

In this study, we have provided an early effort to understandand estimate the vertical profiles of NV hour at major U.S. citiesfrom six climate zones using an in-house ABL meteorology model.The paper is organized as follows. We first describe the climaticcharacteristics in the U.S., and sources of surface observation andupper air weather data. Next we elaborate the in-house boundarylayer meteorology model for estimating the vertical profiles ofmeterological variables. In the second part of the paper, we presentand discuss the results, followed by a summary of key findings.

2. Methodology

2.1. Climate characteristics of selected cities

Climate varies widely across the U.S. due to the massiveexpanse of land and complicated terrain. The classification of cli-mate zones developed by U.S. Department of Energy’s PacificNorthwest National Laboratory (PNNL) is adopted in this study todifferentiate various climates (Table 1). The map shown in Fig. 1is based on analysis of National Oceanic and Atmospheric Admin-istration (NOAA) weather sites, and is a widely accepted classifica-tion of U.S. climate [30]. This map divided the U.S. into eighttemperature-oriented climate zones (Zone 1 to 8, Very Hot toSubarctic). Each zone is further divided into three moistureregimes designated A (moist), B (dry), and C (marine).

We selected one populated city (i.e., Miami, Houston, Los Ange-les, New York City (NYC), Chicago, and Minneapolis), from six cli-mate zones in the U.S. (Fig. 1). The boxplots of temperature andwind speed by month collected at airport surface observation sta-tions at each city for the year of 2011 are presented in Fig. 2. In hotand humid climate, such as Houston and Miami, it is often chal-lenging to utilize natural ventilation. However, large air changerates (ACH) as a result of natural ventilation helps to removeunwanted humidity, and to widen the thermal comfort range. Inhot summer and cold winter climates, such as Chicago and NYC,buildings can be naturally ventilated for the majority of time fromMarch to October. In the summer, night-purge ventilation, coupled

Fig. 2. Box plots of the monthly temperature and wind speed at Miami, Houston, Los Angeles, New York City (NYC), Chicago, and Minneapolis.

Fig. 1. Classification of climate zones in the U.S.

278 Z. Tong et al. / Applied Energy 193 (2017) 276–286

Z. Tong et al. / Applied Energy 193 (2017) 276–286 279

with high thermal mass of the building can be a very effectivestrategy if diurnal temperature differences are significant. Temper-ate climate is the ideal climate for natural ventilation, which can beseen in many cities in California (e.g. Los Angeles). If designedproperly, thermal comfort can be achieved by natural ventilationalone for the majority of time throughout the year, without theneed to operate HVAC system. In cold climates, such as Min-neapolis, natural ventilation can be used extensively in relativelywarm months (April to October), because building heat loss is aprimary concern.

The number of existing high-rise buildings that is greater than100 m at each major city in the U.S. is overlaid on the map(Fig. 1). Based on the 2016 database by the Council on Tall Build-ings and Urban Habitat (CTBUH), NYC, the city with the largestnumber of high rises in the U.S., has 771 buildings with a heightgreater than 100 m, followed by 304 in Chicago. Miami and Hous-ton are ranked 3rd and 4th on the list with 86 and 85 buildings(>100 m), respectively. Among the rest of studied cities that westudied, Los Angeles has 63 buildings (>100 m), and Minneapolishosts 26 buildings (>100 m).

2.2. Meteorology and land cover data

The hourly surface data were obtained from a network of Auto-mated Surface Observation Stations (ASOS) that are located at air-port with anemometer standing 10 m above the ground level. Theupper air data were obtained from the Earth System Research Lab-oratory (ESRL) at National Oceanic and Atmospheric Administra-tion (NOAA). These data were collected using radiosondes, whichis an expendable instrument package attached to a large balloon.The sensors in the package measure profiles of pressure, tempera-ture, and relative humidity as the radiosonde rises. Readings weretaken twice a day at a small number of stations within each state.Surface characteristic values, including albedo, Bowen ratio, andsurface roughness length, were calculated according to land coverdata from U.S. Geological Survey (USGS) National Land Cover Data(NLCD) archives, which is a 21-class land cover classificationscheme that is applied consistently over the U.S. The spatial reso-lution of the data is 30 m and mapped in the Albers Conic EqualArea projection. Each classification is linked to a set of seasonalsurface characteristics that were developed by U.S. EPA [31]. Themeteorology and land cover data used in this study are from theyear of 2011 due to the quality of available data.

2.3. Boundary layer meteorology model

To quantify the thermal environment and NV potential on eachfloor of high-rise buildings, it is critical to understand the verticalstructure of wind, temperature and humidity in the ABL. Thewidely-used logarithmic wind profile is valid only for the neutralatmosphere in the lower ABL. As the atmosphere becomes unstableor stable, the profiles depart from being logarithmic. Atmosphericstability is a measure of the atmosphere’s tendency to encourageor deter vertical motion, which can be classified as stable, neutralor unstable conditions. Fig. 3 displays an example of vertical pro-files of velocity (u) and potential temperature (h) under differentstability conditions. The ABL is unstable or convective when thesurface is warmer than the air, such as during sunny days withlight winds over land. The ABL is stable when the surface is colderthan the air, such as during a clear night over land. The ABLapproaches neutrality when the thermal exchanges between airand surface are minimal, such as windy and overcast conditions.The ABL meteorology model presented here is based on Monin-Obukhov similarity theory (MOST). It is generally accurate for con-structing vertical profiles at low altitude (i.e., less than 300 m)according to experimental observations in the past [32,33].

Entrainment of free atmospheric air and the effect of the coriolisforce are not taken into account in our approach. The neglectionof these factors has little impact on our prediction, because we pri-marily focus on high-rise buildings, not ‘‘supertall” buildings thatare greater than 300 m, as defined by CTBUH [1]. The surroundingbuildings are not explicitly modeled. Instead, we implicitly takeinto account their effect in constructing vertical profiles usingparameterized surface roughness length. In addition, the meteorol-ogy model introduced here can be easily coupled with building-resolved Computational Fluid Dynamics (CFD) simulation at urbanneighourhood-scale through inlet boundary condition [34–37]. Thecomputing cost for coupled CFD simulation is usually a few ordersof magnitudes greater than the parameterization approach. It ishowever suitable for site-specific case studies where the buildingstructures near the location of interest are geometrically complexand building-to-building interactions must be considered.

2.3.1. Vertical profiling functionsA semi-empirical model based on MOST is developed here to

capture the vertical variation of meteorological quantities underdifferent stability conditions. We employed Monin-Obukhovlength (L), the height at which the production of turbulence byboth mechanical and buoyance forces is equal, to quantify the sta-bility of the lower ABL. L is computed by Eq. (1).

L ¼ �qcpTref u3�

kgHð1Þ

where g is the acceleration of gravity, k is the von Karman constant,cp is the specific heat of air, q is the density of air, and Tref is the ref-erence temperature. H is the surface sensible heat flux. u� is the fric-tion velocity, which is determined by the atmospheric stability. Byits definition, L > 0 indicates stable conditions, L < 0 indicatesunstable conditions, and L ¼ 1 applies to neutral conditions. Inthe unstable boundary layer, since u� and L depend on each other,they are estimated using an iterative method (Eq. (2)) developedby Perry [38].

u� ¼ kuref

ln zrefz0

� �� wm

zrefL

� � ð2Þ

where uref is the reference velocity from the weather station at areference height zref , and wm is the similarity function for windspeed. In stable case, u� is determined by solving a quadratic equa-tion shown in Eq. (3) [39].

u2� � CDuu� þ CDu2

0 ¼ 0 ð3Þwhere u2

0 ¼ bmzref gh�=Tref , and the drag coefficient CD is defined ask= lnðzref =z0Þ [40]. The wind speed u, potential temperature h, andhumidity q profiles considering the effect of atmospheric stabilityare represented by Eqs. (4)–(6).

uu�

¼ 1k

lnzz0

� �� wm

zL

� �� �ð4Þ

h� h0h�

¼ 1k

lnzz0

� �� wh

zL

� �� �ð5Þ

q� q0

q�¼ 1

kln

zz0

� �� ww

zL

� �� �ð6Þ

where z=L is the non-dimensional heightwm,wh, andww are similarityfunctions for wind speed, potential temperature, and absolutehumidity, respectively. h� is the temperature scale. h0 is the referencepotential temperature at measurement height. q� is the humidityscale. q0 is the reference absolute humidity at measurement height.The presence of surrounding buildings is not directly modeled.

Fig. 3. Vertical velocity [m/s] and potential temperature [K] profile under unstable, neutral and stable conditions.

280 Z. Tong et al. / Applied Energy 193 (2017) 276–286

Instead, their effect is parameterized in terms of surface roughnesslength, which is obtained from the NLCD archives. h� is calculatedfrom sensible heat flux H and friction velocity u⁄ by Eq. (7).

h� ¼ � Hqcpu�

ð7Þ

In order to estimate of H, parameterizations are made based onthe energy balance approach by Eq. (8) [41].

H ¼ 0:9Rn

1þ 1=B0ð8Þ

where Rn is the net radiation, and B0 is the Bowen ratio, which aredetermined from the USGS National Land Cover Data archives. Thehumidity scale q� is computed by Eq. (9).

q� ¼ � LEqLvu�

ð9Þ

where LE is latent heat flux, and Lv is the latent heat of evaporation.The similarity function wm, wh, ww in Eqs. (4), (5), and (3) can bedetermined from the well-known Businger-Dyer relationships,which is recommended for most practical application [32,42]. Inthe stable condition, a modified similarity function is described inEq. (10) recommended by Van Ulden and Holtslag [43].

wm ¼ wh ¼ wq ¼ �17 1� e�0:29zL�

forzL� 0 ð10Þ

In the unstable condition, the similarity functions of the windvelocity wm are written according to Eq. (11), and the similarityfunctions of the potential temperature wh and humidity wq aredescribed in Eq. (12).

wm ¼ 2 ln1þ x2

� �þ ln

1þ x2

2

� �� 2 tan�1 xþ p

2for

zL< 0

ð11Þ

wh ¼ wq ¼ 2 ln1þ x2

2for

zL< 0 ð12Þ

where

x ¼ 1� 16z=Lð Þ0:25 ð13ÞIn the neutral condition,

wm ¼ wh ¼ wq ¼ 0 ð14ÞSince absolute temperature T is used for the calculation of NV

hours, h is converted to T using Eq. (15).

T ¼ hPP0

� � Rcp

ð15Þ

where R=cp ¼ 0:286. P0 is the pressure on the ground level, and P iscomputed as a function of altitude using Barometric formula.

2.3.2. Boundary layer depthThe profiling functions described in Section 2.3.1 are generally

valid until the height of ABL, which evolves with the diurnal cycleof ABL. During the early morning, the ABL is shallow. After the sun-rise, the warmed ground due to solar radiation heats the air above.The unstable (or convective) ABL depth hCBLðL < 0Þ starts to grow indepth and reaches a maximum depth in late afternoon. After thesunset, the surface cooling starts to create a stable ABL with depthhSBL above which is the residual layer, the leftover part of the day-time ABL. The time evolution of hCBL is simulated using numericalintegration of mixed-layer slab models [44]. This model usespotential temperature sounding prior to sunrise and time-dependent surface heat flux H(t) to calculate the temporal varia-tion of hCBL. The model assumes that the heat flux brought intothe boundary layer due to entrainment is a fraction (20%) of thesurface sensible heat flux H. The height of the stable boundarylayer hSBLðL < 0Þ is determined based on the approach by Zilitinke-vich [45] in Eq. (16).

hSBL ¼ Cu�Lf

� �0:5

ð16Þ

C is an empirical constant with a value of 0.4 according to the sug-gestion from Brost and Wyngaard [46]. f is Coriolis parameterdefined as Eq. (17).

f ¼ 2X sinðuÞ ð17ÞX is the rotation rate of the Earth (7.29 � 10�5 rad/s), and u is

the latitude of each studied city. In the case where the ABL is suf-ficiently neutral u�=ðfLÞj j < 4, the ABL height can be estimatedusing Eq. (18).

hn ¼ Cu�

fð18Þ

In the mixed or residual layer that is above the height of hCBL orhSBL, u is assumed to be equal to uðhCBLÞ or uðhSBLÞ, and the verticalgradients du

dz ;dhdz ;

dqdz can be neglected due to strong vertical mixing

[47].

Z. Tong et al. / Applied Energy 193 (2017) 276–286 281

2.4. Estimation of NV hours

NV hours for each floor of a high-rise building, using number ofhours per year as the unit, is defined as when outdoor weather (i.e.,wind speed, temperature, humidity) is suitable for natural ventila-tion. The overall flow chart of estimating the vertical profiles of NVhours is shown in Fig. 4. The methodology of calculating verticalprofiles of meteorological variables in the ABL is discussed previ-ously in Section 2.3. An in-house program is developed in MATLABto perform the simulation. The thresholds for determining NVhours are elaborated in this section. The temperature threshold ischosen when the outdoor dry-bulb temperature is below the upperthreshold, Tup, of adaptive thermal comfort model defined by deDear and Brager [48], and greater than 12.8 �C (the lowest supplyair temperature specified in ASHRAE 55 and fundamental hand-book to avoid unpleasant draft to occupants). The dew point tem-perature threshold is chosen when the outdoor dew pointtemperature Tdew is below 17 �C for the sake of humidity control[49,50]. The upper threshold of adaptive thermal comfort modelTup on each floor is calculated by Eq. (19).

TupðzÞ ¼ 0:31ToutðzÞ þ 17:8þ 3:5 ð19Þwhere Tout (�C) is mean monthly outdoor air temperature on eachfloor. The upper threshold is based on 80% thermal acceptability,which produces a mean comfort zone band of ±3.5 �C. Tdew is calcu-lated from the vapor pressure e by Eq. (20) [51].

TdewðzÞ ¼243:5 ln eðzÞ

6:112

17:6� ln eðzÞ6:112

ð20Þ

The unit of e is in millibars. The relationship between vaporpressure e and specific humidity q is described in Eq. (21).

eðzÞ ¼ qðzÞPðzÞeþ 0:378qðzÞ ð21Þ

where e ¼ 0:622, which is the ratio of the molecular weight of water(MWH20 ¼ 18) to an average molecular weight for air(MWair ¼ 28:9). The maximum allowable indoor air velocity,uin;max, is chosen at 0.8 m/s, according to ASHRAE [52]. The corre-sponding outdoor wind velocity, uout , is calculated based on anempirical relationship (Eq. (22)) developed by Phaff et al. [53],which takes into account the combined effect of wind, temperature,and turbulence on natural ventilation.

Fig. 4. Flow chart of estimating v

uin ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiC1u2

out þ C2hDT þ C3

qð22Þ

where h is the vertical height of the opening, DT is the differencebetween the outside and the average indoor temperature, whichis estimated using the difference between Tup and the lower limitof outdoor temperature for natural ventilation, C1 is the wind speedcoefficient, C2 is the buoyance coefficient, and C3 is the turbulencecoefficient. Their values are determined from the study by Phaffet al. [53], in which C1 = 0.001, C2 = 0.0035, and C3 = 0.01. In addi-tion, the other threshold for maximum allowable ambient windspeed is 13.8 m/s, according to the classification of the Beaufortwind scale of 6, and based on a study on wind-induced noise ofhigh-rise buildings [54,55].

3. Results and discussion

The vertical profiles of wind speed u, temperature T, and specifichumidity q are critical to the estimation of NV hours. Here, anexample of the vertical profiles generated by the ABL meteorologymodel is presented. The temporal variation of u, T, and q at 1 AM, 8AM, 1 PM and 7 PM on the day of the summer solstice are shown inFig. 5. The diurnal cycle of heating and cooling on ground surfacehas a considerable effect on the stability and vertical structure ofABL. In the daytime, the surface is heated by solar radiation aftersunrise, which enhances the vertical exchange of momentum, heat,and moisture, and creates a mixed layer with nearly constant dis-tribution of u, h, and q above the surface layer (Fig. 5b and c). Eva-poration from the surface adds moisture to the near-surface region,resulting in a rapid attenuation of specific humidity with height(Fig. 5b and c). At night, the ground surface cools down due to longwave radiation. A nocturnal inversion layer near the surface isoften formed in which the temperature increases with height(Fig. 5a). In addition, the near-ground wind speed increases rapidlywith height due to the strong wind shear and the lack of verticalmixing (Fig. 5a and d). At night, although humidity remains nearlyconstant in most of the middle and top portions of the boundarylayer due to diminished turbulence, the radiative cooling at groundsurface causes dew or frost formation, which reduces humidity inthe bottom of the boundary layer (Fig. 5a and d).

Fig. 6 presents the vertical profiles of NV hours at six studiedcities. Los Angeles, which is located in Zone 3, displays significantlylarger NV potential than the remainder of the cities, out of a maxi-

ertical profiles of NV hours.

Fig. 5. Vertical profiles of wind speed [m/s], absolute temperature [K], and specific humidity [g/kg] at different time on the day of summer solstice in Los Angeles.

282 Z. Tong et al. / Applied Energy 193 (2017) 276–286

mum of 8760 NV hours. The temperate and dry climate in SouthernCalifornia is ideal for natural ventilation. The humidity is relativelylow throughout the year. Based on our analysis, only 11 out of8760 hours exceed the humidity threshold at Tdew ¼ 17 �C at groundlevel. Combined with thresholds for temperature and wind speed,there are still 7258 NV hours at ground level, accounting for 83% ofthe year. In general, the number of NV hours is reduced with heightabove the ground due to the decrease in absolute temperature, andan increase in wind speed. The vertical gradient of NV hours in LosAngeles is steep relative to other cities. At z = 300 m,NVhours dropsto 4866 from 7258 hours. Since the near-ground temperatures for asignificant portionof the year are only a fewdegrees greater than thelower NV threshold, the number of NV hours drops considerably asthe absolute temperature decreases with height.

Following Los Angeles, NYC offers the second largest NV poten-tial due to its favorable climate for utilizing natural ventilationamong studied cities. The city features a warm temperate climatewith hot and moist summers, and cold winters with temperaturesoften below freezing. The annual ground-level NV hour is 3360.The NV hour at z = 300 m decreases to 2464 due to sharp temper-ature drop with altitude. The remainder of the four studied citiesdisplay comparable numbers of NV hours, approximately 2500 atground level. Miami, which is located in Zone 1, has hot humidsummers, and short warm winters. The annual ground-level NVhour is 2552, and the vertical distribution of NV hours is nearlyuniform. In the Midwest, both Chicago from Zone 5 and Min-neapolis from Zone 6 demonstrate considerable negative verticalgradient of NV hour as a result of increasing outdoor wind speedwith height above ground level, which gradually exceeds the windspeed threshold.

Three important meteorological parameters that greatly affectNV hours are revealed. Fig. 7 shows the breakdown of the

individual effect of wind speed (u), temperature (T), and humidity(Tdew) on NV hour calculation. In Miami, the effect of humidity isdominant over temperature and wind speed. Nearly 6000 NV hoursare lost due to the humidity effect alone (Fig. 7a). The loss due totemperature is much less, about 150 hours, depending on the alti-tude. The effect of wind speed increases with altitude. It starts toplay a role in calculating NV hours at roughly 120 m above theground level. In comparison, NV potential is dominated by theeffect of temperature in Minneapolis (Fig. 7b). Roughly 5000 NVhours are lost due to unfavorable temperature. The role of humid-ity on NV potentials is much less significant compared to that inZone 1 and 2 (Miami and Houston). The influence of wind speedgenerally determines the vertical profile of NV hours in Minneapo-lis. Unlike Miami, the loss in NV hours from wind speed starts nearthe ground level and increases considerably with height above theground level as a result of the relatively windy climate in the Mid-west of U.S.

Fig. 8 displays the seasonal variations in vertical NV hour pro-files at each studied city. Unsurprisingly, natural ventilation canbe achieved rarely in the summer in cities with hot and humid cli-mate (Zone 1 and 2), such as Miami and Houston. In contrast,spring, winter, and later autumn are periods of the year whenthe weather is generally favorable for natural ventilation, and a sig-nificant amount of NV hours are available in Miami and Houston.The vertical profiles of NV hours are nearly uniform except in thewinter of Houston. A noticeable negative vertical gradient (i.e.,decreasing NV hours with increasing altitude) occurs, becausehumidity no longer plays a dominant role in determining NV hours(Tdew < 17 �C for the most of time). Instead, the vertical profiles aremainly shaped by the temperature and wind speed thresholds. Asopposed to the climate in Miami, the near-ground temperatures inHouston are in the vicinity of the lower NV threshold of

Fig. 6. Vertical profiles of NV hour (out of 8760 hours per year) from ground level to 300 m above the ground at studied cities.

Fig. 7. Breakdown of the NV hours at humidity-dominant climate and temperature-dominant climate. Tdew represents the effect of humidity, T represents the effect oftemperature, and u represents the effect of wind speed.

Z. Tong et al. / Applied Energy 193 (2017) 276–286 283

temperature (12.8 �C). Therefore, increasing the altitude decreasesthe absolute temperature and available NV hours.

On the contrary, in the winter, cities in the north of the U.S. suchas NYC, Chicago, and Minneapolis display almost zero potential for

natural ventilation (Fig. 8) due mostly to the unfavorable outdoortemperature that creates considerable heat loss and unpleasantthermal comfort in buildings. A significant amount of NV hourswith generally uniform vertical distribution are seen in the

Fig. 8. Seasonal variations of vertical NV hour profiles (out of 8760 hours per year) at six studied cities.

284 Z. Tong et al. / Applied Energy 193 (2017) 276–286

summer, suggesting that there are enormous opportunities toreduce the cooling energy load if buildings are properly designed.Los Angeles, which has the ideal climate for natural ventilation,shows the largest amount of NV hours and significant negative gra-dients of NV hours throughout the year. This can be explained bythe relatively moderate variations in temperature and humidityamong different seasons.

4. Conclusion

As architects have already begun designing high-rise buildingswith sustainable strategies, the utilization of natural ventilationthrough mixed-mode ventilation systems is critical for reducingenergy consumption in buildings and maintaining a healthy indoorenvironment. Given that HVAC systems consume about 30–40% ofthe overall building energy consumption, reducing the size ofmechanical systems with natural ventilation becomes a promisingmeans to achieve sustainability in high-rise buildings. Due to thecomplex structure of ABL, the elevated thermal environment thatincludes temperature, wind speed, and humidity, varies greatlywith altitude for each floor, which results in non-uniform vertical

profiles of NV hours. It is therefore essential to incorporate bound-ary layer meteorology in the evaluation of NV potential. In thisstudy, we have provided the early effort to understand and esti-mate the vertical NV hour profiles for major U.S. cities from six cli-mate zones, using an in-house ABL meteorology model. Asdifferent climates pose different environmental challenges, variousNV strategies can be designed accordingly to achieve desirablelevel of thermal comfort.

The diurnal cycle of ABL and local climate characteristics arefound to have a great effect on the vertical structure of NV poten-tial. In general, negative vertical gradients of NV hours areobserved for all studied cities except Miami where the vertical dis-tribution is nearly uniform. For example, the annual NV hourdecreases from 7258 at ground level to 4866 at 300 m above theground in Los Angeles. Outdoor temperature is a key parameterthat determines vertical profiles of NV hours in Los Angeles, NYC,Chicago, and Minneapolis. On the other hand, humidity plays agreater role in Miami and Houston, where the outdoor temperatureis often suitable for natural ventilation except in the summer.Among six studied cities, Los Angeles provides the ideal climate(warm and dry) for using natural ventilation, and displays the

Z. Tong et al. / Applied Energy 193 (2017) 276–286 285

greatest NV potential (7258 NV hours and 83% of the year atground level) among studied cities, followed by NYC with 3360NV hours. The remainder of the four studied cities display compa-rable numbers of NV hours, with approximately 2500 at groundlevel.

Our study investigated the key elements affecting natural ven-tilation in high-rise buildings, providing a valuable reference toarchitects and engineers on ventilation design and the verticalarrangement of floor functionalities. Although site-specific charac-teristics, such as building structure and surroundings, are notdirectly taken into account, the ABL meteorology model in thisstudy can be easily coupled with building-resolved CFD simulationat urban neighborhood scale through inlet boundary conditions toreduce modeling uncertainties. The computing cost of such cou-pled simulation is high, but it is suitable for site-specific case stud-ies where the building structures near the location of interest aregeometrically complex and building-to-building interactions mustbe considered. The methodology and findings from this study areintended to assist architects and policy makers in quantifying thepotential energy savings of natural ventilation, and illustratingthe importance of considering the vertical variations of elevatedthermal environment in high-rise buildings across different cli-mate zones at the early design stage.

Acknowledgment

Z.T. and Y.C. are grateful for the postdoctoral fellowship fromthe Center for Green Buildings and Cities (CGBC) at HarvardUniversity.

References

[1] Wood A, Salib R. Natural ventilation in high-rise office buildings (CTBUHTechnical Guides); 2013.

[2] Cao B, Zhu Y, Li M, Ouyang Q. Individual and district heating: a comparison ofresidential heating modes with an analysis of adaptive thermal comfort.Energy Build 2014;78:17–24.

[3] Chen Y, Samuelson HW, Tong Z. Integrated design workflow and a new tool forurban rainwater management. J Environ Manage 2016;180:45–51.

[4] Chou SK, Chua KJ, Ho JC, Ooi CL. On the study of an energy-efficient greenhousefor heating, cooling and dehumidification applications. Appl Energy2004;77:355–73.

[5] Chua KJ, Chou SK, Yang WM, Yan J. Achieving better energy-efficient airconditioning – a review of technologies and strategies. Appl Energy2013;104:87–104.

[6] Cui C, Wu T, Hu M, Weir JD, Li X. Short-term building energy modelrecommendation system: a meta-learning approach. Appl Energy2016;172:251–63.

[7] Hua Y, Göçer Ö, Göçer K. Spatial mapping of occupant satisfaction and indoorenvironment quality in a LEED platinum campus building. Build Environ2014;79:124–37.

[8] Li X, Wen J. Building energy consumption on-line forecasting using physicsbased system identification. Energy Build 2014;82:1–12.

[9] Li X, Wen J, Bai E-W. Developing a whole building cooling energy forecastingmodel for on-line operation optimization using proactive systemidentification. Appl Energy 2016;164:69–88.

[10] Malkawi A, Yan B, Chen Y, Tong Z. Predicting thermal and energy performanceof mixed-mode ventilation using an integrated simulation approach. BuildSimul 2016;9:335–46.

[11] Oh MS, Ahn JH, Kim DW, Jang DS, Kim Y. Thermal comfort and energy saving ina vehicle compartment using a localized air-conditioning system. Appl Energy2014;133:14–21.

[12] Ramponi R, Angelotti A, Blocken B. Energy saving potential of nightventilation: sensitivity to pressure coefficients for different Europeanclimates. Appl Energy 2014;123:185–95.

[13] Tong Z, Chen Y, Malkawi A, Adamkiewicz G, Spengler JD. Quantifying theimpact of traffic-related air pollution on the indoor air quality of a naturallyventilated building. Environ Int 2016;89–90:138–46.

[14] Tong Z, Whitlow TH, Landers A, Flanner B. A case study of air quality above anurban roof top vegetable farm. Environ Pollut 2016; 208, Part A:256-260.

[15] van Hooff T, Blocken B, Tominaga Y. On the accuracy of CFD simulations ofcross-ventilation flows for a generic isolated building: comparison of RANS,LES and experiments. Building Environ 2017;114:148–65.

[16] Yan J, Chou S-K, Desideri U, Lee D-J. Transition of clean energy systems andtechnologies towards a sustainable future (Part II). Appl Energy2016;162:1109–13.

[17] Hua Y, Oswald A, Yang X. Effectiveness of daylighting design and occupantvisual satisfaction in a LEED Gold laboratory building. Build Environ2011;46:54–64.

[18] Cao B, Luo M, Li M, Zhu Y. Too cold or too warm? A winter thermal comfortstudy in different climate zones in China. Energy Build 2016;133:469–77.

[19] Zhang S, Wu Y, Hu J, Huang R, Zhou Y, Bao X, et al. Can Euro V heavy-dutydiesel engines, diesel hybrid and alternative fuel technologies mitigate NOX

emissions? New evidence from on-road tests of buses in China. Appl Energy2014;132:118–26.

[20] Tong Z, Whitlow TH, MacRae PF, Landers AJ, Harada Y. Quantifying the effect ofvegetation on near-road air quality using brief campaigns. Environ Pollut2015;201:141–9.

[21] Luo M, Cao B, Damiens J, Lin B, Zhu Y. Evaluating thermal comfort in mixed-mode buildings: a field study in a subtropical climate. Build Environ2015;88:46–54.

[22] Oropeza-Perez I, Østergaard PA. Energy saving potential of utilizing naturalventilation under warm conditions – a case study of Mexico. Appl Energy2014;130:20–32.

[23] Tong Z, Chen Y, Malkawi A, Liu Z, Freeman RB. Energy saving potential ofnatural ventilation in China: the impact of ambient air pollution. Appl Energy2016;179:660–8.

[24] Tong Z, Chen Y, Malkawi A. Defining the influence region in neighborhood-scale CFD simulations for natural ventilation design. Appl Energy2016;182:625–33.

[25] Zhou C, Wang Z, Chen Q, Jiang Y, Pei J. Design optimization and fielddemonstration of natural ventilation for high-rise residential buildings. EnergyBuildings 2014;82:457–65.

[26] Pasquay T. Natural ventilation in high-rise buildings with double facades,saving or waste of energy. Energy Buildings 2004;36:381–9.

[27] Niu J. Some significant environmental issues in high-rise residential buildingdesign in urban areas. Energy Buildings 2004;36:1259–63.

[28] Prajongsan P, Sharples S. Enhancing natural ventilation, thermal comfort andenergy savings in high-rise residential buildings in Bangkok through the use ofventilation shafts. Building Environ 2012;50:104–13.

[29] Priyadarsini R, Cheong KW, Wong NH. Enhancement of natural ventilation inhigh-rise residential buildings using stack system. Energy Buildings2004;36:61–71.

[30] Council IC. International energy conservation code; 2009.[31] Epa US. AERSURFACE User’s Guide. NC: US Environmental Protection Agency

Research Triangle Park; 2008.[32] Businger JA, Wyngaard JC, Izumi Y, Bradley EF. Flux-profile relationships in the

atmospheric surface layer. J Atmos Sci 1971;28:181–9.[33] Garratt JR, Wyngaard JC, Francey RJ. Winds in the atmospheric boundary layer-

prediction and observation. J Atmos Sci 1982;39:1307–16.[34] Tong Z, Yang B, Hopke PK, Zhang KM. Microenvironmental air quality impact

of a commercial-scale biomass heating system. Environ Pollut2017;220:1112–20. Part B.

[35] Tong Z, Zhang KM. The near-source impacts of diesel backup generators inurban environments. Atmos Environ 2015;109:262–71.

[36] Wang YJ, Nguyen MT, Steffens JT, Tong Z, Wang Y, Hopke PK, et al. Modelingmulti-scale aerosol dynamics and micro-environmental air quality near a largehighway intersection using the CTAG model. Sci Total Environ2013;443:375–86.

[37] Tong Z, Baldauf RW, Isakov V, Deshmukh P, Max Zhang K. Roadside vegetationbarrier designs to mitigate near-road air pollution impacts. Sci Total Environ2016;541:920–7.

[38] Perry SG. CTDMPLUS: a dispersion model for sources near complextopography. Part I: technical formulations. J Appl Meteorol 1992;31:633–45.

[39] Hanna SR, Chang JC. Hybrid plume dispersion model (HPDM) improvementsand testing at three field sites. Atmos Environt. Part A. General Topics1993;27:1491–508.

[40] Garratt J. The atmospheric boundary layer. In: Cambridge atmospheric andspace science series. Cambridge 416: Cambridge University Press; 1992. 444.

[41] Oke TR. Boundary layer climates. Routledge; 2002.[42] Arya PS. Introduction to micrometeorology. Academic Press; 2001.[43] Van Ulden A, Holtslag A. Estimation of atmospheric boundary layer parameters

for diffusion applications. J Climate Appl Meteorol 1985;24:1196–207.[44] Carson DJ. The development of a dry inversion-capped convectively unstable

boundary layer. Quarterly J R Meteorol Soc 1973;99:450–67.[45] Zilitinkevich SS. On the determination of the height of the Ekman boundary

layer. Boundary-Layer Meteorol 1972;3:141–5.[46] Brost RA, Wyngaard JC. A model study of the stably stratified planetary

boundary layer. J Atmos Sci 1978;35:1427–40.[47] Stull RB. An introduction to boundary layer meteorology. Springer Science &

Business Media; 2012.[48] de Dear RJ, Brager GS. Thermal comfort in naturally ventilated buildings:

revisions to ASHRAE Standard 55. Energy Build 2002;34:549–61.[49] ASHRAE, 2009. ASHRAE handbook - fundamentals.[50] Hiyama K, Glicksman L. Preliminary design method for naturally ventilated

buildings using target air change rate and natural ventilation potential maps inthe United States. Energy 2015;89:655–66.

[51] Bolton D. The computation of equivalent potential temperature. Mon WeatherRev 1980;108:1046–53.

[52] ASHRAE, 2010. Standard 55: Thermal environmental conditions for humanoccupancy.

286 Z. Tong et al. / Applied Energy 193 (2017) 276–286

[53] Phaff J, de Gids W, Ton J, van der Ree D, Schijndel L. The ventilation ofbuildings: investigation of the consequences of opening one window on theinternal climate of a room, Report C 448. TNO Institute for EnvironmentalHygiene and Health Technology (IMG-TNO), Delft, The Netherlands; 1980.

[54] Ploemen J, Nijs L, Pleysier J, Schipper H. Wind-induced sound on buildings andstructures. In: Proceedings of the 13th international conference on wind

engineering, ICWE 13, Amsterdam, 10–15 July 2011. Multi-Science Publishing;2011.

[55] Stewart RH. Introduction to physical oceanography. Texas A & M UniversityTexas; 2008.