Content uploaded by D. N. Asimakopoulos
Author content
All content in this area was uploaded by D. N. Asimakopoulos on May 09, 2018
Content may be subject to copyright.
Solar Energy Vol. 70, No. 3, pp. 201–216, 2001
2001 Elsevier Science Ltd
Pergamon PII: S0038–092X(00)00095–5 All rights reserved. Printed in Great Britain
0038-092X/01 /$ - see front matter
www.elsevier.com/locate/ solener
ON THE IMPACT OF URBAN CLIMATE ON THE ENERGY CONSUMPTION
OF BUILDINGS
,
†
M. SANTAMOURIS* , N. PAPANIKOLAOU*, I. LIVADA*, I. KORONAKIS*,
,
C. GEORGAKIS*, A. ARGIRIOU** and D. N. ASSIMAKOPOULOS* **
*Group of Building Environmental Studies, Section of Applied Physics, Physics Department,
University of Athens, Building of Physics - 5, 157 84, University Campus, Athens, Greece
**National Observatory of Athens, Institute of Meteorology and Physics of the Atmospheric Environment,
Lofos Nimfon, Athens, Greece
Abstract—Climatic measurements from almost 30 urban and suburban stations as well as specific measure-
ments performed in 10 urban canyons in Athens, Greece, have been used to assess the impact of the urban
climate on the energy consumption of buildings. It is found that for the city of Athens, where the mean heat
island intensity exceeds 108C, the cooling load of urban buildings may be doubled, the peak electricity load for
cooling purposes may be tripled especially for higher set point temperatures, while the minimum COP value of
air conditioners may be decreased up to 25% because of the higher ambient temperatures. During the winter
period, the heating load of centrally located urban buildings is found to be reduced up to 30%. Regarding the
potential of natural ventilation techniques when applied to buildings located in urban canyons, it is found that,
mainly during the day, this is seriously reduced because of the important decrease of the wind speed inside the
canyon. Air flow reduction may be up to 10 times the flow that corresponds to undisturbed ambient wind
conditions. 2001 Elsevier Science Ltd. All rights reserved.
1. INTRODUCTION Almost similar situations are found in Europe.
It is reported that during the 1970s, urban tertiary
Cities are increasingly expanding their boundaries buildings in Greece represented about 14% of
and populations and as stated ‘from the clima- these buildings in the country. Due to the
tological point of view, human history is defined dramatic urbanization, the corresponding percent-
as the history of urbanization’. Increased indus- age of new urban tertiary constructions increased
trialization and urbanization in recent years have up to 55% during the 1980s.
affected dramatically the number of urban build- The situation will be even more dramatic in
ings with major effects on the energy consump- developing countries. Already, 23 of the 34 cities
tion of this sector. It is expected that 700 million with more than 5 million inhabitants are in
people will move to urban areas during the last developing countries. Current projections estimate
decade of this century. The number of urban that 11 of those cities will have populations of
dwellers has risen from 600 million in 1920 to 2 between 20 and 30 million by the year 2000.
billion in 1986 and if this growth continues, more It is clear that urban areas without a high
than one-half of the world’s population will live climatic quality use more energy for air con-
in cities by the end of this century. One hundred ditioning in summer and even more electricity for
years ago, only 14% lived in cities and in 1950, lighting. Moreover, discomfort and inconvenience
less than 30% of the world population was urban. to the urban population due to high temperatures,
Today, at least 170 cities support more than one wind tunnel effects in streets and unusual wind
million inhabitants each. As estimated, in the turbulence due to wrongly designed high rise
United States, 90% of the population is expected buildings is very common (Bitan, 1992).
to be living in, or around, urban areas by the year Data on the energy and specific electricity
2000. Estimations show that urban populations consumption of major European cities are given
will occupy 80% of the total world population in by Eurostat, 1995. Data on the electricity consump-
2100. tion in European cities range from 60 GWh/year
for Valetta to 26,452 GWh/year for London.
According to the same data, the average electrici-
†
ty consumption calculated on the basis of avail-
Author to whom correspondence should be addressed. Tel.: able data for cities with more than 1,000,000
130-1-727-6934; fax: 130-1-729-5282; e-mail:
msantam@atlas.uoa.gr inhabitants is around 4500 GWh per year. How-
201
202 M. Santamouris et al.
ever, these data cannot be used to draw any the materials and their properties. The upper
conclusions. boundary of the urban canopy varies from one
Other, statistical data (Stanners and Bourdeau, spot to another because of the variable heights of
1995) show that the amount of energy consumed the buildings and the wind speed.
by cities for heating and cooling of offices and The air dome layer is defined by Oke (1976),
residential buildings in western and southern as ‘that portion of the planetary boundary layer
Europe has increased significantly in the last two whose characteristics are affected by the presence
decades. A recent analysis, (Jones, 1992), showed of an urban area at its lower boundary’ and is
that a 1% increase in the per capita GNP leads to more homogeneous in its properties over the
an almost equal (1.03), increase in energy con- urban area at large.
sumption. However, as reported, an increase of Temperature distribution in urban areas is
the urban population by 1% increases the energy highly affected by the urban radiation balance.
consumption by 2.2%, i.e. the rate of change in Solar radiation incident on the urban surfaces is
energy use is twice the rate of change in urbaniza- absorbed and then transformed to sensible heat.
tion. These data show clearly the impact that Most of the solar radiation impinges on roofs, and
urbanization may have on energy use. the vertical walls of the buildings, and only a
Thus, it becomes increasingly important to relatively small part reaches the ground level.
study urban climatic environments and to apply Walls, roofs and the ground emit long wave
this knowledge to improve people’s environment radiation to the sky. The intensity of the emitted
and decrease the energy consumption in cities. radiation depends on the view factor of the
The present paper presents the results of an surface regarding the sky. Under urban conditions
urban study carried out in Athens aiming, among most of the sky dome viewed by walls and
other things, to investigate the impact of the urban surfaces is blocked by other buildings, and thus
climate on the energy consumption of urban the long wave radiant exchange does not really
buildings. An extended network of measuring result in significant losses.
stations has been installed and climatic data The net balance between the solar gains and the
collected for a 3-year period. Specific air flow and heat loss by emitted long wave radiation de-
temperature distribution experiments have been termines the thermal balance of urban areas.
also carried out in 10 urban canyons. The data Because the radiant heat loss is slower in urban
have been used to evaluate the impact of in- areas the net balance is more positive than in the
creased ambient temperatures on the heating and surrounding rural areas and thus higher tempera-
cooling performance of buildings. Also, air flow tures are presented.
and temperature distribution data in urban Air temperatures in densely built urban areas
canyons have been used to evaluate the impact of are higher than the temperatures of the surround-
canyon geometry and characteristics on the po- ing rural country. The phenomenon known as
tential of natural ventilation techniques to provide ‘heat island’, is due to many factors the more
passive cooling to urban buildings. important of which are summarized by Oke et al
.
(1991).
•Canyon Radiative Geometry contributes to
2. URBAN CLIMATE AND ENERGY decrease the long wave radiation loss from
CONSUMPTION OF BUILDINGS within street canyon due to the complex ex-
A very comprehensive description of the urban change between buildings and the screening of
climate is presented by Landsberg (1981). It is the skyline.
important to describe the mean features by which •Thermal properties of materials that may in-
the urban climate differs from the climatic con- crease storage of sensible heat in the fabric of
ditions of the surrounding rural areas. the city.
According to Oke (1977), the air space above a •Anthropogenic heat released from combustion
city can be divided into the so-called urban air of fuels and animal metabolism.
‘canopy’, and the boundary layer over the city •Urban greenhouse, that contributes to increase
space called ‘the urban air dome’. The urban air the incoming long wave radiation from the
canopy is the space bounded by the urban build- polluted and warmer urban atmosphere.
ings up to their roofs. The specific climatic •Canyon radiative geometry decreasing the
conditions at any given point within the canopy effective albedo of the system because of the
are determined by the nature of the immediate multiple reflection of short wave radiation
surroundings and in particular of the geometry, between the canyon surfaces.
On the impact of urban climate on the energy consumption of buildings 203
•Reduction of evaporating surfaces in the city may offer a more complete view of the impact of
putting more energy into sensible and less into urban temperatures on the energy load of build-
latent heat. ings.
•Reduced turbulent transfer of heat from within Various studies correlating urban heat island
streets. and energy consumption have been carried out in
The heat island phenomenon may occur during the United States. As reported, (Akbari et al
.
,
the day or the night period. The intensity of the 1992), for US cities with population larger than
heat island is mainly determined by the thermal 100,000 the peak electricity load will increase 1.5
balance of the urban region and can result in up to to 2% for every 18F increase in temperature.
10 degrees of temperature difference. Taking into account that urban temperatures
Higher urban temperatures have a serious im- during summer afternoons in the US have in-
pact on the electricity demand for air conditioning creased by 2 to 48F during the last 40 years, it can
of buildings, increase smog production, while be assumed that 3 to 8% of the current urban
contributing to increased emission of pollutants electricity demand is used to compensate for the
from power plants, including sulfur dioxide, heat island effect alone.
carbon monoxide, nitrous oxides and suspended Comparisons of high ambient temperatures to
particulates. utility loads for the Los Angeles area have shown
Unfortunately, very few studies have been that an important correlation exists. It is found
carried out on the impact of the urban climate to that the net rate of increase of the electricity
the energy consumption of urban buildings for demand is almost 300 MW per 8F. Taking into
heating and cooling purposes. Existing studies account that there is a 8F increase of the peak
either correlate increased urban temperatures and temperature in Los Angeles since 1940, this is trans-
the corresponding electricity demand for selected lated into an added electricity demand of 1.5 GW
utility districts or either use sets of local tempera- due to heat island effect.
ture data to calculate the breakdown of the Similar correlation between temperatures and
cooling and heating load in a city suffering from electricity demand have been established for
increased temperatures. selected utility districts in the USA. The rate of
Both methodologies and techniques present increase of the electricity demand in MW/8F, as
important advantages. When correlations between well as the increase of the electricity demand, in
temperatures and energy use are established by percentage, for each utility, has been calculated
comparing utility-wide electricity loads to tem- for selected cities (Akbari et al
.
, 1992). Based on
peratures at the same time of the day, a very clear the above increase rates, it has been calculated
picture on the real impact of high urban tempera- that for the USA the electricity costs for summer
tures is established. However, to achieve it, it is heat island alone could be as much as $1 million
necessary to minimize the non-climate-related per hour, or over $1 billion per year (Akbari et
effects on the electricity demand, which is not al
.
, 1992). Computer studies have shown for the
always possible, and when possible is not always whole country the possible increase of the peak
accurate. This technique, although giving an cooling electricity load due to heat island effect
estimation of the increase of the energy consump- could range from 0.5 to 3% for each 18F rise in
tion in an integrated way, does not permit to temperature.
investigate local effects and the impact of the Heat island studies in Singapore reported by
specific urban layout and characteristics to the Tso (1994), show a possible increase of the urban
energy consumption of the buildings. temperature close to 18C. According to the reports
When temporally extended data of the tempera- if there were to be similar changes in tempera-
ture breakdown in a city are used to calculate tures 50 years from now, the anticipated increase
either the energy load of a reference building in building energy consumption, mainly in air
located in a city, or the distribution of the energy conditioning, is of the order of 33 GWh per
consumption in a city, very useful information on annum for the whole island.
the relative energy consumption of the various Watanabe et al
.
(1990/91), using LANDSAT-5
urban sub-regions having different layout and data have analyzed the land temperature distribu-
climatic characteristics is established. However, tion and the thermal environment of the Tokyo
the overall impact of the high urban temperatures metropolitan area for some days of September
on the global energy consumption of the city is 1973. Based on these data as well as the results of
not possible or it is very difficult to evaluate. It is a field survey by the metropolitan government,
obvious that a combination of both techniques energy consumption distribution maps have been
204 M. Santamouris et al.
prepared. Although no further information is internal power supply and high data storage
provided the overall analysis shows clearly that capacity. In order to protect the instruments from
much higher energy consumptions are calculated solar radiation and rain, white wooden boxes with
for the central Tokyo area. lateral slots were constructed, approximating the
Other studies on the Tokyo area reported by Stevenson screen. The detailed and specific
Ojima (1990/91), show clearly that during the 10 characteristics of the sensors used are given in
years from 1965 to 1975, due to the heat island Santamouris et al
.
(1996).
phenomenon, the cooling load of existing build- As the study had to cover the Athens basin,
ings has increased by 10–20% on average. If it measurement points were selected with the fol-
continues to increase at the same rate, it will make lowing criteria:
more than a 50% increment in 2000. (a) to get information about the boundary
conditions around the basin,
(b) to study densely built areas with heavy
3. EXPERIMENTAL PROCEDURE traffic,
In the frame of the urban climate experiment (c) to study densely built areas with less traffic,
carried out in Athens, 20 automatic temperature (d) to study the conditions in green areas of the
and humidity stations have been installed in the city center,
major Athens area during spring 1996. At a later (e) to study medium density built areas.
phase the number of stations has been increased Therefore an attempt was made to select areas
to 30. The instrumentation used was selected to according to these criteria, that additionally are
satisfy several criteria like acceptable cost, in located on the north–south and east–west axes of
order to cover as many locations as possible, the city. A detailed map showing the relative
satisfactory performance according to the interna- position of the stations is given in Fig. 1. In
tional meteorological standards, low maintenance, general, seven stations have been placed in the
Fig. 1. Relative position of the measuring stations.
On the impact of urban climate on the energy consumption of buildings 205
very central area of Athens, 26 stations are placed the wind speed and direction out of the
in urban areas and in a radial configuration around canyon. Measurements have been performed
Athens while one station is placed in an almost every 12 s.
rural region out of the city in order to be used as
the reference station. 4. TEMPERATURE AND DEGREE DAYS
Measurements are taken on a minute basis DISTRIBUTION
while all sensors are calibrated between them and
against high precision sensors. The collected data have been analyzed in detail
In order to investigate the specific air flow and in order to assess the heat island intensity in the
thermal phenomena in an urban canyon, experi- city of Athens as well as the specific distribution
ments were performed during summer 1997 in 10 of the ambient temperature in the city. Specific
different canyons having dissimilar layouts, and detailed statistical and climatological analyses
orientation, anthropogenic heat and vegetation. have been performed, however, presentation of
Three types of measurements were performed. this analysis is out of the scope of the present
(a) Air temperature measurements. Miniature paper.
ambient air temperature sensors have been During the summer period much higher tem-
used. The sensors were shielded inside a white peratures have been recorded in the central
painted wooden cylinder opened on two parts Athens area especially during the daytime. Fig. 2
to permit air circulation. The length of the plots the relative temperature difference between
cylinders was 20 cm while their internal and 12 urban stations located in and around the center
external diameter was 9 and 8 cm, respective- as a function of the temperature of the urban
ly. Sensors were completely protected from station. Plotted data refer to the summer of 1996.
solar radiation. The cylindrical wooden boxes As shown, temperature increase in the very
including the sensors have been fixed in the central area may reach values up to 158C. Daily
exterior facades of the buildings and in various heat island intensity for most of the central urban
heights in the canyon. The distance between stations is close to 108C. Heat island intensity is
the cylindrical box and the exterior wall was much lower in suburban areas and ranges between
between 5 cm and 2 m, thus the temperature 6 and 28C. During the night period, heat island
sensors were between 12 and 205 cm from the intensity, Fig. 3, varies between 2 and 58Casa
walls. Measurements were performed every 15 function of the station characteristics. Urban green
min. areas present 2–38C lower than the reference
(b) Surface temperature measurements. An station.
infrared thermometer equipped with a laser Cooling degree hours determine to a high
beam has been used. The surface temperature degree the cooling load of buildings (Santamouris
of the exterior facades of the buildings are and Assimakopoulos, 1996), thus knowledge of
measured. Measurements are performed from their spatial and temporal distribution as well as
the bottom to the top of both facades of the of their absolute value in a city is of high interest
canyon using a step of 3–3.5 m. All measure- for designers and climatologists. The spatial dis-
ments have been performed from the street tribution of the cooling degree hours for 13:00 h
level. The pavement and road temperature and for the whole of August 1996 are given in
were measured as well at five different points Fig. 4. Cooling degree hours have been calculated
along the width of the canyon in both sections for a temperature base of 268C. As shown, during
defined above. All measurements have been noon hours, the center of the city presents almost
performed on an hourly basis during day and two times higher cooling degree hours than the
night. city surroundings. In particular, in the central
(c) Wind speed measurements. A three-axis Athens area cooling degree hours are close to 385,
anemometer has been used to measure the while the corresponding value for the reference
three components of the wind speed inside the suburban areas is close to 132.
canyon. The anemometer was mounted on the A very interesting spatial distribution of cool-
exterior facade of a building in the canyon at a ing degree hours has been calculated for the night
distance of 1–2 m from the wall and 25 m period. Fig. 5 shows this distribution for 1:00 h
from the intersection of the canyon with a during the whole of August 1996. As shown,
perpendicular road. A cup anemometer has higher cooling degree hours have been calculated
been also placed on the top of the canyon at a in the Western Athens area than in the central one.
distance of 6 m from its top level to measure In particular while cooling degree hours in the
206 M. Santamouris et al.
Fig. 2. Temperature difference between urban stations and the reference one as a function of the absolute temperature of the
urban station during summer days.
Fig. 3. Temperature difference between urban stations and the reference one as a function of the absolute temperature of the
urban station during summer nights.
On the impact of urban climate on the energy consumption of buildings 207
Fig. 4. Iso-cooling degree hour lines for 268C temperature base, at 13:00 h, for August 1996.
reference area are close to 26, the corresponding degree hours in the central Athens area are about
values for the western and central Athens area are 3000–4000 while in the surrounding areas can be
close to 104 and 85, respectively. This is mainly high as 5000 which correspond to a decrease of
due to the high density of western Athens as well about 40%.
as the lack of green spaces and other cool sinks.
The daily spatial distribution of cooling degree 5. CALCULATION METHODOLOGY
hours in a day in Athens, (August 1996), is given
in Fig. 6. As shown, the central Athens area starts The collected data have been used to calculate
to be heated at about 10:00 h, gets the maximum the distribution of the cooling and heating needs
difference compared to the surrounding area at of a representative office building for all locations
about 14:00 to 15:00 h, while the phenomenon is where climatic data were available.
amortized at about 19:00 h. The considered building is constructed in seven
2
During the winter period, the intensity of heat different levels, and has a total surface of 500 m .
island in the central area is not significantly It is used by 25 people and is a low energy
reduced compared to the summer period. During building involving many energy conservation
the day period, Fig. 7, heat mean island intensity features to decrease its heating and cooling needs,
in the central Athens area is close to 108C, while Fig. 10. A full description of the building is given
it is reduced to 3–68C in the surrounding subur- in Allard (1998). The building was monitored for
ban areas. During the night period, Fig. 8, heat about 2 years while specific experiments involv-
island intensity is up to 108C, and varies as a ing tracer gas injections were carried out. To
function of the characteristics of the area. study the performance of the building and to
Heating degree hours determine to a high analyze in detail the obtained experimental data, a
degree the heating load of a building. An indica- theoretical model of the building has been created
tive spatial distribution for December 1997, base in TRNSYS (Solar Energy Laboratory, 1997)
18.38C, is given in Fig. 9. As shown, heating simulation tool. Comparison of experimental with
208 M. Santamouris et al.
Fig. 5. Iso-cooling degree hour lines for 268C temperature base, at 1:00 h, for August 1996.
Fig. 6. Spatial and hourly distribution of the cooling degree hours in Athens for August 1996.
On the impact of urban climate on the energy consumption of buildings 209
Fig. 7. Temperature difference between urban stations and the reference one as a function of the absolute temperature of the
urban station during winter days.
the calculated performance of the building show placed locally. Fig. 11 gives the iso-cooling load
an excellent agreement between the two sets of lines, (in kWh per square metre and month),
data (Geros et al
.
, 1998). indicating the spatial variation of the cooling load
Using hourly data of the ambient temperature for the whole region of Athens. As shown, the
collected at all urban climate stations, simulations cooling load at the center is about double that in
of the cooling and heating load of the reference the surrounding Athens region. The maximum
building have been performed for the whole cooling load always corresponds to the very
period. The same solar radiation data have been central area of Athens and especially to a station
used for all considered stations as a non-signifi- very close to a high traffic road. Minimum values
cant spatial variation of solar radiation has been were calculated in the south east Athens region, a
observed in Athens (Psioglou, 1997). All other mean density residential area close to the
operational data, like internal gains, have been Hemetus forest. Much higher cooling loads have
selected to correspond exactly to the measured been calculated for the Western Athens region.
conditions. Thus, the spatial distribution of the This area is characterized by high density plots,
heating and cooling load of the reference building lack of green spaces, important industrial activity
has been calculated. and higher traffic than the Eastern Athens region.
Apart from increased energy loads for cooling
of buildings, high ambient temperatures increase
6. IMPACT ON THE SUMMER peak electricity loads and put a serious stress on
PERFORMANCE OF URBAN BUILDINGS the local utilities. Thus, knowledge of the possible
The calculated spatial variation of the cooling increase of the peak electricity load due to higher
load of the reference building for a set point of urban ambient temperatures may be very im-
278C and for August 1996 is given in Fig. 11. portant
Values are in kWh per square metre and month. Using TRNSYS, the instant peak cooling load
Very local phenomena and conditions are not of the reference building has been calculated for
shown in the maps unless a measuring station was August 1996 and for various set point tempera-
210 M. Santamouris et al.
Fig. 8. Temperature difference between urban stations and the reference one as a function of the absolute temperature of the
urban station during winter nights.
Fig. 9. Spatial distribution of the heating degree days in Athens for December 1997.
On the impact of urban climate on the energy consumption of buildings 211
Fig. 10. Plan of the reference building.
tures ranging between 26 and 288C. The obtained 27.5 kWwhile the minimum one is close to 13.7 kW.
spatial variation of the peak cooling load for the Thus, the impact of higher urban temperatures is
whole Athens region, and for a set point of 268C extremely important and almost doubles the peak
is given in Fig. 12.Values are in kW. It should be cooling load of the reference building.
noticed that the reported peak cooling loads refer When the set point temperature is 288C, much
to the whole month and do not occur during the higher differences have been found. In this case
same time in all stations. As expected, much the maximum cooling load is close to 23.5 kW
higher peak cooling loads have been calculated while the minimum one is close to 7.3 kW. Thus,
for the central Athens area. For a set point while the maximum peak cooling load is reduced
temperature equal to 268C, the highest peak load to about 4.3 kW, the minimum peak load is
of the reference building is calculated close to reduced by 6.4 kW. The results show that for the
Fig. 11. Distribution of the cooling load in the city of Athens for a set point temperature of 278C and for August 1996.Values are
2
in kWh/ m .
212 M. Santamouris et al.
Fig. 12. Spatial variation of the peak cooling load of a reference building in Athens during August 1996 and 268C set point
temperature. Values are in kW.
central Athens area the peak cooling load is 102% for the South East area of Athens. Results
mainly due to persisting very high ambient tem- show clearly that except for high cooling loads
peratures and is not sensitive to the change of the and peak electricity problems, the heat island
set point temperature. On the contrary, the calcu- effect reduces significantly (to about 25%) the
lated minimum peak cooling load changes efficiency of the air conditioning systems and thus
dramatically as the set point temperature in- may oblige designers to increase the size of the
creases. This is mainly due to the minimization of installed A/C systems and thus intensify peak
the load induced by the indoor–outdoor tempera- electricity problems and energy consumption for
ture differences as the set point increases. In this cooling purposes.
case the calculated peak cooling load is mainly
due to solar radiation and the other sources. 7. IMPACT ON THE WINTER PERFORMANCE
High ambient temperatures have a very serious OF URBAN BUILDINGS
impact on the efficiency of conventional air
conditioners. The coefficient of performance Increased urban temperatures may have a seri-
(COP) is directly affected by relative humidity ous impact on the heating load of urban buildings.
and ambient temperature, and thus it is of interest Calculations of the heating load of the same
to investigate a possible decrease of the COP due reference building have been performed for all
to heat island effect. stations and the whole monitoring period. All
Using hourly temperature and humidity data stations have been grouped in three cluster sta-
from all stations in and around Athens, for the tions located in the very central Athens area,
whole summer 1996, the distribution of the COP suburban stations and urban parks and green
value of a conventional A/C system has been areas. The corresponding heating load for each
calculated. The minimum value for each station station is given in Fig. 14. The heating load in the
has been calculated. Fig. 13 gives the spatial central Athens region was estimated close to
2
distribution of the minimum COP values in 3.7 kWh/ m /month, while the corresponding
Athens. As shown, the absolute minimum COP load of the suburban stations was close to
2
values correspond to the very central area of 5.1 kWh/m /month, and the mean load of the
Athens (close to 75%), because of the high stations located in green areas was close to 7.3
2
ambient temperatures as well as for the coastal kWh/ m /month.
area because of the high humidity. The highest of As shown, increased urban temperatures de-
the minimum COP value is calculated close to crease the heating load of urban temperatures to
On the impact of urban climate on the energy consumption of buildings 213
Fig. 13. Spatial distribution of the minimum COP of a conventional air conditioner in Athens (values in percentage).
Fig. 14. Heating load of the reference office building calculated using data from various urban and suburban stations as well as
2
from stations located in urban areas. Values are in kWh/ m .
214 M. Santamouris et al.
about 30% while the maximum difference be- having a window of 1.5 31.5 m, in each canyon
tween suburban and urban stations was close to facade is also considered.
55%. Two types of simulations have been performed
for each configuration. The first was based on the
wind and temperature data measured inside the
8. IMPACT ON THE NATURAL VENTILATION canyon, while the second one was based on the
POTENTIAL undisturbed temperature and wind speed mea-
Natural ventilation of buildings located in sured over the buildings. Comparison of both
urban canyons is seriously reduced because of the simulation results should permit to assess the
important decrease of the wind velocity inside the decrease of the natural ventilation potential in
canyons. Air flow phenomena associated with urban canyons.
urban canyons are extensively discussed in San- Simulations have been performed using the
tamouris (1999). Experiments in 10 deep canyons AIOLOS natural ventilation simulation code (Al-
during the summer 1997, have shown that mean lard, 1998). The software used is well validated in
wind speed inside the canyon rarely exceeds the frame of the PASCOOL research project
1 m/s, independent of the free wind speed above against a high number of experiments (Limam et
the buildings. Fig. 15 shows, as an example, the al
.
, 1997).
variation of the air speed inside and outside a Figs. 16 and 17 give the air flow rate for the 10
canyon having an aspect ratio close to one during canyons, and for the single side and cross ventila-
the whole experimental period. tion configuration, respectively. The two flow
Natural ventilation of buildings is due either to rates, one when the ambient temperature and wind
the wind forces or to the temperature difference speed is used, and the second corresponding to the
between the indoor and outdoor environment or in inside canyon measured data, are given. An
a combination of both. Design of urban buildings analysis of the results permits the following
to improve natural ventilation potential should conclusions to be extracted.
consider the appropriate wind data and not routine (a) During the day, when the ambient wind
meteorological observations collected in open speed is considerably higher than wind speed
fields. Also, the specific temperature regime in a inside the canyon and inertia phenomena domi-
canyon should be considered. nate the gravitational forces, the natural venti-
In order to evaluate the natural ventilation lation potential in single and cross ventilation
potential of urban buildings, as well as its possible configurations is seriously decreased inside the
decrease because of the canyon related phenom- canyon. In practice this happens when the
ena, simulations of the air flow processes have ambient wind speed is higher than 4 m/s. For
been carried out for 10 different canyons where single side ventilation configurations the air
wind speed and temperature data have been flow is reduced up to five times, while in cross
collected. Two configurations have been consid- ventilation configurations the flow is some-
ered: single as well as a cross ventilation configu- times reduced up to 10 times.
23
ration. A typical zone of 36 m , and 144 m (b) During the day time and when the ambient
Fig. 15. Measured wind speed inside and outside a representative urban canyon.
On the impact of urban climate on the energy consumption of buildings 215
Fig. 16. Air flow rates calculated for 10 different canyons and for single side building configurations.
Fig. 17. Air flow calculated for 10 different canyons and for cross ventilation configurations.
216 M. Santamouris et al.
Acknowledgements—The present research is partly financed
wind speed is lower than 3–4 m/s, gravitation- by the European Commission, Directorate General for Science,
al forces dominate the air flow processes. In Research and Technology under the contract JOR3-CT95-
this case the difference in wind speed inside 0024. The contribution of the Commission is gratefully
acknowledged.
and outside the canyon, do not play any
important role and especially in single side
configurations. REFERENCES
(c) During the night time the ambient wind Akbari H., Davis S., Dorsano S., Huang J. and Winert S.
speed is seriously decreased and is comparable (1992). Cooling our Communities —A Guidebook on Tree
to the wind speed inside the canyon. In this Planting and Light Colored Surfacing, US Environmental
Protection Agency, Office of Policy Analysis, Climate
case the air flow calculated for inside and Change Division, January.
outside the canyon is almost the same. Allard F. (Ed.) (1998). Natural Ventilation of Buildings, James
(d) The calculated reduction of the air flow and James, London.
Bitan A. (1992) The high climatic quality of city of the future.
inside the canyon is mainly a function of the Atmos
.
Environ
.
26B, 313–329.
wind direction inside the canyon. When the GerosV., Santamouris M., Tsagrassoulis A. and Guaraccino G.
ambient flow is almost vertical to the canyon (1998) Experimental evaluation of night ventilation phe-
nomena. Energy Build
.
, In press.
axis, the flow inside the canyon is almost Jones B. G. (1992) Population growth, urbanization and
vertical and parallel to the window. In this case disaster risk and vulnerability in metropolitan areas: a
a much higher pressure coefficient corresponds conceptual framework. In Environmental Management and
Urban Vulnerability,World Bank Discussion Paper no. 168.
to the conditions outside the canyon, and thus Landsberg H. E. (1981) The urban climate. In International
a much higher flow is calculated when the Geophysics Series, Vol. 28, Academic Press, New York.
ambient conditions are considered and inertia Limam K., Allard F. and Dascalaki E. (1997) Natural ventila-
tion research activities undertaken in the framework of
forces are dominating. When the ambient flow Pascool. Int
.
J
.
Solar Energy 19, 81–119.
is parallel to the canyon axis, a similar flow is Ojima T. (1990/91) Changing Tokyo metropolitan area and its
observed inside the canyon; thus the pressure heat island model. Energy Build
.
15–16, 191–203.
Oke T. R. (1976) The distance between canopy and boundary
coefficients are almost similar. layer urban heat island. Atmosphere 14(4), 268–277.
Oke T. R. (1977) The significance of the atmosphere in
planning human settlements. In Ecological and Classifica-
9. CONCLUSIONS tion in Urban areas, Ecological Land Classification Series,
Vol. 3, Wilkin E. B. and Ironside G. (Eds.), Canadian
The present paper discusses the energy impact Government Publishing Center, Ottawa, Ontario.
of higher ambient temperatures in the city centers. Oke T. R., Johnson D. G., Steyn D. G. and Watson I. D.
(1991) Simulation of surface urban heat island under ideal
Data from the Athens urban climate have been conditions at night – Part 2: Diagnosis and causation.
used to evaluate the increase of the cooling load Boundary Layer Meteorology 56, 339–358.
of urban buildings. It is found that for a repre- Psioglou V. (1997). Development of an Atmospheric Solar
Radiation Model, University of Athens, Physics Depart-
sentative building, the cooling load almost dou- ment, Athens, Greece, Ph.D. dissertation.
bled in the central Athens area, while peak Santamouris M., Argiriou A. and Papanikolaou N. (1996)
electricity load may be tripled for higher set point Meteorological stations for microclimatic measurements. In
Report to the POLIS project, Commission of the European
temperatures. In parallel, the minimum COP value Commission, Directorate General for Science Research and
of the air conditioners in the central Athens area Technology, September; available through the authors.
is reduced by up to 25%. Santamouris M. and Assimakopoulos D. (Eds.) (1996) Passive
Cooling of Buildings, James and James, London.
Higher temperatures in the city centers affect Santamouris M. (Ed.) (1999) Energy and Climate in the
the heating load of urban buildings as well. Urban Built Environment, James and James, London, UK.
Calculations based on urban climatic data have Solar Energy Laboratory (1997). University of Wisconsin and
Transsolar Stuttgart
:
TRNSYS version
14
-A Transient Sys-
shown that the heating load of urban buildings tem Simulation Program.
may be reduced up to 30–50% compared to Stanners D. and Bourdeau P. (Eds.) (1995) Europe
’
s Environ-
buildings located in suburban areas. ment
:
the Dobris Assessment, European Environmental
Agency, Copenhagen.
Finally, an assessment of the potential of Tso C. P. (1994) The impact of urban development on the
natural ventilation in urban areas and in particular thermal environment of Singapore. In The Report of the
in urban canyons are presented. Calculations are Technical Conference on Tropical Urban Climates, WMO,
Dhaka.
based on air flow and temperature measurements Watanabe H., Yoda H. and Ojima T. (1990/ 91) Urban
taken in 10 different canyons in Athens. A very environmental design of land use in Tokyo metropolitan
serious reduction of the natural ventilation po- area. Energy Build
.
15–16, 133–137.
tential inside canyons is calculated especially
during the day time. Compared to the air flow
rates when undisturbed ambient meteorological
data are used, air flow rates inside canyons may
be reduced up to 10 times.
