Effects of the large-scale atmospheric circulation on the

Theor Appl Climatol (2014) 117:73–87
DOI 10.1007/s00704-013-0989-7
ORIGINAL PAPER
Effects of the large-scale atmospheric circulation
on the onset and strength of urban heat islands: a case study
Admir Créso Targino & Patricia Krecl &
Guilherme Conor Coraiola
Received: 3 March 2013 / Accepted: 29 July 2013 / Published online: 16 August 2013
# Springer-Verlag Wien 2013
Abstract Air temperature was monitored at 13 sites across
the urban perimeter of a Brazilian midsize city in winter 2011.
In this study, we show that the urban heat island (UHI)
develops only at night and under certain weather conditions,
and its intensity depends not only on the site's land cover but
also on the meteorological setting. The urban heat island
intensity was largest (6.6 °C) under lingering high-pressure
conditions, milder (3.0 °C) under cold anticyclones and almost vanished (1.0 °C) during the passage of cold fronts. The
cooling rates were calculated to monitor the growth and decay
of the UHI over each specific synoptic setting. Over four
contiguous days under the effect of a lingering high-pressure
event, we observed that the onset of cooling was always at
about 2 h before sunset. The reference site attained mean
cooling rate of −2.6 °C h−1 at sunset, whilst the maximum
urban rate was −1.2 °C h−1. Under a 3-day cold anticyclone
episode, cooling also started about 2 h before sunset, and the
difference between maximum rural (−2.0 °C h−1) and urban
(−1.0 °C h−1) cooling rates diminished. Under cold-front
conditions, the cooling rate was homogeneous for all sites
and swang about zero throughout the day. The air temperature
has a memory effect under lingering high-pressure conditions
which intensified the UHI, in addition to the larger heat
storage in the urban area. Cold anticyclone conditions promoted the development of the UHI; however, the cold air pool
and relatively light winds smoothed out its intensity. Under
A. C. Targino (*) : G. C. Coraiola
Department of Environmental Engineering, Federal Technological
University of Paraná (UTFPR), Av. dos Pioneiros 3131,
86036-370 Londrina, PR, Brazil
e-mail: [email protected]
P. Krecl
Institute of Hydraulic Research, Federal University of Rio Grande do
Sul (UFRGS), Av. Bento Gonçalves 9500, 91501-970 Porto Alegre,
RS, Brazil
the influence of cold fronts, the urban fabric had little effect on
the city's air temperature field, and the UHI was imperceptible.
1 Introduction
It is estimated that by 2030, more than two thirds of the
world's population will live in urban areas. The fast pace of
human population growth over the last decades has been
observed especially in developing countries experiencing rapid industrialization. Interestingly, the urban population of these countries is projected to grow at a faster rate when compared to the developed world, with an average annual rate of
2.4 against 1.2 % (Population Division of the Department of
Economic and Social Affairs of the United Nations Secretariat
2007). Although urban areas take up just 2 % of the Earth's
surface, they represent a large burden on natural resources. For
example, urban areas account for about 75 % of industrial
wood use, and 60 % of the water withdrawn is for human use
(O'Meara 1999).
The environmental implications of increasing the urban
population without proper planning affect several areas, like
freshwater resources (Shiklomanov and Rodda 2003), food
security (Crush and Frayne 2011) and air quality (Duh et al.
2008). A confounding issue is that the increase in urban
population occurs irrespective of the size of the city.
However, Brazilian midsize cities (with populations between
100 and 500 thousand inhabitants) grew between 2000 and
2007, whilst small cities shrank, and large cities showed a
modest decline in population. In 2007, midsize cities concentrated 25.0 % of the country's urban population, compared to
23.8 % in 2000. The population of small and large cities
accounted for 46.4 and 29.8 % of the urban population,
respectively, and declined to 45.2 and 29.7 % over the same
period (Institute for Applied Economic Research 2008).
As regions transform from rural to urban, various environmental changes become evident. The vegetation that provides
74
moisture for evapotranspiration and cooling are removed,
and the introduction of impervious surfaces increases
risks of flash flooding when heavy rains fall over a city,
since runoff from an acre of pavement is about 10–20
times greater than the runoff from an acre of grass (Frazer
2005). Engineered materials used to build shelter, mobility and other modern facilities transform natural surfaces
into high-energy absorbing surfaces, rendering thermal
properties that retain more heat, impact radiative characteristics, reduce turbulent transfer of heat within streets
and ultimately rise the air temperature (Taha 1997). These
alterations favour the occurrence of the urban heat island
(UHI) phenomenon, whereby the temperature in the urban core is higher than in the surrounding suburbs and
rural areas. The UHI is a complex phenomenon, involving the interaction of various elements of the urban atmospheric system with on-site variables, such as the
geometry of urban canyons, thermal properties of materials, the urban greenhouse effect, reduction of albedo by
canyon geometry, reduction of turbulent transfer, reduction in evaporating surface and the production of anthropogenic heat (Oke 1987; Taha et al. 1988). The anthropogenic heat is released by both mobile (fuel combustion)
and stationary sources (electricity and fuel consumed by
houses, industrial and commercial activities) and by human and animal metabolisms (Ferreira et al. 2010).
Most urban sprawls have low surface albedo, which is
largely due to a high proportion of concrete or asphalt
paving in the open areas. Replacing surfaces of lowenergy absorption potential, like water and vegetation
areas, by surfaces of high-energy absorption potential,
such as concrete and asphalt, changes the partition of the
energy flux at surface level. In such scenario, the energy
that would be used in evapotranspiration, through the latent heat flux, will be used in surface heating through the
sensible heat flux. As impervious surfaces have a high
capability to store heat, the temperature of the air in contact
with the heated surface will consequently increase. The
UHI is also conditioned by off-site variables such as incoming solar radiation and large-scale wind field, which in
turn are related to the synoptic weather conditions (Kim
and Baik 2005; Hart and Sailor 2009). High wind speeds
and/or cloud cover, usually observed during the passage of
cold fronts, disrupt the cooling differences between the
urban and rural areas and reduce the urban heat island
intensity (UHII) (Kidder and Essenwanger 1995), whilst
calm conditions with clear skies usually observed in anticyclonic systems favour the development of stronger UHII
(Nkemdirim 1980; Yague et al. 1991; Unger 1996;
Gedzelman et al. 2003).
The UHI is associated with adverse health aspects (e.g.
thermal discomfort; Giles et al. 1990; Pantavou et al. 2011;
Matzarakis and Nastos 2011; Nastos and Matzarakis 2012),
A.C. Targino et al.
enhancement of air pollution (Jonsson et al. 2004; Zhao
et al. 2006; Sarrat et al. 2006) and precipitation (through
convection that replaces the hot air of the urban areas by
colder and dense air of the rural areas; Jauregui and
Romales 1996) and higher electricity consumption for
cooling buildings (Hekkenberg et al. 2008). On the positive side, some studies identified a few benefits in the
UHI phenomenon. Kolokotroni et al. (2010) found that
the UHI in London reduces the energy demand for
heating between 65 and 85 % in winter compared to sites
outside the UHI area. Hirano and Fujita (2012) reported
that the UHI increases commercial energy consumption in
the Tokyo metropolitan area, but decreases residential
energy use; nevertheless, they observed a total net decrease in energy consumption. Searle et al. (2012)
showed that red oak seedlings in New York's Central
Park accumulated eight times the biomass in the same
period as those grown in more rural, cooler settings in the
Hudson Valley and Catskill Mountains. The differences in
average maximum and minimum temperatures between
New York's Central Park and Catskill Mountains were
of 2.4 and 4.6 °C, respectively.
Previous investigations to assess the UHII focused on
comparing the temperature difference between pairs of urban
and rural sites. However, the heterogeneities found in urban
areas may result in different UHI intensities (Upmanis et al.
1998; Unger 2004), which is why extending the results of a
pair of stations to other urban points with different land cover
is not advisable. Moreover, it is important to study the weather
conditions which favour the UHI and their frequency of
occurrence, so that strategies of mitigation can be streamlined
for that particular study region. Many studies reported on
monthly (Unger et al. 2001) and even yearly (VelazquezLozada et al. 2006) values of UHII. Despite the important
contributions of these investigations to identifying local climate variability trends, these approaches hinder the assessment of the processes that control the development of the UHI.
For example, mean monthly values of UHII mask the daily
variability imposed by the atmosphere dynamics. From this
standpoint, assessing the onset and strength of UHI should
integrate both high-frequency air temperature data and description of the atmospheric conditions.
In this work, we utilised air temperature data gathered in
dedicated field measurements at 13 sites across a midsize city
in southern Brazil, covering a range of synoptic conditions
and land cover surfaces to characterise the spatio-temporal
distribution of the UHI. We focused on nocturnal intra-site
cooling rates and atmospheric circulation to pinpoint the conditions that favoured the development of the UHI. To investigate the effect of the large-scale atmospheric circulation on
the UHII, the data collected during the experiment were
segregated into the prevailing synoptic regimes observed during the campaign.
Effects of the large-scale atmospheric circulation on UHI
2 Methodology
2.1 Study area
Londrina is a city with 510,000 inhabitants located in the
northern part of the state of Paraná, Brazil (lat 23°19′S, long
51°08′W). The city was founded in 1934 by the British in an
area of semi-deciduous native forest that was mostly eradicated by slash-and-burn agriculture to grow coffee. In the 1960s
and 1970s, the city experienced a rapid urbanisation and
increased the housing density. The city's core is compact with
tall buildings (up to 20 storeys) and a few green areas. In the
region surrounding the city centre, the buildings are lower
with variable heights, whereas detached houses are mostly
present in the residential suburbs. A relatively new urbanised
area developed in the southern part of the city with even taller
buildings (up to 30 storeys). With each new development,
native vegetation was sacrificed as this area tries to keep up
with its ever increasing demand for dwellings and urban
infrastructure (shopping areas, sidewalks, driveways and
parking lots). The city's urban green areas are scattered and
amount to only 16 % (ca. 3.64 km2) (Polidoro et al. 2011).
Londrina has a humid subtropical climate (Cfa in the
Köppen–Geiger classification) with an annual mean temperature of 21.0 °C and annual mean precipitation of 1,630 mm.
Rainfall occurs throughout the year, with summer (Dec to
Feb) being the rainiest season and winter (Jun to Aug), the
driest one. In winter, the average temperature ranges from 11.6
to 25.8 °C, whereas the range varies from 19.0 to 29.7 °C in
summer. Insolation is greatest during winter time (mean value
of 225 h) when cloud coverage is at its lowest level. The city
has a gentle relief with altitudes ranging between 520 and
610 m above mean sea level, oriented on a northwest–southeast direction following the hydrographic basin, and with low
declivity (up to 10º).
2.2 Experimental set-up and data analysis
Between 7 Jun and 9 Aug 2011, the ICALON experiment
(acronym for the project's original name in Portuguese Ilha de
Calor em Londrina) was conducted to map out the air temperature field across the city perimeter. The study was divided
into three stages: (1) assessment of land cover around sites
strategically chosen to represent different surface covers on
the ground, (2) deployment of sensors and continuous measurements of air temperature at these sites and, further, (3)
separate data analysis in relation to synoptic scale circulations.
In the first stage, we performed an image segmentation and
unsupervised classification using the Georeferenced
Information Processing System (SPRING) software (Camara
et al. 1996). We used high-spatial resolution Google Earth
images from the QuickBird satellite to evaluate the land cover
in a squared area of 40,000 m2 around each monitoring site. In
75
short, the image classification with SPRING was performed as
follows (Rodríguez-Yi et al. 2000): (a) imagery importing and
geo-referencing with aid of a high-resolution image of the
Shuttle Radar Topography Mission; (b) performing supervised segmentation (identification of regions in the image that
are made up of pixels that have similar spectral properties); (c)
training by selecting regions (ground truth) of known land
cover categories. The images were trained to identify five
categories: vegetation, asphalt, roof, water and bare soil; and
(d) performing classification.
The air temperature was monitored at 13 sites (Fig. 1). The
data at the sites marked with an asterisk in Table 1 were
gathered using HOBO-U23 temperature data loggers (Onset
Computer, Bourne, MA). These sensors operate in an air temperature range of −40 to 70 °C, with ±0.2 °C accuracy and
0.02 °C resolution. In this study, the data acquisition interval
was set up to 1 min, and the sensors were deployed at 1.5–2.0 m
high, except for the thermometer at the city's Public Library
which was installed on the roof top (ca. 7.0 m) due to security
reasons. Air temperature at the Brazilian Agricultural Research
Corporation (EMBRAPA) was measured with a S-THB-M002
sensor (Onset Computer, Bourne, MA); at Londrina's Airport
(LDB), the temperature was sampled using a thermistor model
5190C (Qualimetrics, Inc., now All Weather, Inc., Sacramento,
CA); and at the State Meteorological Service (SIMEPAR), with
a HMP155 probe (Vaisala, Helsinki). These sensors have an
accuracy of ±0.2 °C. The wind speed and direction, and incoming short-wave radiation (SWR) were measured at EMBRAPA
using a pyranometer model S-LIB-M003 and propeller anemometer model S-WCA-M003, respectively (Onset Computer,
Bourne, MA).
The urban heat island intensity was calculated by the
following:
UHII ¼ T Ui −T ref
ð1Þ
where TUi is the temperature of the i-th station, and Tref is the
temperature of the reference station, here taken as the
Londrina State University (UEL) Campus located on the
western part of the city, spread over a 2.2-km2 area of which
90 % is covered by vegetation, comprising mainly native
forest and some cropland. We chose this site as reference
because it captures the patterns of the region's native vegetation and landscape, providing a natural framework for
assessing the impact of species depletion due to clearing on
the city's microclimate over the last decades. The species on
the campus are predominantly evergreen trees with large
trunks and well-developed crowns. Note that the vegetation
cover reported in Table 1 for UEL is for a restricted area
around the measurement point and, thus, differs from the
campus total vegetation percentage.
Ideally, the UHII is best assessed over areas with flat
terrain, since differences in terrain features are reflected into
76
A.C. Targino et al.
Fig. 1 Geographical location of
Londrina in Brazil (right upper
corner) and spatial distribution of
the monitored sites across the city
(left). The grey area delimits the
urban sprawl
not available, a constant value of 6.5 °C km−1 is generally
used from the Earth's surface up to 11 km. However, two
reservations should be made: (a) the lapse rate is affected by
wind speed and solar radiation and, thus, varies considerably
close to the Earth's surface along the day. For example, Pepin
et al. (1999) found nocturnal lapse rate as low as 0.2 °C km−1
increasing to 9.8 °C km−1 during the day in the English
horizontal variations of surface temperature and turbulent
fluxes. The altitude within the study area ranges from 534 m
(at IATE Clube (IAT)) to 633 m (at EMBRAPA (EMB)), with
the reference site located at 587 m. Hathway and Sharples
(2012) suggested to control the altitude influence on the air
temperature by applying a correction factor based on the
environmental lapse rate. When lapse rate measurements are
Table 1 Location of the monitored sites and land cover percentage in 40,000-m2 areas around each site obtained using the SPRING software. The sites'
acronyms are shown in parenthesis
Sites
Position
Land cover (%)
Lat (S)
Long (W)
Alt (m)
Vegetation
Roof
Asphalt
Water
Bare soil
SIMEPAR (SIM)
EMBRAPA (EMB)
Londrina State University (UEL)*
Technological University (UTF)*
City's Public Library (BIB)*
IATE Clube (IAT)*
Exposition Park (PQE)*
Servino Freitas St (SER)*
Belo Horizonte St (BHZ)*
Londrina's Airport (LDB)*
23°21′34″
23°11′38″
23°19′40″
23°18′27″
23°18′46″
23°19′44″
23°17′47″
23°15′34″
23°18′47″
23°19′59″
51°09′54″
51°11′02″
51°11′56″
51°06′52″
51°09′32″
51°09′59″
51°13′44″
51°10′35″
51°10′01″
51°08′04″
587
633
587
552
592
534
606
553
592
566
86
83
69
51
38
33
31
31
21
19
2
6
17
9
40
14
22
49
61
7
0
11
13
24
21
20
28
12
17
64
2
0
0
0
0
32
0
0
1
0
10
0
0
16
0
0
18
8
0
0
Lucia Helena Ave (LUC)*
Odilon Carvalho St (ODI)*
Lessence Building (LES)*
23°17′18″
23°17′53″
23°19′58″
51°09′17″
51°10′26″
51°10′34″
535
571
584
17
16
7
60
46
47
22
38
43
0
0
1
0
0
2
*Indicates data acquired with HOBO-U23 temperature dataloggers
Effects of the large-scale atmospheric circulation on UHI
77
uplands; (b) radiative cooling could produce temperature inversions in clear weather and cause warmer temperature close
to the ground. In this scenario, the urban–rural temperature
difference would be weakened rather than strengthened. By
applying a conservative correction of 6.5 °C km−1, our temperature measurements would be offset by a maximum of +
0.34 °C (adjustment based on the height difference between
IAT and reference site) and by a minimum of −0.30 °C (based
on the height difference between EMB and reference site),
which would still fall within the accuracy of the equipment.
Moreover, the adjustments would be lost in the overall calculations of the average air temperature and cooling rates. Thus,
our dataset was not corrected for altitude difference, given the
lack of air temperature vertical profiling.
3 Results
3.1 General overview
During the campaign, mean daily temperatures ranged between 7.4 °C (at UEL) and 26.3 °C (at BIB) (Fig. 2), and
the mean relative humidity was between 40 and 100 %. The
mean wind speed was 2.1 m s−1, and the dominant winds were
from west (34 %), northwest (17 %) and northeast (11 %).
Total monthly rainfall measured at EMB was 67.8 mm in Jun,
and 81.0 mm in Jul (Aug 2011 was not included since measurements span only a few days into this month). Compared to
the climatological period of 1976–2010, Jun and Jul 2011
presented monthly precipitation anomalies of −19.3 and
12.1 mm, respectively.
The anomaly of mean air temperatures for Jun was 1.4 °C
and for Jul, −0.9 °C; however, the differences between the
measurements in 2011 and the climatological period was not
statistically significant, according to the Mann–Whitney U
IAT
BIB
LES
SER
LUC
UTF
ODI
BHZ
PQE
UEL
07/07
12/07
17/07
22/07
27/07
EMB
LDB
01/08
06/08
27
23
Air Temp. [°C]
Fig. 2 Daily mean air
temperature in Londrina during
ICALON experiment. The
hatched grey areas highlight
periods in which the intra-site air
temperature differences are
largest
test at 5 % significance level. To analyse tendencies in extreme
temperature events over the period of 1976–2010, we calculated the warm day (Tx90) and warm night (Tn90) indices.
They are defined as the number of days in which the maximum temperature is larger than 90th percentile of daily maximum temperature and the number of days in which the
minimum temperature is larger than the 90th percentile of
the daily minimum temperature, respectively (Frich et al.
2002). We observed an uptrend in the values of Tx90 and
Tn90. For example, in 1976, Tx90 and Tn90 were 6 and 17,
respectively, jumping to 42 and 57 in 2010, with some falls in
the time series that match the occurrence of El Niño events,
which feature enhanced rainfall in southern South America
and, thus, a decrease in air temperatures. These results suggest
that in the twenty-first century, periods of hot sustained temperatures will be more frequent in Londrina. This pattern is in
agreement with model results reported by the IPCC (2012)
which project increased dryness for southern Brazil in the
twenty-first century, with increases in the frequency and magnitude of warm daily temperature extremes and decreases in
cold extremes.
Broadly speaking, the air temperature curves during
ICALON followed one another (Fig. 2), with a few outstanding features. Over some periods, large daily intra-site temperature differences (hatched grey areas) were recorded, whilst
the difference was modest during other measurement periods.
For example, on 15 Jul, the average daily UHII between a
street canyon site (defined as a relatively narrow street with
buildings lined up along both sides) on Belo Horizonte St
(BHZ), surrounded by roof, cement and asphalt, and the
reference site UEL was 2.7 °C, whilst on 2 Jul, the UHII
between the same sites was only 0.3 °C. Given that the city's
fabric makeup was unaltered over this period, the different
UHIIs suggest that the air temperature was controlled not only
by the local but also by external variables related to the
19
15
11
7
12/06
17/06
22/06
27/06
02/07
Date [2011]
SIM
78
A.C. Targino et al.
The measurements were segregated into three prevailing synoptic situations observed during the experiment: lingering
high-pressure conditions (LHP), cold anticyclones (CA) and
passage of cold fronts (CF). Lingering high-pressure conditions occurred at 47 % of the time and featured clear skies and
wind speed <2 m s−1. These conditions usually locked in place
across Londrina between 4 and 6 days, producing scarce or no
rainfall, weak flows and pushing temperatures up. The cold
fronts (21 % of occurrence) were accompanied by rainfall, a
shift in wind direction from north to south, wind speed larger
than 4 m s−1 and a temperature change of at least 2 °C. Cold
anticyclones (32 %) occurred either due to surface and upper
level anticyclogenesis or in the wake of cold fronts and
brought clear skies, a sharp drop in temperature and wind
speed <4 m s−1.
South America has a strong frontogenetic region over the
east-central part, between the two semi-permanent subtropical
anticyclones (Satyamurty and Mattos 1989). The fronts in this
region have a southwest–northeast trajectory and are more
frequent below 30° S during winter (average of 45 cold fronts
per year) with the frequency of occurrence decreasing towards
the equator. The passage of frontal systems over Londrina is
between only 10 and 15 per year (Cavalcanti and Kousky
2009). This relatively low number is influenced by strong and
lingering blocking highs which prevent the rain-bearing cold
fronts from penetrating the region. As a consequence, rainfall
is weakest and insolation is greatest during the winter months,
which contribute to the occurrence of heat waves. With little
rainfall to alleviate the hot and dry spells, the UHII is expected
to be largest in this season.
UHII [°C]
4
a
3.3 Spatial distribution of minimum temperatures and UHII
Since the intra-site temperature differences and the urban heat
island intensity were more pronounced at night, we focused
on the spatial distribution of average minimum air temperature
and the nocturnal UHII (1800–0700 hours) under the three
synoptic situations (Figs. 4 and 5).
b
500
c
SWR
2.5
375
1
250
−0.5
125
−2
0
2
4
6
8 10 12 14 16 18 20 22 0
2
4
hours
IAT
BIB
6
8 10 12 14 16 18 20 22 0
2
4
hours
LES
SER
LUC
UTF
6
8 10 12 14 16 18 20 22
−2
3.2 UHII and synoptic situations
Figure 3 shows the diurnal cycle of UHII for all monitored
sites in the three atmospheric regimes, along with the incoming SWR. The intra-site air temperature differences under CF
conditions are within 1.5 °C of one another, with the curves
swinging about 1.0 °C throughout the day. In contrast, under
the influence of CA and LHP systems, the UHII was largest in
the night hours at urbanised sites and converged towards zero
between 9 and 16 h. Under calm, clear-sky anticyclonic conditions, the engineered urban materials absorb much solar
radiation and convert it mostly into sensible heat flux and
storage heat flux, and thus increase the air temperature in
contact with the surface. The more vegetated sites (for example, UEL campus), unlike buildings and asphalt, absorb part of
the solar radiation and divert it from warming the soil into
evaporating water off of the leaves and soil to naturally cool
the surrounding air. However, this cooling mechanism is more
efficient when water is available in the soil. For example, the
State Meteorological Service (SIM) site showed a cool island
during daytime, which may be partly attributed to the irrigation of the crops at the experimental farm surrounding the
meteorological station. Sacks et al. (2009) and Lobell and
Bonfils (2008) reported that irrigation increases evapotranspiration and alters the surface energy partition to favour latent
over sensible heating. This shift is usually accompanied by a
decrease in soil and near surface air temperatures, typically
affecting the diurnal maximum more than the minimum
temperature.
Incoming SWR [Wm ]
meteorological setting. These features will be discussed in
more detail throughout the paper.
0
hours
ODI
BHZ
PQE
EMB
LDB
SIM
Fig. 3 Average diurnal cycles of UHII for the three prevailing synoptic situation: a cold front, b cold anticyclones and c lingering high pressure. The
incoming short-wave radiation (dashed line) is also shown
Effects of the large-scale atmospheric circulation on UHI
79
Fig. 4 Spatial distribution of mean minimum air temperature in Londrina for a cold front, b cold anticyclones and c lingering high-pressure conditions in
winter 2011
A common feature in these figures is the almost regular
concentric shapes with larger values in the city's urban core,
decreasing towards the outskirts. The minima are more homogeneously distributed under CF (between 9.3 and 10.8 °C) and
CA (between 7.3 and 9.1 °C) conditions. The cold air masses
advected during and after the passage of the cold front smooth
out the air temperature spatial distribution, resulting in a more
homogeneous thermal field. The largest minima (15.8 °C)
occurred under LHP conditions at sites that concentrate buildings, pavements and anthropogenic activities which release
heat and contribute to increase the air temperature (e.g. BHZ
and BIB sites). A deviation from this concentric shape occurs
in the south around LES site where another branch of relatively high minima was observed. This area was formerly
occupied by farmlands until the 1990s when high-rise developments started to emerge. The consequence of this recent
land cover alteration is seen on the southern branch of the UHI
penetrating at about 23.35° S and surrounded by (still) relatively cooler and more rural areas (Figs. 4c and 5c). Another
hallmark of Fig. 4c is a sharp temperature gradient between
the city centre and the southeastern sites (ca. 1.8 °C km−1
compared to IAT, for example). The IAT site is located on the
fringes of Igapó Lake—an artificial lake about 3.5 km long
and 200 m wide, with an average depth of 2 m. Suzuki (1999)
discussed that the water surfaces exert a cooling influence on
local air temperature by evaporating water and absorbing heat,
and via the wind path effect, in which ventilation is more
active in the open space of river area than in the crowded
Fig. 5 Spatial distribution of the mean nocturnal (1800–0700 hours) urban heat island intensity in Londrina for a cold front, b cold anticyclones and c
lingering high-pressure conditions in winter 2011
80
A.C. Targino et al.
street. However, not all water body has a cooling effect.
Sugawara et al. (2009) showed that the temperature of a
shallow water channel (a few centimetres deep) in Tokyo
actually exceeds the air temperature during daytime, but the
air temperature over deeper water channels (about 5 m) is
cooler that the surrounding air, which helps alleviate the urban
heat island effect. The Igapó Lake has an average depth
comparable to the deep water channels investigated by
Sugawara et al. (2009) and has also a horizontal free path that
allows the wind to distribute the cooler air to areas in the
vicinity of the lake.
Some sites with large percentage of vegetation, such as
Federal Technology University Campus (UTF) and EMB,
did not present the expected lowest minima. These sites
are surrounded by croplands, which dry out more rapidly
during periods of scarce rainfall, like the ones observed
under LHP conditions. In this scenario, the soil moisture
depletion occurs rapidly, and water-stressed vegetation
controls transpiration through a physiological mechanism
whereby their stomata shut. When the stomata are closed,
even vegetated lands behave like dry soil and lose their
cooling capacity (Avissar 1992). Another aspect that
should be taken into account is the composition on the
vegetation. Teuling et al. (2010) showed that forests are
more effective in keeping a constant evaporative rate,
which sustains its cooling effect for a longer period. The
city's public library (BIB) is located in front of a 4.0-ha park
but, in spite the 38 % of vegetation cover around the monitored point, the site received little influence of the surrounding
vegetation and ends up inserted in the heat island core
(Fig. 5c). Spronken-Smith (1994) and Upmanis et al.
(1998) discussed that the park size is important in determining their influence on the surroundings (the effect of
larger parks extends at larger distances outside the park).
The built-up area around the BIB site consists almost
entirely of streets with buildings that run parallel to the
park's borders, which may prevent the cooler air in the
park from intruding into the built-up area. Upmanis et al.
(1998) recorded similar behaviour when they observed that
a 2.4-ha park in Sweden exerted no influence on the air
temperature of its surroundings. A summary of the maximum UHII and dates of occurrence is exhibited in Table 2.
3.4 Thermal diurnal amplitude and cooling rates
To analyse in more detail the relationship between synoptic
situation and the UHII, we selected three periods representative of each synoptic category. Prefrontal cloudiness followed
by a cold-front passage was observed between 30 Jun and 3
Jul; a postfrontal cold anticyclone entered the region on 4 Jul;
and finally, a period with anticyclonic weather condition
persisted between 12 and 17 Jul. We explored the behaviour
of four out of the 13 sites monitored in this study, representing
very different surface cover:
The BHZ site is located in a street canyon in a dense builtup area in the city's urban core, with paved sidewalks,
little vegetation and with intense light-duty vehicle traffic
during weekdays.
The UTF is located on the eastern edge of the city, in a
cropland area with the closest neighbourhood consisting
of sparsely built, detached houses about 400 m away,
across wheat and soybean fields.
The Servino Freitas St (SER) site is located in a residential suburb with detached houses further out the city. It is
close to the city's urban limit, with croplands to the north
of the site.
The weather station of the SIM is located at the experimental farm of the Agronomic Institute of Paraná in the
south part of the city. The farm includes 740 acres of
cropland (coffee, soybean, maize and wheat among
others) for research and demonstration projects.
Figure 6a clearly shows modest intra-site temperature differences with a weak diurnal cycle signal during CF conditions for all stations. The daily thermal amplitude (ΔT) for 1
Jul, for example, lies between 2.0 and 2.7 °C (Table 3). The
reduced amount of incoming SWR observed during the coldfront passage (reaching a maximum value of 165 W m−2 on 1
Jul) lessened the absorption of heat by the surfaces. In this
situation, the air temperature is governed primarily by the cold
air mass. On 4 Jul, a postfrontal anticyclone transported a pool
of cold air over the region, bringing relatively clear skies. This
intensified anticyclone activity in conjunction with the enhanced incoming SWR conferred a new behaviour on the
diurnal air temperature cycle. First, the thermal amplitude
Table 2 Maximum hourly urban heat island intensity observed during ICALON in winter 2011
Sites
Max UHII
[°C]
Date and
time
SIM
EMB
UTF
BIB
IAT
PQE
SER
BHZ
LDB
LUC
ODI
LES
4.00
5.20
3.30
5.70
2.60
3.60
4.30
6.60
3.70
3.70
4.40
6.10
15 Jul
2200
15 Jul
2200
13 Jun
2000
9 Aug
0300
28 Jun
0200
9 Aug
0300
8 Aug
2100
9 Aug
0300
12 Jun
1800
6 Aug
2100
8 Jul
0300
9 Aug
0300
Effects of the large-scale atmospheric circulation on UHI
81
600
UEL (ref)
BHZ
SER
UTF
SIM
SWR
a
Air Temp. [°C]
26
500
22
400
18
300
14
200
10
100
6
30/06
01/07
Incoming SWR [W m−2]
30
Fig. 6 Diurnal cycles of air
temperature and incoming SWR
for three synoptic situations: a
cold front, b cold anticyclone and
c lingering high pressure
0
03/07
02/07
Date [2011]
600
30
26
500
22
400
18
300
14
200
10
100
6
04/07
05/07
Incoming SWR [W m−2]
Air Temp. [°C]
b
0
07/07
06/07
Date [2011]
30
600
26
500
22
400
18
300
14
200
10
100
6
12/07
13/07
15/07
14/07
Date [2011]
16/07
0
17/07
Incoming SWR [W m−2]
Air Temp. [°C]
c
82
A.C. Targino et al.
increased irrespective of the location, reaching the highest
value at the SIM site (10.1 °C). The enhanced amplitude is
due to the decrease in the minimum temperature rather
than increase in the maxima. Actually, on 6 Jul, the maxima is only slightly larger (between 19.6 and 20.4 °C)
compared to the CF of 1 Jul (between 18.2 and
18.5 °C). Second, the UHI becomes evident when intrasite temperatures start to diverge in the evening and night.
The UTF, UEL and SIM sites registered the lowest nocturnal temperatures, whilst BHZ and SER register the
largest. Even though urban heat islands are usually associated with adverse microclimate conditions, this configuration represents a situation in which the UHI may be
beneficial since it will promote offsetting of the cold surge,
pushing the nocturnal temperatures up in the urban core.
A high-pressure system established itself over the study
area between 12 and 17 Jul, producing sustained dry weather
and clear or relatively clear skies (maximum value of incoming SWR of 540 W m−2 on 12 Jul). The thermal amplitude
increased at all sites and is more prominent at the rural sites
due to the drop in the minima. What becomes clear in periods
affected by anticyclonic conditions is that the maximum temperatures attain similar values irrespective of the site (Table 3),
which suppress the development of diurnal UHI in Londrina.
This behaviour has been ubiquitously observed worldwide,
with only a few studies reporting on daytime UHI (e.g.
Nkemdirim and Truch 1978; Jauregui 1997). A plausible
explanation for this is that during prolonged periods with lack
of precipitation, rural and urban sites use the net radiation in
similar ways: when the vegetation is water stressed during the
day, the stomata are closed, and a considerable portion of the
net radiation is spent for heating the surface just like in the
urban area. In this scenario, the inhibition of evaporationmediated cooling at the rural sites was instrumental in reducing the difference in maximum temperatures between urban
and rural areas.
The hourly cooling rates were calculated to monitor the
growth and decay of UHI over each specific synoptic setting.
Figure 7 shows the mean cooling rates, urban heat island
intensity and wind speed for the LHP episode described previously. The change from daytime warming to nocturnal
cooling started 2–3 h before sunset, and until about 1600 hours,
the sites were cooling at the same rate (between 0.13 and
−0.27 °C h−1). At 1700 hours, the cooling rate at the reference
site started to diverge, attaining its highest rate at 1800 hours
(−2.6 °C h−1), whilst the street canyon station presented a
lower rate (−1.25 °C h−1). At this time, the cooling rates for
the other land-use types ranged between −0.8 and −2.1 °C h−1.
During the rest of the night, the cooling slowed down, and at
about 4 h after sunset until sunrise, there is little difference in
cooling rates between the sites, although the absolute air temperature differs and remains that way throughout the night.
These results are commensurable with values reported in
Vancouver (Oke and Maxwell 1975) and Gothemburg
(Holmer et al. 2007).
Nocturnal cooling processes are governed by outgoing
long-wave radiation. Urban areas have larger heat capacity,
whilst cities are typically more effective at storing the sun's
energy as heat within their infrastructure. At night, these
materials continue to emit heat to the adjacent air, which
may be absorbed and re-emitted by other urban elements
and slow down the cooling rate.
This hysteresis lag effect results in large amounts of energy
stored in the urban canopy during the day and released after
sunset at a different cooling rate compared to the rural site
(Holmer et al. 2007). The mean cooling rates for the other
situations depicted in Fig. 7 are shown in Table 4. We have
highlighted the hours when cooling is largest for LHP (between 1700 and 2300 hours) to facilitate comparison amongst
the three weather regimes. Note that despite the smaller
cooling rates under CA conditions, they remained sitedependent, with the reference site exhibiting the largest
values. The cold pool of air introduced by the postfrontal
anticyclone served to decrease the overall minimum temperatures; however, the effect of differential heating of the surfaces is still observed.
The UHII grew after sunset as a result of these diverging
dissipation rates (Fig. 7b) reaching maximum values of 3–5 h
later. Similar results were observed in various studies
conducted in Canada (Oke and Maxwell 1975), South Korea
(Kim and Baik 2002), Sweden (Holmer et al. 2007) and the
UK (Kolokotroni and Giridharan 2008). The cooling pattern
for CF was uniform with all sites having similar, modest rates
(mean of −0.25 °C h−1).
Table 3 Summary of maximum, minimum and thermal amplitude at selected sites in Londrina in winter 2011 for days representative of cold front, cold
anticyclone and lingering high-pressure situations
CF (1 Jul)
Max [°C]
Min [°C]
ΔT [°C]
CA (6 Jul)
LHP (15 Jul)
UEL
BHZ
SER
UTF
SIM
UEL
BHZ
SER
UTF
SIM
UEL
BHZ
SER
UTF
SIM
18.2
15.6
2.6
18.3
16.3
2.0
18.5
15.9
2.6
18.2
15.9
2.3
18.4
15.7
2.7
19.8
10.5
9.3
19.6
13.1
6.5
19.7
12.1
7.6
19.8
10.8
9.0
20.4
10.3
10.1
27.6
11.5
16.1
27.7
15.7
12.0
27.8
15.0
12.8
27.2
12.7
14.5
27.6
10.4
17.2
ΔT [°C]
Fig. 7 a Mean cooling rates, b
UHII and c wind speed for
selected sites under LHP
conditions in the period 12–17 Jul
2011 between 1500 and
0900 hours
Cooling rate [ºC hr−1]
Effects of the large-scale atmospheric circulation on UHI
1
83
a
0
−1
−2
UEL (ref)
−3
4
BHZ
SER
SIM
UTF
b
2
Wind speed [m s−1]
0
−2
c
2
1
Sunset
0
15
17
19
Sunrise
3
1
23
21
5
7
9
Hour
Table 4 Hourly cooling rates observed during cold front, cold anticyclone and lingering high-pressure situations at selected sites in Londrina
Hour
Cooling rates (°C h−1)
CF
CA
LHP
UEL
BHZ
SER
UTF
SIM
UEL
BHZ
SER
UTF
SIM
UEL
BHZ
SER
UTF
SIM
15
16
17
18
19
20
21
22
23
00
0.07
0.07
−0.03
−0.27
−0.30
−0.30
−0.20
−0.27
−0.10
−0.10
0.07
0.07
−0.17
−0.03
−0.20
0.07
−0.23
−0.23
0.00
−0.20
0.10
0.10
−0.17
−0.27
−0.30
−0.37
−0.23
−0.23
−0.20
−0.17
0.07
0.07
−0.33
−0.20
−0.23
−0.30
−0.10
0.03
−0.27
0.00
−0.10
−0.10
0.00
−0.27
−0.23
−0.27
−0.18
−0.22
−0.13
0.00
−0.10
−0.90
−1.57
−1.03
−0.90
−0.67
−0.23
−0.50
0.03
−0.47
−0.13
−0.53
−0.70
−0.47
−0.50
−0.93
−0.33
−0.40
−0.43
−0.17
−0.20
−0.80
−1.13
−0.97
−0.70
−0.37
−0.27
−0.23
−0.13
−0.13
−0.20
−0.83
−0.80
−0.37
−0.40
−0.93
−0.33
−0.57
−0.23
−0.30
−0.10
−0.13
−1.53
−1.30
−0.67
−1.00
−1.03
−0.20
−0.77
0.00
0.38
−0.27
−1.50
−2.63
−1.28
−0.93
−0.92
−1.17
−0.72
−0.92
0.08
−0.27
−0.78
−1.25
−0.77
−0.53
−0.67
−0.85
−0.68
−0.70
0.72
0.07
−0.90
−2.12
−1.73
−1.12
−0.73
−0.57
−0.47
−0.60
0.20
−0.23
−1.07
−0.83
−1.13
−0.83
−0.93
−1.30
−1.17
−1.02
0.75
0.13
−0.38
−1.48
−1.53
−0.82
−1.63
−0.97
−1.42
−0.92
01
02
03
04
05
06
07
08
09
−0.10
−0.07
−0.10
0.07
0.10
0.03
0.10
0.37
0.43
−0.10
−0.23
0.03
0.33
−0.17
0.03
0.37
−0.17
0.30
−0.07
−0.03
−0.03
0.00
−0.03
0.20
0.10
0.27
0.47
−0.20
0.10
−0.03
−0.03
0.03
0.07
0.03
0.17
0.40
−0.33
0.00
0.03
−0.03
0.07
0.07
0.03
0.07
0.40
−0.40
−0.20
−0.53
−0.37
−0.27
−0.07
0.83
1.27
1.73
−0.20
−0.37
−0.20
−0.17
−0.20
−0.20
0.10
0.53
1.60
−0.43
−0.37
−0.27
−0.23
−0.40
−0.03
0.60
1.17
1.57
−0.50
−0.40
−0.53
−0.47
−0.33
0.03
0.87
1.37
1.70
−0.10
−0.40
−0.40
−0.47
−0.23
−0.33
−0.20
1.07
1.63
−0.45
−0.67
−0.60
−0.27
−0.38
−0.37
−0.48
1.18
2.97
−0.93
−0.45
−0.60
−0.92
−0.20
0.05
−0.32
−0.07
1.60
−0.77
−0.60
−0.42
−0.52
−0.20
−0.28
−0.40
0.75
1.78
−0.53
−0.43
−0.77
−0.47
−0.38
−0.50
−0.07
1.53
2.78
−1.00
−0.37
−0.68
−0.72
−0.47
0.10
−0.55
−0.57
2.08
Hours when cooling is largest for LHP are highlight with italic
84
100
2.5
CA
CF
Vegetation
Asphalt
Roof
Water
Soil
75
1.75
50
1
25
0.25
0
IAT
SIM
EMB
UTF
BIB
SER
PQE
BHZ
LDB
LUC
ODI
LES
UHII [°C]
LHP
Percentage [%]
Fig. 8 Relationship between land
cover (vertical bars) and UHII for
the three meteorological regimes
observed during the experiment:
lingering high-pressure (solid
line), cold anticyclones (dashed
line) and cold front (dasheddotted line)
A.C. Targino et al.
−0.5
Sites
The urban–rural cooling rates observed in Londrina seem
to follow the conceptual model suggested by HaegerEugensson and Holmer (1999), in which a differential cooling
phase accompanied by a decrease in wind speed starts before
sunset. It is followed by a transition phase, characterised by a
slight increase in wind speed and decrease in cooling rates. In
the final phase (stabilisation), the pressure gradients between
urban and rural areas trigger advective transport between these
environments. As pulses of air are exchanged, the temperature
difference and the cooling rates lessen and remain small until
the morning when a new cycle begins.
3.5 Relation between land cover and UHII
We discussed that the UHII is controlled by both on-site and
off-site variables. Figure 8 summarises this analysis by showing the relationship between land cover and UHII. The vertical
bars represent the land cover percentage at each monitoring
site (sites are sorted by increasing vegetation cover, except for
IAT which is plotted as the first station due to its particular
behaviour in relation to water cover), and the lines correspond
to the mean UHII for each synoptic situation. We calculated
the mean UHII utilising all data for each synoptic category and
not only data for the nocturnal period as explored in Fig. 5,
which yields slightly smaller intensities. In the overall trend,
we observe a clear negative relationship between vegetation
cover and UHII. This relationship is more evident under LHP
and CA conditions, with urban stations showing maximum
values of 2.0 °C. It is interesting to note that under these
conditions, the UHII curves follow the same pattern, which
illustrates that the surfaces are responding to the solar heating
even during the incursion of cold air by the cold anticyclones.
The urban sites BHZ, LDB and Lucia Helena Ave (LUC)
have different mean UHI intensities, in spite of the similar
combined asphalt and roof percentages (78, 71 and 82 %,
respectively). The UHII was largest at BHZ (2.0 °C) and
decreased at urban sites with lower aspect ratio (height of
the buildings divided by the street width) like LUC
(1.1 °C) and at sites with relatively open space like LDB
(0.8 °C). The large storage heat of impervious surfaces in
the daytime and the low evaporation rates have been considered to be one of the main causes of the UHI (Kusaka and
Kimura 2004; Taha 1997). However, the descriptive aspects
of the relationship between surface cover and UHII within
the urban environment must go beyond the two-dimensional
representation of urban surfaces. Due to the vertical extensions of the buildings, land cover in urban environments is
best described as three-dimensional geometry, which encompasses causative factors of the UHI such as heat stored
in the vertical walls, radiation trapping and wind speed
reduction (Ryu and Baik 2012). The absorption of solar
radiation is larger in urban canyons because of the multiple
reflections from the vertical surfaces, and increases with
increasing building height (Kusaka and Kimura 2004). At
night, the canyon structure traps outgoing long-wave radiation, reducing radiative cooling and contributing positively
to the nocturnal UHII (Kusaka et al. 2001; Martilli 2002).
The penetration of the wind into the street is also reduced
because of the drag caused by the city's structure, which
causes warm air over urban areas to stagnate. In a relatively
open space with small roughness length, like at LDB, radiative loss is facilitated due to enhanced turbulence. Thus, in
spite of the similarity of the urban fabric at BHZ, LDB and
LUC, the temperature regimes at an urban canyon site like
BHZ may be quite different from that in open spaces or in
areas of residential low-rise buildings (such as LUC).
4 Conclusions
The air temperature field was monitored at 13 sites in
Londrina between Jun and Aug 2011. We investigated the
onset, development and intensity of the UHI as a function of
land cover and prevailing large-scale atmospheric circulation.
Effects of the large-scale atmospheric circulation on UHI
Different patterns of cooling rates and UHI spatial distribution
emerged when the dataset was segregated and analysed
according to the synoptic situations. One important result
from this work is that the diurnal intra-site air temperature
differences are small during the day, inhibiting the development of the daytime UHI. Our analyses showed that in winter,
the UHI develops only during the night and under certain
atmospheric conditions. During the passage of cold fronts,
the cooling was uniform across the city with all sites having
similar, modest rates. The UHII reached maximum 1.0 °C
during the passage of cold fronts. Under the influence of
anticyclonic circulation, cooling was site-specific and more
intense around sunset at the reference and suburban sites than
at urban sites, where cooling remained low and almost constant throughout the night. As result of the diverging nocturnal
cooling rates under anticyclonic conditions, the UHII reached
its largest intensity about 3 to 5 h after sunset in the city's
central core (in the order of 6.0 °C for lingering high-pressure
conditions and 3.0 °C for cold anticyclones). Our results
showed clearly that there exists a strong relationship between
land cover and UHI features in Londrina. It is under calm and
clear-sky anticyclones conditions when the surface cover
played a major role in determining the onset, time and intensity of the UHI, with predominantly urbanised sites having
larger UHII, whereas vegetation-rich areas have lower
UHII. We also identified that the urban geometry, albeit
not explicitly quantified in this study, affects the UHII at
sites within the urban core. We suggest that because of
the heat stored in the vertical walls, the radiation trapping
and wind speed reduction within the urban canyon, the
mean UHII is largest at BHZ (2.0 °C) and decreases for
sites in which the urban structure has lower aspect ratio,
such as LUC (1.1 °C) and LDB (0.8 °C), in spite of the
similar combined asphalt and roof percentages at these
sites. The composition of urban vegetation and proximity
to urban elements also affected the degree to which
vegetation cooled the environment. The city's public library exhibited large UHII despite its proximity to a park.
We hypothesise that the thermal effect of the densely
built-up area around the site overtakes the park's cooling
effect.
We concluded that any study aiming to characterise the
UHI and pinpoint the conditions that favour its development
should incorporate segregated observations over a range of
atmospheric conditions, since the surface-mediated heating
and cooling differ considerably depending on the meteorological setting and surface cover.
With anticyclones (either CA or LHP) accounting for about
80 % of the synoptic conditions during the campaign, it
becomes evident that the city's urban core is susceptible to
increase in minimum temperature during wintertime. With
prolonged periods of excessively hot weather predicted to
occur in the twenty-first century, intense nocturnal UHII like
85
those observed under anticyclonic conditions will be more
common.
Urban planning encompasses a complex network of disciplines to make the city aesthetically appealing, environmentally healthy and functional year-round. In this respect, it is of
utmost importance to have a cohesive understanding on the
UHI during other periods of the year to streamline strategies
for mitigation or use of the thermal energy trapped in the city's
engineered structure. An investigation on the UHI in Londrina
during summer is underway and will be object of a future
publication.
Acknowledgments The authors would like to thank Fundação Araucária (grant number 470/2010) for funding the air temperature sensors
used in this study. G. Coraiola is an undergraduate research bursary
financed by CNPq (grant number 555768/2010-4). We thank those individuals and companies that hosted the temperature sensors during the
measurement campaign, namely the students of the Federal Technology
University of Paraná, the managers of IATE Clube Londrina, Lessence
and Boulevard Park buildings and State University of Londrina (UEL).
EMBRAPA Soja, IAPAR, SIMEPAR and INFRAERO are acknowledged for furnishing meteorological data. The authors express their
sincere appreciation to the two anonymous reviewers who offered very
relevant suggestions to improve the quality of this work.
References
Avissar R (1992) Conceptual aspects of a statistical-dynamical approach
to represent landscape subgrid-scale heterogeneities in atmospheric
models. J Geophys Res 97:2729–2742
Camara G, Souza RCM, Freitas UM, Garrido J (1996) SPRING: integrating remote sensing and GIS by object-oriented data modelling.
Comput Graph 20(3):395–403
Cavalcanti IFA, Kousky VE (2009) Frentes frias sobre o Brasil (in
Portuguese). In: Silva Dias MAF, Cavalcanti IFA, Ferreira NJ,
Justi da Silva MGA (eds) Tempo e clima no Brasil. Oficina de
Textos, São Paulo, pp 135–148
Crush J, Frayne B (2011) Urban food insecurity and the new international
food security agenda. Dev South Afr 28(4):527–544
Duh J-D, Shandas V, Chang H, George LA (2008) Rates of urbanisation
and the resiliency of air and water quality. Sci Total Environ
400:238–256
Ferreira MJ, Oliveira AP, Soares J (2010) Anthropogenic heat in the city
of São Paulo. Braz Theor Appl Climatol 104(1–2):43–56. doi:10.
1007/s00704-010-0322-7
Frazer L (2005) Paving paradise: the peril of impervious surfaces.
Environ Health Perspect 113(7):A456–A462
Frich P, Alexander LV, Della-Marta P, Gleason B, Haylock M, Klein Tank
A, Peterson T (2002) Global changes in climatic extremes during the
2nd half of the 20th century. Clim Res 19:193–212
Gedzelman SD, Austin S, Cermak R, Stefano N, Partridge S,
Queensberry S, Robinson DA (2003) Mesoscale aspects of the
urban heat island around New York City. Theor Appl Climatol
75:29–42
Giles DB, Balafouts C, Maheras P (1990) Too hot for comfort: the
heatwaves in Greece in 1987 and 1988. Int J Biometeorol 34:98–104
Haeger-Eugensson M, Holmer B (1999) Advection caused by the urban
heat island circulation as a regulating factor on the nocturnal urban
heat island. Int J Climatol 18:681–700
86
Hart MA, Sailor DJ (2009) Quantifying the influence of land use and
surface characteristics on spatial variability in the urban heat island.
Theor Appl Climatol 19:975–988
Hathway EA, Sharples S (2012) The interaction of rivers and urban form
in mitigating the urban heat island effect: a UK case study. Build
Environ 58:14–22
Hekkenberg M, Benders RMJ, Moll HC, Schoot Uiterkamp AJM (2008)
Indications for a changing electricity demand pattern: the temperature dependence of electricity demand in the Netherlands. Energ
Policy 37:1542–1551
Hirano Y, Fujita T (2012) Evaluation of the impact of the urban heat
island on residential and commercial energy consumption in Tokyo.
Energy 37:371–383
Holmer B, Thorsson S, Eliasson I (2007) Cooling rates, sky view factors
and the development of intra-urban air temperature differences.
Geogr Ann A 89:237–248
Institute for Applied Economic Research (Ipea) (2008) http://www.redbcm.
com.br/arquivos/bibliografia/pesquisa%20ipea.pdf. Accessed 2 Jun
2013
IPCC (2012) Summary for policymakers. In: Barros V, Stocker TF, Qin
D, Dokken DJ, Ebi KL, Mastrandrea MD, Mach KJ, Plattner G-K,
Allen SK, Tignor M, Midgley PM, Field CB (eds) Managing the
risks of extreme events and disasters to advance climate change
adaptation. a special report of Working Groups I and II of the
Intergovernmental Panel on Climate Change. Cambridge
University Press, Cambridge, pp 1–19
Jauregui E, Romales E (1996) Urban effects on convective precipitation
in Mexico city. Atmos Environ 30:3383–3389
Jauregui E (1997) Heat island development in Mexico city. Atmos
Environ 31:3821–3831
Jonsson P, Bennet C, Eliasson I, Lindgren ES (2004) Suspended particulate matter and its relations to the urban climate in Dar es Salaam,
Tanzania. Atmos Environ 38:4175–4181
Kidder SQ, Essenwanger OM (1995) The effect of clouds and wind on
the difference in nocturnal cooling rates between urban and rural
areas. J Appl Meteorol 34:2440–2448
Kim YH, Baik JJ (2002) Maximum urban heat island intensity in Seoul. J
Appl Meteorol 41:651–659
Kim YH, Baik JJ (2005) Spatial and temporal structure of the urban heat
island in Seoul. J Appl Meteorol 44:591–605
Kolokotroni M, Davies M, Croxford B, Bhuiyan S, Mavrogianni A
(2010) A validated methodology for the prediction of heating and
cooling energy demand for buildings within the urban heat island:
case-study of London. Sol Energy 84:2246–2255
Kolokotroni M, Giridharan R (2008) Urban heat island intensity in
London: an investigation of the impact of physical characteristics
on changes in outdoor air temperature during summer. Sol Energy
82:986–998
Kusaka H, Kondo H, Kikegawa Y, Kimura F (2001) A simple singlelayer urban canopy model for atmospheric models: comparison with
multi-layer and slab models. Bound-Layer Meteorol 104:261–304
Kusaka H, Kimura F (2004) Thermal effects of urban canyon structure on
the nocturnal heat island: numerical experiment using a mesoscale
model coupled with an urban canopy model. J Appl Meteorol
43:1899–1910
Lobell D, Bonfils C (2008) The effect of irrigation on regional temperatures: a spatial and temporal analysis of trends in California. J Clim
21:2063–2071
Martilli A (2002) Numerical study of urban impact on boundary layer
structure: sensitivity to wind speed, urban morphology and rural soil
moisture. J Appl Meteorol 41:1246–1266
Matzarakis A, Nastos PT (2011) Human-biometeorological assessment
of heat waves in Athens. Theor Appl Climatol 105:99–106
Nastos PT, Matzarakis A (2012) The effect of air temperature and human
thermal indices on mortality in Athens. Theor Appl Climatol
108:591–599
A.C. Targino et al.
Nkemdirim L, Truch P (1978) Variability of temperature fields in Calgary,
Alberta. Atmos Environ 12:809–822
Nkemdirim LC (1980) A test of lapse rate/wind speed model for estimating heat island magnitude in an urban airshed. J Appl Meteorol
19:748–756
O'Meara M (1999) Reinventing cities for people and the planet.
Worldwatch paper 147. Worldwatch Institute, Washington
Oke TR, Maxwell GB (1975) Urban heat island dynamics in Montreal
and Vancouver. Atmos Environ 9:191–200
Oke TR (1987) Boundary layer climates. Methuen, New York
Pepin N, Benthan D, Taylor K (1999) Modelling lapse rates in maritime
uplands of northern England: implications for climate change. Arct
Antarc Alp Res 31:151–164
Pantavou K, Theoharatos G, Mavrakis A, Santamouris M (2011)
Evaluating thermal comfort conditions and health responses
during an extremely hot summer in Athens. Build Environ
46:339–344
Polidoro M, Lollo JA, Barros MVF (2011) Environmental impacts of
urban sprawl in Londrina, Paraná, Brazil. J Urban Environ Eng
5:73–83
Population Division of the Department of Economic and Social Affairs of
the United Nations Secretariat (2007) World urbanization prospects:
the 2007 revision. http://www.un.org/esa/population/publications/
wup2007. Accessed 3 Mar 2013
Rodríguez-Yi JL, Shimabukuro YE, Rudorff BFT (2000) Image segmentation for classification of vegetation using NOAA-AVHRR data. Int
J Remote Sens 21(1):167–172
Ryu Y-H, Baik J-J (2012) Quantitative analysis of factors contributing to
urban heat island intensity. J Appl Meteorol Clim 51:842–854
Sacks W, Cook B, Buenning N, Levis S, Helkowski J (2009)
Effects of global irrigation on the near-surface climate. Clim
Dyn 33:159–175
Sarrat C, Lemonsu A, Masson V, Guedalia D (2006) Impact of urban heat
island on regional atmospheric pollution. Atmos Environ 40:1743–
1758
Searle SY, Turnbull MH, Boelman NT, Schuster WSF, Yakir D, Griffin K
(2012) Urban environment of New York City promotes growth in
northern red oak seedlings. Tree Physiol 32:389–400
Satyamurty P, Mattos LF (1989) Climatological lower tropospheric
frontogeneis in the midlatitudes due to horizontal deformation and
divergence. Mon Weather Rev 117:1355–1364
Shiklomanov IA, Rodda JC (2003) World water resources at the beginning of the twenty-first century. Cambridge University Press,
Cambridge
Spronken-Smith RA (1994) Energetics and cooling in urban parks. Ph.D.
Thesis, University of British Columbia
Sugawara H, Narita K, Kim MS (2009) Cooling effect by urban river. The
7th International Conference on Urban Climate; 29 June–3 July 3,
Yokohama, Japan
Suzuki C (1999) A climatological study of the cooling effect of urban
rivers on heat island phenomena. Tokyo Metropolitan University,
Dissertation
Taha H (1997) Urban climates and heat islands: albedo, evapotranspiration, and anthropogenic heat. Energ Build 25:99–103
Taha H, Akbari H, Rosenfeld A, Huang J (1988) Residential cooling
loads and the urban heat island—the effects of albedo. Build
Environ 23:271–283
Teuling AJ, Sonia I, Seneviratne SI, Stöckli R, Reichstein M, Moors E,
Ciais P, Luyssaert S, van den Hurk B, Ammann C, Bernhofer C,
Dellwik E, Gianelle D, Gielen B, Grünwald T, Klumpp K,
Montagnani L, Moureaux C, Sottocornola M, Wohlfahrt G (2010)
Contrasting response of European forest and grassland energy exchange to heatwaves. Nat Geosci 3:722–727
Unger J (1996) Heat island intensity with different meteorological conditions in a medium-sized town: Szeged, Hungary. Theor Appl
Climatol 54:147–151
Effects of the large-scale atmospheric circulation on UHI
Unger J, Sümeghy Z, Zoboki J (2001) Temperature cross-section features
in an urban area. Atmos Environ 58:117–127
Unger J (2004) Intra-urban relationship between surface geometry
and urban heat island: review and new approach. Clim Res
27:253–264
Upmanis H, Eliasson I, Lindqvist S (1998) The influence of green areas
on nocturnal temperatures in a high latitude city (Göteborg,
Sweden). Int J Climatol 18:681–700
87
Velazquez-Lozada A, Gonzalez JE, Winter A (2006) Urban heat island
effect analysis for San Juan, Puerto Rico. Atmos Environ 40:1731–
1741
Yague C, Zurita E, Martinez A (1991) Statistical analysis of the Madrid
urban heat island. Atmos Environ 25B:327–332
Zhao SQ, Da LJ, Tang ZY, Fang HJ, Song K, Fang JY (2006) Ecological
consequences of rapid urban expansion: Shanghai, China. Front
Ecol Environ 4(7):341–346