Theor. Appl. Climatol. 79, 151–164 (2004)
DOI 10.1007/s00704-004-0070-7
1 Meteorological Research Institute, Korea Meteorological Administration, Seoul, Korea
2 School of Earth and Environmental Sciences, Seoul National University, Seoul, Korea
Daily maximum urban heat island intensity
in large cities of Korea
Y.-H. Kim1 and J.-J. Baik2
With 8 Figures
Received December 23, 2003; revised April 4, 2004; accepted April 28, 2004
Published online August 27, 2004 # Springer-Verlag 2004
Summary
This study investigates the characteristics of the daily
maximum urban heat island (UHI) intensity in the six
largest cities of South Korea (Seoul, Incheon, Daejeon,
Daegu, Gwangju, and Busan) during the period 1973–2001.
The annually-averaged daily maximum UHI intensity in all
cities tends to increase with time, but the rate of increase
differs. It is found that the average annual daily maximum
UHI intensity tends to be smaller in coastal cities (Incheon
and Busan) than in inland cities (Daejeon, Daegu, and
Gwangju), even if a coastal city is larger than an inland city.
A spectral analysis shows a prominent diurnal cycle in the
UHI intensity in all cities and a prominent annual cycle in
coastal cities. A multiple linear regression analysis is under-
taken in order to relate the daily maximum UHI intensity to
the maximum UHI intensity on the previous day (PER), wind
speed (WS), cloudiness (CL), and relative humidity (RH). In
all cities, the PER variable is positively correlated with the
daily maximum UHI intensity, while WS, CL, and RH vari-
ables are negatively correlated with it. The most important
variable in all cities is PER, but the relative importance of
the other three variables differs depending on city. The total
variance explained by the multiple linear regression equation
ranges from 29.9% in Daejeon to 44.7% in Seoul. A multi-
dimensional scaling analysis performed with a correlation
matrix obtained using the daily maximum UHI intensity data
appears to distinguish three city groups. These groupings are
closely connected with distances between cities. A multidi-
mensional scaling analysis undertaken using the normalized
regression coefficients obtained from the multiple linear
regression analysis distinguishes three city groups. Notably,
Incheon and Busan form one group, whose points in the two-
dimensional space are very close. The results of a cluster
analysis performed using the multivariate data of PER, WS,
RH, and CL are consistent with those of the multidimen-
sional scaling analysis. The analysis results in this study
indicate that the characteristics of the UHI intensity in a
coastal city are in several aspects different from those in
an inland city.
1. Introduction
One of the most known phenomena associated
with inadvertent climate change is the urban heat
island (UHI), in which the air temperature in the
urban canopy is higher than that in the surround-
ing rural area. The UHI intensity varies with
urban size, urban surface characteristics, anthro-
pogenic heat release, topography, and meteorolog-
ical conditions (e.g. Landsberg, 1981; Oke,
1987). Many observational studies indicate that
the UHI is prominent on calm, clear nights and
its intensity can exhibit diurnal and seasonal
cycles (e.g. Ackerman, 1985; Jauregui, 1997;
Montavez et al., 2000). The UHI intensity is
influenced by synoptic and mesoscale circu-
lations (Yague et al., 1991; Yoshikado, 1994;
Runnalls and Oke, 2000; Gedzelman et al.,
2003). For example, in coastal cities under the
influence of sea breeze circulation, the UHI

intensity might be reduced by the intrusion of
relatively cold air into the land and=or the
enhanced wind speed due to the sea breeze.
Many investigators showed that the maximum
UHI intensity is well correlated with the popula-
tion (e.g. Oke, 1973; Park, 1986). Also, it has
shown that a strong relationship exists between
the maximum UHI intensity and urban param-
eters such as sky view factor and built-up ratio
(e.g. Park, 1986; Bottyan and Unger, 2003).
In a recent study, Kim and Baik (2002) inves-
tigated the daily maximum UHI intensity in
Seoul, Korea using data measured at two meteo-
rological observatories (an urban site and a rural
site) during the period 1973–1996. Their results
show that the average daily maximum UHI is
weakest in summer and strong in autumn and win-
ter. Similar to previous studies for other cities of
the world, the daily maximum UHI intensity is
more frequently observed in the nighttime than
in the daytime, decreases as the wind speed in-
creases, and is pronounced with clear skies. A
multiple linear regression analysis was under-
taken to relate the daily maximum UHI intensity
to the meteorological elements of maximum UHI
intensity on the previous day, wind speed, cloudi-
ness, and relative humidity. The analysis results
show that among those four elements the maxi-
mum UHI intensity on the previous day is the
most important and that the relative importance
among the elements varies depending on time of
day and season.
This study extends our previous work (Kim
and Baik, 2002) to characterize the daily maxi-
mum UHI intensity in the six largest cities of
Korea and find its characteristic similarities and
differences among the cities. In particular, we
focus on characteristic differences in the UHI
intensity between coastal and inland cities. For
this, spectral analysis, multiple linear regression
analysis, multidimensional scaling analysis, and
cluster analysis are performed using observed
data. In Section 2, data used and analysis method
are described. In Section 3, analysis results are
presented. Finally, a summary and conclusions
are given in Section 4.
2. Data and analysis method
The data used in this study are from the archives
of the Korea Meteorological Administration
(KMA). The data contain near-surface air tem-
perature (z ¼ 1.2
1.5 m, z: height from the
ground), wind speed, cloudiness, and relative
humidity measured at meteorological observa-
tories in the six largest cities of Korea (Seoul,
Incheon, Daejeon, Daegu, Gwangju, and Busan)
and surface air temperature at nearby meteo-
rological observatories (Yangpyong, Ganghwa,
Geumsan, Yeongcheon, Suncheon, and Geoje,
respectively). The data used span the years from
1973 to 2001 and are at 6-h intervals (03, 09, 15,
and 21 local times). In this study, the daily max-
imum UHI intensity of a city is defined as the
maximum temperature difference between the
city and its nearby observatories during a day.
Figure 1 shows the locations of the six paired
observatories. These pairs were chosen in pre-
vious studies on UHIs in Korea (Kim et al.,
2000; Kim and Baik, 2002).
Table 1 lists the climatological values of meteo-
rological elements during the period 1971–2000
in the six cities, together with the locations of
observatories. Also, the populations of years 1973
and 2001 are listed. Korea belongs to a temperate
climate zone with four distinct seasons. It is
warm=hot and humid in summer with precipita-
tion being concentrated in this season, and cold
Fig. 1. The locations of meteorological observatories
selected for this study. The distance between Seoul and
Yangpyong observatories is 60 km, Incheon and Gwanghwa
33 km, Daejeon and Guemsan 32 km, Daegu and
Yeongcheon 38 km, Gwangju and Suncheon 41 km, and
Busan and Geoje 54 km
152 Y.-H. Kim and J.-J. Baik