www.elsevier.comrlocateratmos
Temperature cross-section features in an urban area
Janos Unger ´
a,), Zoltan Sumeghy ´ ¨
a, Judit Zoboki
baDepartment of Climatology and Landscape Ecology, UniÕersity of Szeged, PO Box 653, 6701 Szeged, Hungary
bDiÕision for Climatological Research and DeÕelopment, Hungarian Meteorological SerÕice, PO Box 38, 1525 Budapest, Hungary
Received 17 November 2000; accepted 8 May 2001
Abstract
This study examines the connection between the built-up urban surface and near-surface air
Ž .
temperature. The studied city Szeged, Hungary is located on a low and flat flood plain with a population of 160,000. Data were collected by mobile measurements under different weather conditions between March 1999 and February 2000. The efforts have been focussed on investigat-
Ž .
ing the maximum development of the urban heat island UHI along an urban cross-section.
According to the results, the UHI intensity changed according to season and month, as a consequence of the prevailing weather conditions. The role of cloudiness and wind speed on the temporal variation of the largest UHI, which represents the increasing effect of Szeged on temperature, is clearly recognized during most of the time in the studied period. The seasonal profiles follow remarkably well the general cross-section of the typical UHI described by Oke ŽOke, T.R., 1987. Boundary Layer Climates. Routledge, London who defines its characteristic. parts as ‘cliff’, ‘plateau’ and ‘peak’. The usefulness of the normalized values in the investigation is proved, the form of the seasonal mean UHI profile is independent of the seasonal climatological conditions, and is determined to a high degree by urban surface factors. As a conclusion, we suggest a modified model describing the metropolitan temperature variable for cities situated in simple geographical conditions: it is equal to the sum of components of the basic climate of the region and of the production of urbanization at the surface, where this last term is a multiplication of weather and urban surface factors.q2001 Elsevier Science B.V. All rights reserved.
Keywords: Urban heat island; Mobile measurements; Cross-section; Land-use features; Climatological parame- ters
)Corresponding author. Fax:q36-62-544-158.
Ž .
E-mail address: unger@geo.u-szeged.hu J. Unger .
0169-8095r01r$ - see front matterq2001 Elsevier Science B.V. All rights reserved.
Ž .
PII: S 0 1 6 9 - 8 0 9 5 0 1 0 0 0 8 7 - 4
1. Introduction
Human settlements modify the materials, the structure and the energy-balance of the surface and the composition of the atmosphere compared to the surrounding ‘natural’
terrains, although they also suffer from some effects of human economy e.g. agricul-Ž ture, forestry , mainly in the highly industrialized countries. These artificial factors. determine a distinguished local climate in the cities, which is the so-called urban climate. With the development of an urban climate, among other parameters, tempera-
Ž .
ture shows the most obvious alteration mainly an increase , which is manifested as an
Ž .
urban heat island UHI . Much research has been done on this phenomenon; therefore, it is impossible to enumerate the relevant bibliography within the scope of this study onlyŽ some subjective examples are given: Landsberg, 1981, pp. 83–126; Oke, 1982; Kuttler et al., 1996; Jauregui, 1997 ..
According to their geographical location, large cities in Hungary can be classified into three orographical types, namely: located in a valley, between mountains and a plain, and on a plain. The first two types are more complicated than the third one from the point of view of urban climate development, because in these types, the influence of artificial factors and the orographical conditions is very difficult to separate. As for the third city-type, detailed climatological measurements can provide a basis for the general deductions on urban climate.
Szeged is an important cultural and trade centre in Hungary, which is located in the
Ž .
southeastern part of the country 468N, 208E on a large and flat flood plain. The Tisza River is a narrow stream and according to our earlier investigation its influence is
Ž . Ž .
negligible Unger et al., 2000, 2001 . These orographical circumstances third city-type make Szeged a favourable place for urban climate investigations. According to Koppen’s¨ classification, the area belongs to the climatic region Cf temperate warm climate with aŽ fairly uniform annual distribution of precipitation . Within the administration district of. 281 km2, the number of inhabitants is 160,000. The basis of the city street network structure is a boulevard–avenue system.
The purpose of this paper is to reveal similarities or deviations in the spatial and
Ž .
temporal distributions of the maximum UHI intensity peak diurnal development by season along a numerously measured urban cross-section, to explain these features using land-use and climatological parameters and their inter-relationships.
2. Area and methodology
In the frame of a large research project, the efforts focussed only on the inner part of the administration district. The area of investigation was divided into two sectors Žnorthern and southern and subdivided into 500. =500 m square grid cells Fig. 1 . TheŽ . same grid size was employed in other urban climate analyses e.g. Jendritzky andŽ Nubler, 1981; Park, 1986 . The study area consists of 107 grid cells, total area 26.75¨ . km2, covering the urban and suburban parts of Szeged. The outlying parts of the city, mainly characterized by village and rural features, are not included in the grid network
Fig. 1. Built-up areas of Szeged, division of the study area into 500=500 m grid cells and the selected urban
Ž . Ž . Ž . Ž .
cross-section with numbers a built-up area, b open area, c border of the original study area.
except for four cells on the western side of the area. These cells are necessary to determine the temperature differences between urban and rural areas.
The measurements of air temperature were based on automobile traverses once a
Ž .
week during the period of March 1999–February 2000 altogether 48 times , which is a common process for detecting certain characteristics of the urban climate e.g. ConradsŽ and van der Hage, 1971; Oke and Fuggle, 1972; Johnson, 1985; Moreno-Garcia, 1994;
Klysik and Fortuniak, 1999 . This 1-week frequency of traverses provided sufficient. information under different weather conditions, except for rain. Table 1 summarizes the numbers of measurements by months and seasons. Temperature readings were obtained
Ž .
using a radiation-shielded resistance temperature sensor resolution 0.018C connected to a portable data logger for digital sampling. Since the data were collected every 16 s, at an average car speed of 20–30 km hy1 the average distance between measuring points was 89–133 m. The temperature sensor was mounted 0.60 m in front of the car at 1.45 m above the ground to avoid engine and exhaust heat. This is similar to the measure-
Ž .
ment system used by Ripley et al. 1996 in Saskatoon, Saskatchewan. The car speed
Table 1
Ž .
Monthly and seasonal numbers of mobile measurements in Szeged March 1999–February 2000
M A M J J A S O N D J F
3 4 4 5 4 5 4 4 3 4 4 4
Spring Summer Autumn Winter
11 14 11 12
was sufficient to provide adequate ventilation for the sensor to measure the momentary ambient air temperature. The traffic density in the late hours of measurements was rather
Ž .
low. The logged values at forced stops e.g. at traffic lights were rejected from the data set.
Having averaged the measurement values by cells, time adjustments to a reference time 4 h after sunset, which is the likely time of the occurrence of the strongest UHI inŽ the diurnal course in Szeged were applied. Urban–rural air temperature differences. ŽUHI intensity were determined by cells referring to the temperature value of cell 1. ŽTcell–TcellŽ1.., where the synoptic weather station of the Hungarian Meteorological
Ž .
Service is located, so this cell was regarded as rural Fig. 1 . The main, generalized parameters of land-use by grid cells were determined by GIS calculation of the Normalized Vegetation Index using SPOT XS images.
More detailed description of the study area and of the applied methods for the whole
Ž .
research project can be found in recently published papers Unger et al., 2000, 2001 . In this paper, only a small but very important part of the original study area is examined Žaltogether 10 cells . This part is an overlapping area between the above mentioned.
Ž .
northern and southern sectors Fig. 1 , namely, a cross-section in the urban area, which
Ž . Ž .
consists of cells stretching from the rural cell 1 in the west to the city core 10 . It
Ž . Ž .
measures a distance of 4.5 km between the centres of the first 1 and the last 10 cells.
The spatial and temporal profiles of the maximum UHI intensity along the cross-sec- tion was investigated by comparison of absolute and normalized seasonal means taking land-use and climatological features into consideration. The normalized values by cells are the ratios of the absolute means of a cell and of cell 10 where the UHI intensity isŽ
. Ž
the largest in all cases . Since meteorological conditions first of all, wind speed and
. Ž . Ž
cloudiness influence the magnitude in8C of the mean UHI intensity e.g. Landsberg, 1981, pp. 112–117; Park, 1986; Yague et al., 1991; Unger, 1996 , the seasonal¨ . comparison of spatial variation of UHI during the studied period is more effective using normalized values. Namely, the profile of the normalized mean UHI intensity is expected to be independent of the prevailing weather conditions in the studied period;
nevertheless, it is expected to be dependent of the urban surface factors e.g. land-useŽ features, distance from the city centre . Finally, a further investigation is focussed on the. relationship between temporal variations of the monthly mean largest value of maximum UHI intensity measured in the city and that of some climatological parameters windŽ speed and cloudiness ..
3. Seasonal patterns
Along the urban cross-section, three generalized land-use types were distinguished:
Ž .
covered surface built-up , which consists of streets, pavements, buildings, parking lots,
Ž . Ž .
etc., vegetated or bare areas open and water surfaces. Then the areal ratios % of these land-use types were determined by cells. One can see from Table 2 that the proportion
Ž .
of the water surface is rather negligible. The largest built-up density more than 90%
Ž .
can be found around the geometric centre of the city see cells 9 and 10 in Fig. 1 , but
Table 2
Generalized land-use characteristics by grid cells along the selected urban cross-section in Szeged Land-use type Cell number
1 2 3 4 5 6 7 8 9 10
Ž .
built-up % 0 0 18.9 70.4 54.2 85.6 71.7 77.8 91.4 90.5
Ž .
open % 100 100 81.1 23.5 45.3 11.1 28.3 22.2 8.6 9.1
Ž .
water % 0 0 0 6.1 0.5 3.3 0 0 0 0.4
the variation from the urban edge to the core is not uniform, it both increases and decreases.
Before the detailed seasonal analysis of the mean maximum UHI intensity, it can be seen at a first glance in Fig. 2 that the profiles in every season show a marked increase
Ž .
from a rural level immediately reaching the edge of built-up areas cell 3, see Table 2
Ž .
and the largest values are in the city centre cell 10 .
The absolute values of spring, summer and autumn profiles are almost the same in
Ž .
every cell point ranging between 08C and 3.028C , which is shown clearly by the fact that the maximum difference is only 0.218C between spring and summer values in cellŽ 5 . That is, the mean temperature along the cross-section varies closely together in these. three seasons. On the other hand, the values of the winter temperature profile do not reach even half of the values in other seasons. In winter, the largest mean maximum UHI intensity is only 1.448C. Because of the low winter values, the mean annual profile is a bit moderate compared to the spring, summer and autumn ones the largest intensityŽ is 2.588C ..
This seasonal magnitude variation of the urban temperature anomaly is attributed
Ž .
mainly to differences in the weather conditions Table 3 . The highest wind speed and
Fig. 2. Seasonal and annual profiles of the absolute mean maximum UHI intensity across Szeged MarchŽ 1999–February 2000 ..
Table 3
Monthly, seasonal and annual means of selected meteorological parameters in Szeged March 1999–FebruaryŽ 2000.
Ž y1. Ž .
Period Wind speed m s Cloudiness okta
M 3.5 4.7
A 3.5 4.9
M 2.6 4.3
J 2.5 4.6
J 2.6 3.8
A 2.1 3.6
S 2.7 3.6
O 2.9 4.5
N 2.6 5.9
D 3.4 5.6
J 3.5 6.1
F 3.4 3.7
Spring 3.2 4.6
Summer 2.4 4.0
Autumn 2.7 4.7
Winter 3.4 5.1
Year 2.9 4.6
cloudiness can be found in the winter months, at that time the UHI intensity is the weakest. In summer, when the above mentioned parameters have the smallest values, the extent of the UHI is not the largest in every point along the cross-section: it is slightly under or over the values in spring and autumn. It suggests that the effects of the climatological parameters are fairly complex and the investigation of their influence on the development of the UHI cannot be restricted to only two, although vital important, parameters.
Continuing the examination of the seasonal UHI profiles using normalized values ŽFig. 3 , we found that the differences between the seasonal profiles tend to diminish.. These patterns—with some local peculiarities—follow remarkably well the general
Ž .
cross-section of the typical UHI described by Oke 1987 who defines the characteristic parts of the profile as follows: ‘Cliff’ is a steep temperature gradient at the ruralrurban boundary, and much of the rest of the urban area appears as a ‘plateau’ of warm air with a steady but weaker horizontal gradient of increasing temperature towards the city centre. The urban core shows a final ‘peak’ to the heat island where the largest
Ž .
temperature difference is observed Fig. 4 .
Ž .
In Szeged, in all four seasons Fig. 3 the ‘cliff’ with a large temperature gradient is
Ž .
located between cells 2 and 4 a distance of 1 km . The profile revealed evidence of a 1.5-km-long thermal ‘plateau’ characterized by a very low temperature increment
Ž .
through its four cells from 4 to 7 . After the ‘plateau’ there is a second, very steep
Ž .
‘cliff’ between cells 7 and 8 0.5 km , which indicates the onset of the ‘peak’ region.
Ž . Ž
The areal extent of the largest more than 0.77 values are rather wide three cells, 1
. Ž .
km , so the prominence of the real ‘peak’ value of 1 is relatively less sharp. This is explicable by reference to the extent and homogeneity of the central urban part, which is
Fig. 3. Seasonal and annual profiles of the normalized mean maximum UHI intensity across Szeged MarchŽ 1999–February 2000 ..
dominated by 3–5-storey buildings built around 1900 and in the first decades of the 20th century after the so-called ‘Big Flood’ in 1879, which almost completely destroyed the earlier city. So, it can be established that the normalized mean maximum UHI along the cross-section varied rather strongly in the seasons and in the whole 1-year-period theŽ largest deviations are only around 0.13 ..
After the utilization of the normalized values to reveal the cross-section temperature distribution in an urban area, we can state that the form of seasonal mean UHI profile depends only on urban surface factors. Among them, the built-up ratio may not be the most important factor, because the steady, but not uniform increment of temperature
Ž .
Fig. 4. Generalized cross-section of the typical UHI after sunset modified after Oke, 1987 .
towards the city core does not exactly follow the built-up variations by cells see TableŽ 2 . However, another parameter, the distance from the city centre, seems to be more.
Ž .
dominant in this general increasing tendency of urban temperature Unger et al., 2000 . Naturally, the influence of these two urban parameters and of others possible, but notŽ investigated here: building density and height, sky view factor, etc. on the magnitude of. UHI intensity is not displayed one by one but in a complex manner.
Here, we have to touch upon the general dilemma on the surrogate estimate of urban effects on climate, which is, in general, comparing measured values inside the city with
Ž . Ž .
those at surrounding rural sites Oke, 1997 . Landsberg 1981, pp. 10–11 and Oke Ž1984 cite Lowry’s 1977 basic model, in which three elements influence the measured. Ž .
Ž .
value of a metropolitan variable, M for instance, temperature . The model is a sum of the three components:
MsCqLqU, Ž .1
where C is the measure of the basic climate of the region, L is a difference induced by
Ž .
the geographical location topography, bodies of water and U is the production of urbanization in surface land-use, material, geometry, building height and mass, positionŽ inside the city, etc. ..
Applying our results, we can specify and modify Eq. 1 for the metropolitan variableŽ . of temperature. If we omit the term L, which is a proper process in case of simple and plain terrain in and outside of a city, namely there is no complex geographical location Žlike Szeged , we get a simpler sum for the model:.
MsCqU. Ž .2
If M is temperature, the UHI is the urban heat excess, that is the production of
Ž .
urbanization UHIsU . In our case, we can mark the production of the urbanization in the cell n and in a given time or in a given season t by U . As our results on theŽ . n t seasonal mean normalized profiles of UHI suggest, the shape of a profile, which is
Ž .
represented by UN n sthe normalized value of the UHI in the cell n , is independent of the seasonal weather conditions and its variation determined by the urban surface factors Žun. in a high rank: UN nsf u . On the other hand, the absolute UHI intensity in a1Ž n.
Ž .
given urban place and in a given time Un t is a function not only of the urban surface
Ž . Ž . Ž .
factors un but of the weather situations in the region c , too: Ut n tsf c , u . That is,2 t n the influence of the surface factors by cells is weakened and strengthened by weather factors. Therefore, our suggestion for a modification of Eq. Ž .2 is that with the multiplication of components c and u , we can describe the real situation of thet n determination of U more exactly: Un n tsc u , and in generalt n
MsCqcu. Ž .3
Ž Ž ..
We emphasize that the modification Eq. 3 is valid in the case of a specific city-location for temperature but not valid by all means in the case of other meteorologi- cal variables.
4. Monthly variations
The last examination is directed to the connection between the annual variation of the mean largest value of maximum UHI intensity measured in the city and the variation of
Ž .
Fig. 5. Annual variations of the mean largest UHI intensity TcellŽ10.– TcellŽ1. measured at the maximum diurnal
Ž y1 .
development, monthly means of cloudiness and wind speed in8C, in okta and in m s , respectively in
Ž .
Szeged March 1999–February 2000 .
Ž .
some climatological parameters wind speed and cloudiness in the study area using monthly means. The largest UHI intensity is defined as the temperature difference
Ž . Ž .
between the central 10 and rural 1 cells, which represents the increasing effect of Szeged on temperature. The relationship between this UHI intensity and the above mentioned parameters is shown in Fig. 5.
The influence of cloudiness and wind speed can be clearly recognized on the monthly mean values in the period June 1999–February 2000 and in March 1999. In these months, the magnitudes of the UHI intensity and the climatological elements varied contrastedly, which support the earlier assumption on the negative role of cloudiness and wind speed on the development of UHI. In the cases of April and May, this negative relationship cannot be seen, which may be explained by the fact that the weather situations on the days selected for measurements were not entirely representative from the point of view of the average climatological conditions in these months. In addition, as mentioned in the previous section, wind speed and cloudiness are only two segments of those meteorological factors which affect the UHI intensity.
5. Conclusions
In this paper, the distribution of maximum urban–rural temperature difference and their seasonal variation along a cross-section in a Hungarian city were analyzed in the hopes of gaining some general deduction on urban climate. It was found that:
ØThe UHI effect in the period March 1999–February 2000 in the case of Szeged, which is a medium-sized city situated on a low and plain area, is evident.
ØThe heat island phenomenon always formed in the study area, even though the UHI intensity changed during the year, as a consequence of the prevailing weather condi-
tions, which vary significantly by seasons in case of a city located in the temperate climatic region.
ØThe role of cloudiness and wind speed on the temporal variation of the largest UHI, which represents the increasing effect of Szeged on temperature, is clearly recognized during most of the time in the studied 1-year-period.
ØThe seasonal profiles follow remarkably well the general cross-section of the
Ž .
typical UHI described by Oke 1987 who defines its characteristic parts as a ‘cliff’ at the ruralrurban boundary, as a ‘plateau’ of slightly increasing temperature towards the city centre and as a ‘peak’, where the largest temperature difference is found.
ØThe usefulness of the normalized values in the investigation of the cross-section temperature distribution in an urban area is proved. It came to light that the shape of the seasonal mean UHI profile is independent of the seasonal weather conditions and it is determined by the urban surface factors in a high rank.
Ž .
ØTherefore, we suggest a simple modified model describing the metropolitan
Ž .
temperature variable M for cities situated in special geographical conditions: M is
Ž . Ž
equal to the sum of components C basic climate of the region and U production of
. Ž
urbanization in surface , where Uscu multiplication of weather and urban surface factors ..
Acknowledgements
The research was supported by grants from the Hungarian Scientific Research Fund ŽOTKA Tr023042 and Tr034161 and the Ministry of Education FKFP-0001. Ž r2000 .. The authors are grateful to the Hungarian Meteorological Service for providing data on wind speed and cloudiness.
References
Conrads, L.A., van der Hage, J.C.H., 1971. A new method of air-temperature measurement in urban climatological studies. Atmos. Environ. 5, 629–635.
Jauregui, E., 1997. Heat island development in Mexico City. Atmos. Environ. 31, 3821–3831.
Jendritzky, G., Nubler, W., 1981. A model analysing the urban thermal environment in physiologically¨ significant terms. Arch. Meteorol., Geophys. Bioklimatol., Ser. B 29, 313–326.
Johnson, D.B., 1985. Urban modification of diurnal temperature cycles in Birmingham. J. Climatol. 5, 221–225.
Klysik, K., Fortuniak, K., 1999. Temporal and spatial characteristics of the urban heat island of Lodz, Poland.´ Atmos. Environ. 33, 3885–3895.
Kuttler, W., Barlag, A.-B., Rossmann, F., 1996. Study of the thermal structure of a town in a narrow valley.
Atmos. Environ. 30, 365–378.
Landsberg, H.E., 1981. The Urban Climate. Academic Press, New York.
Lowry, W.P., 1977. Empirical estimation of urban effects on climate: a problem analysis. J. Appl. Meteorol.
16, 129–135.
Moreno-Garcia, M.C., 1994. Intensity and form of the urban heat island in Barcelona. Int. J. Climatol. 14, 705–710.
Oke, T.R., 1982. The energetic basis of the urban heat island. Q. J. R. Meteorol. Soc. 108, 1–24.
Ž .
Oke, T.R., 1984. Methods in urban climatology. In: Kirchofer, W. et al. Eds. , Applied Climatology. Zurcher¨ Geogr. Schr., vol. 14, pp. 19–29.
Oke, T.R., 1987. Boundary Layer Climates. Routledge, London.
Ž .
Oke, T.R., 1997. Urban climates and global environmental change. In: Thompson, R.D., Perry, A. Eds. , Applied Climatology. Routledge, London, pp. 273–287.
Oke, T.R., Fuggle, R.F., 1972. Comparison of urbanrrural counter and net radiation at night. Bound. Lay.
Met. 2, 290–308.
Park, H.-S., 1986. Features of the heat island in Seoul and its surrounding cities. Atmos. Environ. 20, 1859–1866.
Ripley, E.A., Archibold, O.W., Bretell, D.L., 1996. Temporal and spatial temperature patterns in Saskatoon.
Weather 51, 398–405.
Unger, J., 1996. Heat island intensity with different meteorological conditions in a medium-sized town:
Szeged, Hungary. Theor. Appl. Climatol. 54, 147–151.
Unger, J., Bottyan, Z., Sumeghy, Z., Gulyas, 2000. Urban heat island development affected by urban surface´ ¨ ´ factors. Idojaras 104, 253–268.˜ ´ ´
Unger, J., Sumeghy, Z., Gulyas, A., Bottyan, Z., Mucsi, L., 2001. Land-use and climatological aspects of the¨ ´ ´ ´
Ž .
urban heat island. Meteorol. Appl. 8 in press .
Yague, C., Zurita, E., Martinez, A., 1991. Statistical analysis of the Madrid urban heat island. Atmos. Environ.¨ 25B, 327–332.
