• Monash home
• Arts home
• Schools & Centres
• Campuses
• About Arts
• Contact Arts

• Site map
• A-Z index
• Staff directory

Climate Group

Urban Climate Research

Overview

Data and sites

Links

Acknowledgements

Report

Project Overview:

The urban heat island (UHI) effect has been documented and researched for over a century (Howard 1833; Oke 1987). The UHI is a significant part of anthropogenic climate change studies (Karl et al. 1988) and its importance as a modifier of the urban environment has lead to ongoing measurement and modeling (Jauregui 1997; Hafner and Kidder 1999).

(http://eetd.lbl.gov/HeatIsland/HighTemps/UrbanProfile.gif)

There has been a simultaneous growth in the developed urban and suburban land area of Melbourne, which is expected to have increased the magnitude of the UHI and its spatial extent. The long-term urban warming component of UHI trends can be evaluated using differences between urban and rural climate stations. However, many cities, including Melbourne, do not have temperature records at a sufficiently fine spatial scale for mapping purposes. Instead, such spatial data are collected occasionally by direct field measurement or by inference from remotely sensed satellite data. It is important to understand the spatial details of local climate patterns and processes, because they combine complexly over time to determine the climate variability and trends measured at particular points. Therefore, the primary research goals of this study are (i) to quantify temporal urban warming trends in Melbourne over the last few decades, and (ii) to examine the spatial details of temperature over the Melbourne metropolitan area.

In addition to this scientifically important research, the project is also stimulated by the desire to integrate some real, if modest, research with undergraduate learning specifically as part of the Earth System Interactions: From biogeochemical cycles to Global change subject (GES3890/ATM3261). This innovate project reflects an international trend in science education (National Science Foundation 2001) and in the atmospheric sciences (Yarger et al. 2000). This project aims to provide a hands-on experience to include inquiry, the processes of science, and the excitement of research, which also develops communication and teamwork skills, critical thinking, and lifelong learning (National Science Foundation 2001). What were once separate research and teaching agendas are being combined into a common pursuit of learning. This view promotes the simultaneous and collaborative nature of learning about the world as a link between undergraduates, graduate students, and faculty members.

Acknowledgements:

This research was supported by funds from the Faculty of Arts (Monash University). Andrew Coutts and Cheryl Hickman have been instrumental in this project. The GES3890/ATM3261 class contributed substantially to the project.

Report:

THE URBAN HEAT ISLAND OF MELBOURNE

(Contributed by GES3890/ATM3261 students)

Abstract

The increase of urban temperature as compared to rural temperature is known as the Urban Heat Island effect (UHI) and is the result of the influence of urbanisation and landuse change on climate and its related systems. In Melbourne, the second largest city in Australia, it is expected that an Urban Heat Island exists. This paper presents results of automobile transect data and historical data from ground stations and satellite remote sensing, all of which indicate the presence of an Urban Heat Island in Melbourne and its surrounding suburbs.

Introduction

Examining the Urban Heat Island (UHI) in Melbourne.

For almost 200 years climatical differences between urban and rural environments have been recognised (Taha 1997), of which temperature is the most obvious (Unger et al. 2001). This phenomenon, known as the Urban Heat Island effect (UHI) has been well documented and found to be universally typical by many researchers (Shashua-Bar and Hoffman 2000, among many others). Of the many factors contributing to UHI, changes in surface physical characteristics (including geometry, thermal conductivity and wind speed) as well as the concentrated release of anthropogenic heat, are believed to be of great import (Stewart et al. 2000; Unger et al. 2001; Dousset and Gourmelon 2002; to name a few). Interactions between these and other factors are complex (Stewart et al. 2000) and as such, the influence of artificial (urban) factors and natural topographical conditions are difficult to separate (Unger et al. 2001).

Melbourne, the second largest city in Australia, has a population of 3.2 million and is one of the oldest urban regions in the southern hemisphere (ABS 2003). In the five years between 1996 and 2001, the city's population grew by 7.2%, with at least half of that growth resulting from outer suburb development and increases in medium and high density housing (ABS 2003). Melbourne's urban/rural boundary is characterised by the 'spread' of suburbs and industry into the surrounding landscape. Many different land use types are present. The city enfoldsPort Phillip Bay, with the commercial centre on the northern banks of the Yarra River (Geoscience Australia 2002). The whole area is part of a broad open basin bordered by the Dandenong ranges to the east and the Western Plains district in the west (Geoscience Australia 2002). Melbourne sits 31.2m ABSL at 37.81°S, 144.97°E, with an annual rainfall of 657.3mrn/year (Australian Bureau of Meteorology -ABOM, 2001). Average daily maximum temperature, taken from the Melbourne Regional office is 19.8 oC. Melbourne has a four-season basic weather cycle with winter/spring being the wettest months (mid-year), but with annual rainfall usually not varying by more than 10mm per month (ABOM 2001).

In this study, the presence or absence of a UHI in Melbourne is to be evaluated with the use of data obtained through automobile transects, satellite remote-sensing and accumulated data sourced from multiple ground stations of the Australian Bureau of Meteorology.

Methods: Data Collection and Analysis

Automobile transects were the main focus of temperature data collection in this study. Six transect routes were chosen to give wide coverage of the city; Melton in the South-West, Melbourne Airport in the North-West, Cragieburne in the North, Upper Ferntree Gully in the East, Dandenong in the South-East and Carrum in the South-South-East along the bay.

Data were collected by driving along the transect line with a Vaisala HMP45C temperature and relative humidity probe attached to a side window of the vehicle. A Garmin GPS and electronic data logger were attached to the temperature probe and were used to record temperature, elevation, speed and relative humidity at intervals of 20-seconds. Consequently, the data points were recorded at irregular spatial intervals, but allow a relatively dense sample set in a short time. Contemporaneous satellite data were to be obtained for greater spatial coverage, but due to information rationing (US Wartime Protocol) only historic data (2001) were available.

Two automobile sample sets were taken: a daytime set (begun at 1130, 25 March 2003) and a night-time set (begun at 0100, 26 March 2003). Data were normalised to eliminate non-UHI-related variation in temperature over the course of sampling.

Historical data were used to further the critical evaluation of transect data, and to provide an historical context to the study. Data were obtained from Australian Bureau of Meteorology archives for relevant weather stations (1965-2001) and the USA/Japan ASTER satellite (2001, Tan and Hook, 2001). Temperature data have been presented as spatial cross-sections and temporal charts to allow more effective evaluation.

All the temperature data from the automobile transects showed evidence of an urban heat island. Normalised data showed lower temperatures over all, however both sets of data clearly showed the effects of an Urban Heat Island (UHI). For the purposes of this analysis, normalised data were used.

Results and Discussion

The UHI effect is manifest by increased temperature existing in business and urbanised areas of land when compared to less well-populated or rural areas. Data collected on March 25, 2003 showed elevated temperature in the CBD of Melbourne and an overall trend towards lower temperature in the 'rural' centres chosen as end points for the transects. The UHI effect appeared to be stronger at night, which is in keeping with Roth's (Roth et al. 1989) satellite-based um evaluation. Satellite data available to the researcher measured land surface temperature and showed UHI presence (Figure 1), but it was less pronounced that in other data sets. Australian Bureau of Meteorology data showed evidence of um since 1965 and an overall increase in temperature and intensity.

Figure1: An example of data from satellite land surface temperature readings: Bayside (Frankston): Short Wave Infrared Spectral Analysis.

Graphs of spatial temperature changes obtained from vehicle transect data varied in addition to the overall UHI trend. UHI are largely a result of landuse change (urbanisation) (Saaroni et al. 2000) and temperature variation, even on a small scale, is closely linked to landuse characteristics. Landuse type was recorded manually at the same time as temperature data was being recorded. Some ambiguities were encountered when trying to reconcile temperature and landuse records as a result of the former being recorded as a value at distance from the CBD and the later at a distance from the start of the transect.

To overcome this problem, a detailed local map was consulted ('Melways') thus the two data sets were aligned. Once the alignment had been made, it was clear that landuse variation accounted for temperature variation even on a small scale.

Figure 2: Automobile transect data for daytime temperatures in Frankston, Dandenong and Ferntree Gully Transects.

Beginning in the CBD, the Frankston transect (Figure 2) shows an initial increase in temperature (peak) followed by a sharp drop at the approximate location of AlbertPark Lake -a substantial lake and park area just outside the Melbourne CBD. There is a subsequent increase to a plateau occurring in conjunction with residential suburban land use. A steep drop in temperature, or cliff (Oke 1987, in Unger 2000), is then seen as the sampling road follows the bay.

The end of the Frankston transect (Frankston side -35km away from the Melbourne CBD) coincides with a more developed commercial area (the Frankston City CBD) as is indicated by a consequent rise in temperature. The overall lower temperatures shown in the Frankston data are concurrent with an increased coolness usually found in coastal areas (Roth et al.1989), Landuse influence is evident in the night-time data set however, it is less pronounced (Figure 3).

Figure 3: Automobile transect data for night-time temperatures: Frankston, Ferntree Gully and Melton.

Dandenong transect data (Figure 2) show a similar relationship between landuse and temperature to that found in the Frankston data set. There were lower temperatures where the transect passed parks (~5km from CBD), higher temperatures in commercial, CBD and industrial areas ( 22km), and medium level temperatures in suburban residential areas (12-22km), as verified through mapped landuse infomation. As with the Frankston transect, the final data collection points for the Dandenong line were in the Dandenong CBD and were marked by a small increase in temperature.

The Fern tree Gully transect (Figures 2 & 3) showed the most localised variation out of all transects measured. When landuse types were consulted, it became clear that temperature was very susceptible to localised landuse variation Parks and factories, often alternating with each other, seem to account for the 'see- saw' appearance of temperature data. The influence of parks and other 'small green sites' has also been indicated by Shashua-Bar and Hoffman (2000) and also in Akbari's study (2002) of the role of urban trees in reducing energy emissions and carbon combustion

In Los Angels, researchers have predicted that an increase in trees will lower temperature (and thus the UHI) and when used as shade for buildings, by directly lowering overall temperatures, trees will indirectly reduce the need for air-conditioning (Rosenfeld et al 1996). Increases in temperature result in increased rates of ozone formation (Duefias et al. 2002), which is another aspect of anthropogenic climate modification. Saito et al. (1990) attribute most of the 'heat island problem' to reduced tree density in urban areas. The inference is that the um effect can be mitigated, or at least reduced by strategic placement of green spaces within the urban environment.

Much other research has recognised the importance of surface geometry in determining the extent of UBI formation (Saaroni et al. 2000: Stewart 2000: Unger et al. 2001: Dousset and Gourmelon 2002)

Canyon geometry is responsible for changes in radiative fluxes within the urban setting (Dousset and Gourmelon 2002), resulting in overall reduced losses of long wave radiation and the multiple reflection of short wave radiation (Santamouris et at. 2001 ). Simply, anthropogenic materials have increased storage and absorption of sensible heat, so much so that in some cases a daytime Urban Heat Sink is present (Camahan and Larson 1990)

Data from the Australian Bureau of Meteorology

Temporal data, for the same day as the automobile transect data. were collected from the Australian Bureau of Meteorology. These data showed variation in temperature over the course of a day (pm March 23 -am March 25, 2003) and Relative Humidity values for the same time period. Similarities in relations between humidity and temperature at the rural and urban sites uphold Stewarts' (2000) findings that change in humidity do not greatly influence UHI intensity.

Figure 4: Diurnal variation in temperature and relative humidity: March 23-25, 2003.

A graph of theMelbourne data (Figure 4) has been included as it is representative of the patterns shown for Met. Stations with the same location as automobile transect end-points (Frankston, Airport, etc)., Over the course of the day, with a peak in temperature at around 1700hrs and a low at around 0500hrs, the Relative Humidity displayed an inverse relationship with temperature. The maximum level of humidity may be attributed to dew accumulation at ground level during the early morning and minimum relative humidity to evaporation that would occur during the hottest part of the day. Interestingly, the design of urban environments is aimed at removing surface water with maximum efficiency (Dousset and Gourmelon 2002). the result of this, coupled with the general lack of green spaces, means a decrease in surface moisture available for evaporation and a subsequent increase in latent heat (Santamouris et al. 2001). While relative humidity does not appear to influence UBI. Duefias et al. (2002) found that high levels of relative humidity favoured the production of ozone, as do high temperatures and wind speeds.

Figure 5: Wind speed variation, 23-25 March 2003: Australian Bureau of Meteorology.

Wind speeds were lower during the night-time transect than during the daytime transect (Figure 5). Wind speeds, however, have been found to account for 20% less variation in nocturnal heat island occurrence than does cloud cover (Stewart 2000).

Historical Data

Average maximum daily temperatures of Rural and Urban MET station data since 1965 show the presence of a UHI effect (Figure 6). Maximum urban temperature is shown to be rising at a higher rate than maximum rural temperature (Figure 6), but both show an overall increase over time, as is consistent with many other urban areas (Dixon et at. 2003, Among others). In average minimum temperatures (Figure 7), the difference between rural and urban sites was more pronounced: this is to be expected since the minimum temperature will usually reflect night-time temperature, which is when a greater UBI effect is usually observed (Roth et al. 1989: Oke et a1. 1991 ).

Figure 6: Average Melbourne Rural/Urban maximum temperature since 1965.

Both night-time and daytime temperatures showed an overall increase in temperature with time, and a greater increase in urban temperature than in rural temperature. Greater increases in urban temperature may be a function of increasing urbanisation (BOM 2003).

Figure 7: Average Melbourne Rural/Urban temperatures since 1965.

The ABOM data also suggest that the urban heat island is more pronounced in winter than in summer (Figure 8). This interpretation is supported by the presence of increased anthropogenic heat sources in urban settings during winter (Streutker 2003). In winter, the UHI can decrease the heating loads and help to make temperatures more bearable for humans (Taha 1997). In summer the UHI effect can lead to uncomfortable conditions and health risks for the elderly, infirm and very young and markedly increases cooling loads (Santomouris et al. 2001). Urban planning and management strategies are increasingly taking into account the influence of UHI on the different climatic and anthropogenic systems present in an urban environment (Rosenfeld et at. 1996: Gomez et at. 1998: Lazar and Podesser 1999).

Figure 8: Comparisons of urban and rural temperature in summer (January) and winter (June), in Melbourne: Australian Bureau of Meteorology Data 1965-2001.

Conclusions

The Urban Heat Island phenomenon is widely recognised across the globe as a result of changed land surface characteristics caused by anthropogenic activity. This paper concluded that a UHI effect was present in Melbourne in autumn 2003 (March) and is evidenced through historical data as early as 1965. The UBI topic as such is not original, in that it ratifies previous work in the field, however the orogenic setting of most cities and urban centres varies, and within this variability may lie information that could not be found elsewhere and is of relevance to issues of climate management. Issues arising from the mu are almost as complex as interactions between factors contributing to its extent. Urban planning, management, energy consumption and pollution are all potential applications of any information gleaned through studies of the UHI effect. Increased understanding of the UHI phenomenon and ways in which to manage it should be an integral part of government policy formation. As a direct result of human influence on the environment, the urban heat island effect now significantly influences anthropogenic activity and should be considered in planning new urban centres, ameliorating existing cities, and in general climatological evaluation.

References:

Akbari H. (2002) Shade trees reduce building energy use and CO2 emissions from power plants.Environmental Pollution 116(Supplement-1), S 119-S 126.

Australian Bureau of Meteorology: (http:l /www .bom.gov .au/climate/averages/tables/cw_086071.shtml). Averages for Melbourne Regional Office (1855-2001). Accessed 3 April 2003.

Australian Bureau of Statistics: (http://www .abs.gov .auf Ausstats/abs@ .nsf).. Media release: Australian Bureau of Statistics: 2001 Census social Atlas For Melbourne. Last Updated 11 February 2003, accessed 3 April 2003.

Camahan, W.H., and Larson, R.C (1990) An analysis of an urban heat sink. Remote Sensing Environment 33(1): 65-71

Dixon W., Delworth T.L., Knutson T.R., Spelman M.j., and Stouffer R.j. (2003) A comparison of climate change simulations produced by two GDFL coupled climate models. Global and Planetary Change 37(1-2),81-102.

Dousset, B., and Gourmelon, F (2002) Satellite multi-sensor data analysis of urban surface temperatures and landcover. ISPRS Journal of Photogrammetry and Remote Sensing, 58(1-2), 43-54.

Duefias C., Femtindez, Cafiete S., Carrentero I. and Liger E (2002) Assessment of ozone variations and meteorological effects in an urban area in the Mediterranean Coast, The Science of Total Environment 299 (1-3),97-113.

Geoscience Australia: (www .auslig.gov .aulfacts/capital/v .htm). Geoscience Australia, Melbourne, Victoria. Last updated 17 October 2002. Accessed 3 April 2003.

Gomez F., Gaja E., and Reig A. (1998) Vegetation and climatic changes in a city, Ecological Engineering 10(4), 355-360.

Lazar R. and Podesser A (1999) An urban analysis of Graz and its significance for urban planning in the tributary valleys east of Graz (Austria), Atmospheric Environment 33(24-25),4195-4209.

Oke T.R., Johnson D.G., Steyn D.G. and Watson I. D. (1991) Simulation of surface urban heat island under ideal conditions at night -Part 2: Diagnosis and causation. Boundary Layer Meteorology 56, 339-358.

Oke, T.R (1987) Boundary Layer Climates, Routledge, London.

Rosenfeld A.H., Romm J.J., Akari H., Pomerantz M. and Taha H.G. (1996) Policies to reduce heat islands: magnitudes of benefits and incentives to achieve them. Proceeding of 1996 ACEEE Summer Study on Energy Efficiency in Buildings, Pacific Cove. CA. USA Vol 9, 177.

Rosenfeld A.H., Romm J.J., Akari H., Pomerantz M. and Taha H.G. (1996) Policies to reduce heat islands: magnitudes of benefits and incentives to achieve them. Proceedings of 1996 ACEEE Summer Study on Energy Efficiency in Buildings, Pacific Cove, Ca, USA, Vo19, 177.

Roth M., Oke, T.R., and Emery W.J (1989) Satellite-derived urban heat islands from three coastal cites and the utilization of such data in urban climatology. International Journal of Remote Sensing 10(11): 1699-1720.

Saaroni, H., Ben-Dor, E., Bitan, A., and Potchter, O (200) Spatial distribution and microscale characteristics of the urban heat island in Tel-A viv , Israel. Landscape and Urban Planning 48 (1-2):1-18.

Saito I., Ishimara 0. and Katayama T (1990) Study of the effect of green areas on the thermal environment in an urban area. Journal of Energy and Buildings 15-16, 445- 446.

Santamouris M., Papanikolaou N., Livada I., Koronakis I., Georakis C., Argiriou A and Assimakopopoulos D.N. (2001) On the impact of urban climate on the energy consumption of buildings, Solar Energy 70(3),201-216.

Shashua-Bar, L., and Hoffman, M.E (2000) Vegetation as a climatic component in the design of an urban street; an empirical model for predicting the cooling effect of urban green areas with trees. Energy and Buildings 31(3): 221-235.

Stewart I.D. (2000) Influence of meteorological conditions on the intensity of the urban heat island effect in Regina. Canadian Geographer- Geographe Canadien 44(3),271-285.

Streutker D.R. (2003) Satellite-measured growth of the urban heat island of Houston, Texas. Remote Sensing of Environment, 85(3), 282-289.

Taha, H ( 1997) Urban climates and heat islands: albedo, evapotranspiration and anthropogenic heat. Energy and Buildings 25(2): 99-103.

Tan H.L. and Hook S.J: California Institute of Technology, accessed 17.04.2003, Last modified 16.10.2001 at htt ://asterweb.. I.nasa. ov/instrument/band.htm).

Unger, J., Si.imeghy., and Zoboki, J (2001) Temperature cross-section features in an urban area.Atmospheric Research 58(2): 117-127.

GES Home

• Climatology Home
• Research
• Current Weather
• Courses
• Field Trips
• Honours & Postgrad
• Resources
• Links
• People

• GES Research

Copyright © 2008 Monash University ABN 12 377 614 012 - Caution - Privacy - CRICOS Provider Number: 00008C
Last updated: 29 October 2008 - Maintained by arts.webmaster@arts.monash.edu.au - Accessibility information