- Email: [email protected]
Urban heat island and its impact on climate change resilience in a shrinking city: The case of Glasgow, UK
Building and Environment 53 (2012) 137e149
Contents lists available at SciVerse ScienceDirect
Building and Environment journal homepage: www.elsevier.com/locate/buildenv
Urban heat island and its impact on climate change resilience in a shrinking city: The case of Glasgow, UK Rohinton Emmanuel*, Eduardo Krüger School of Engineering and Built Environment, Glasgow Caledonian University, 70 Cowcaddens Road, Glasgow G4 0BA, UK
a r t i c l e i n f o
a b s t r a c t
Article history: Received 20 July 2011 Received in revised form 13 December 2011 Accepted 25 January 2012
Given its long urban history of growth and decline, Glasgow, UK, provides a historically significant opportunity to study the local climatic changes brought about by urban variables. This study investigates the changes in air temperature within the central area of Glasgow using three data sources: the UK Meteorological Office historical data for Glasgow (climate normals and running data for a 50-year period), the Weather Underground network; MIDAS Surface Weather Stations network of the British Atmospheric Data Centre (BADC). Three approaches were used to evaluate Glasgow’s local climate change: assessment of mean air temperature increases based on two concurrent climate normals, traditional UHI approach (i.e. differences between a ‘rural’ and an ‘urban’ site) and observed temperatures in locations with different land cover characteristics (using the Local Climate Zone LCZ concept. Planning and building scale implications for other shrinking cities are explored. Ó 2012 Elsevier Ltd. All rights reserved.
Keywords: Urban climate Urban heat island Local climate zone Climate change Shrinking cities
1. Introduction Cities are home to over half of the global population [1]. The current growth trajectories indicate that by 2050, nearly 70% of humanity (6.3 billion out of an estimated global population of 9.1 billion) will live in urban areas [1]. The local, regional and global climate implications of such rapid urban growth are complex: On the one hand, cities are major consumers of energy and materials as well as generate vast amounts of waste. They are also major centres of innovation and finance e which could enable them to be at the forefront of climate mitigation actions. At the same time, the nature of climate change with its long lag-times mean that cities need to prepare themselves now, to act as the first line of defence against catastrophic effects of local and global changes expected in the near future. Given the twin realities of potential innovation in mitigatory action as well as adaptive capacity and need in the face of burgeoning urban population, urban areas are beginning to receive a long overdue attention from climate change scientists and policy makers [2,3]. In these efforts, the role of urban climate change in both contributing to and augmenting global change is a key unknown that needs careful attention. Could there be opportunities to use the mitigatory potential of urban fabric to reduce the effects of urban warming with a view to providing some relief to the wider
* Corresponding author. Tel.: þ44 (0) 141 331 3217. E-mail address: [email protected] (R. Emmanuel). 0360-1323/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.buildenv.2012.01.020
and more extreme weather events spawned by global climate change? A further reality confounding the issue is urban decay. Major centres of urban population have already begun to lose population and even those that are yet growing continue to sprawl, leading to lowering urban densities in many parts of the world. What effects will shrinking cities as well as de-densification have on local climate change? What lessons can we learn from the local climate change trajectories of mature cities to plan better the still growing cities of the world? In this paper, we analyse the trends and local differences in air temperature in and around the mature urban area of Glasgow, UK (55 510 N, 04120 W). Using historical weather data from three different sources we determine the effect of local climate zones [4] on microclimate and postulate what lessons could be learnt by other mature and/or shrinking cities and how the city of Glasgow could benefit from these changes in its quest for a sustainable and low carbon future. 2. Background 2.1. Urban heat island in mature cities Urban warming and its links to regional and global warming have been well documented in several mature cities in Japan, Europe (especially in the UK) North America and Sweden. Fujibe’s work in Japan [5,6] epitomises these efforts. Reviewing data from
138
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
561 meteorological stations established in Japanese cities in the early 20th Century, Fujibe [6] found that there is a warming trend of 0.3e0.4 C/decade even for locations with low population density (<100 people per square kilometre), indicating that the recent temperature increase is largely contributed by background climate change. However, an anomalous warming trend is detected for stations with larger population density (in the order of 100e300/ km2) where the anomalous trend is 0.03e0.05 C/decade [6]. Furthermore the recorded rate of temperature increase tends to be larger at night than during the daytime, although in the case of a megacity (Tokyo) widespread urban warming in the hinterland during afternoons of the warm season is seen as a result of extensive urbanization that enhances daytime surface heating. Among the European mature cities, the heat island effect in London is one of the most well studied (see Ref. [7] for a comprehensive review). On the positive side, heating energy consumption in central London is 65e85% of the heating required for the same building based outside the Urban Heat Island [8]; conversely, cooling energy consumption is 32e42% higher in the city. Given the fuel mix for heating and cooling (gas for heating and coal for electricity, [8]), the carbon implications of UHI to London are negative. An estimate of the heat island effect in major population centres in the UK (based on the UK Met Office historical weather data, see UK Met Office, 2011) was developed by Kershaw et al. [9]. Table 1 shows the seasonal and annual average UHIs for several UK cities. The heat island effect on outdoor thermal comfort at a mature city (Göthenburg, Sweden) was studied by Thorsson et al. [10]. Statistically downscaling climate change projections to the city street-levels, Thorsson et al. [10] showed that urban geometry could cause large intra-urban differences in Mean Radiant Temperature (Tmrt), on hourly, daytime and yearly time scales. In general, open areas are warmer than adjacent narrow street canyons in summer, but cooler in winter. The combination of regional warming coupled with augmentation by urban geometries will triple strong/extreme heat stress (to approx. 20e100 h a year, depending on geometry). Conversely, the number of hours with
Table 1 Seasonal and annual average UHI values ( C) for UK cities. City
Winter (DJF)
Spring (MAM)
Summer (JJA)
Autumn (SON)
Annual
New castle upon-Tyne Portsmouth Central London Liverpool Glasgow Edinburgh London Suburbs Plymouth Sheffield Manchester Bristol Middlesbrough York Cardiff Leeds-Bradford Nottingham Bournemouth Birmingham Coventry Belfast Leicester
1.8 1.3 1.3 1.1 1.0 1.1 0.9 0.9 0.9 0.8 0.6 0.8 0.8 0.5 0.6 0.5 0.6 0.4 0.4 0.4 0.0
1.9 1.73 1.7 1.4 1.5 1.4 1.2 1.2 0.9 1.1 1.0 1.0 0.8 0.9 0.7 0.7 0.8 0.5 0.5 0.5 0.2
2.0 2.0 1.9 1.7 1.5 1.4 1.4 1.2 1.1 1.2 1.2 0.9 1.0 1.1 0.8 0.9 0.8 0.7 0.7 0.3 0.1
1.8 2.0 1.6 1.5 1.2 1.3 1.1 1.1 0.9 0.8 0.9 0.8 0.9 0.6 0.8 0.6 0.5 0.6 0.4 0.3 0.1
1.9 1.8 1.6 1.4 1.3 1.3 1.1 1.1 1.0 0.9 0.9 0.9 0.9 0.8 0.7 0.7 0.7 0.6 0.5 0.4 0.1
Note: UHI values are given as three-month average temperature difference between a city centre weather station and a nearby rural station. Winter ¼ December, January and February (DJF); Spring ¼ March, April and May (MAM); Summer ¼ June, July and August (JJA); Autumn ¼ September, October and November (SON). Source: Based on Kershaw et al., 2010.
strong/extreme cold stress will decrease by 400e450 h. Furthermore, the number of hours with no thermal stress will increase by 40e200 h a year. Considering both winter warming and summer cooling potential of judicious arrangement of urban geometry, Thorsson et al. [10] conclude that a densely built urban structure will mitigate extreme swings in Tmrt and in the generally adopted thermal comfort index ‘physiologically equivalent temperature’ (PET), improving outdoor comfort conditions both in summer and in winter. 2.2. Climate change and urban heat island Given the small fraction of land occupied by cities (all urban sites including green as well as built-up areas cover only 2.8 per cent of the Earth’s land area [11]) it is unlikely that cities have a direct bearing on global climate change. However, cities indirectly drive global climate change on account of their insatiable appetite for energy and material (and associated waste and pollution). Furthermore, an increasing urban population will lead to the expansion of land covered by cities, at which point the direct influence of cities on regional and perhaps global climate may not be insignificant. There is increasing evidence to the scale of urban influence on the global climate to be in the order of El-Nino Southern Oscillation (ENSO) [12]. The significance of the signal from urban climate change on the global climate led Hansen et al. [13] to term it as “urban warming.” A key difficulty in untangling the urban warming from global climate changes is the computational and parametric difficulties associated with representing urban areas in climate models. Yet it is increasingly recognised that climate models provide a good approach for assessing the global consequences of the urban climate modifications [14]. Given the lack of detailed land cover information and computing power, most climate models generally do not include a representation of urban areas, and therefore their climate projections are likely to underestimate the heat island phenomenon [9]. The general consensus appears to be that the UHIs are likely to augment the temperature anomaly arising from global warming [2,9]. A recent study commissioned by the World Bank shows that modern patterns of city growth are increasingly land intensive [15]. Average urban densities have been declining for the past two centuries. As transportation continues to improve, the tendency is for cities to use up more and more land per person [15]. In developing countries, cities of 100,000 or more are expected to triple their built-up land area to 600,000 km2 in the first three decades of this century. Cities in developed countries expand at an even faster rate per resident, despite their smaller population size and lower rate of population growth. They will increase their built-up land area by 2.5 times between 2000 and 2030. At that point, they will occupy some 500,000 km2 [15]. These may have direct consequences to global climate. Even if the urban effects on global climate change remain weak, the influence of UHIs on regional climates is clearer. Lamptey [16] attempted to untangle the sensible and latent heat partitioning associated with different land cover classes in Chester County and surroundings near Philadelphia, Pennsylvania, USA. The urban effect became more important as the fraction of urban land cover to the total increased. When the urban land cover increased from 11% to 19% in 9 years it led to the largest proportionate sensible (21.4 Wm2) and latent (14.2 Wm2) heat fluxes during winter. During summer, urban and vegetation land cover produced the largest proportionate sensible heat (59.2 Wm2) while urban land cover produced the second largest proportionate latent heat flux (39.5 Wm2). The regional climatic implications of these energy partitioning cannot be ignored.
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
139
2.3. Urban evolution of Glasgow
Fig. 1. Glasgow’s population trend during the last 200 years. Source: Population data for 1801e1931 from: www.histpop.org.uk. Data for 1951e2001 from the decennial census records at the National Records of Scotland (there was no census taken in 1941).
Given this increasing body of evidence, the role of urban climate change as a significant part of the human experience of climate change is increasingly being recognised. While pointing out the ‘minor impact on estimates of global trends of land surface-air temperature (LSAT), Parker [17] recognises the combined effect of global and urban warming on human health and welfare. Fujibe [5] showed the differential nature of warming experienced by Japanese cities of different sizes and densities: cities with over 3000 persons/ km2 or urban land cover over 50% of the total land, showed a warming trend of 0.1 C/decade, compared to the centennial global trend estimated by the IPCC [18] of 0.74 C/century for 1906e2005.
From its medieval ecclesiastical origins Glasgow expanded into a major port in the 18th century and with the advent of industrial revolution, added a massive industrial base to its already well developed built fabric. This industrial base was responsible for the intense concentrations of people, wealth, poverty, and eventual decline of the city. The industrial heritage of Glasgow gave it both the notoriety for pollution as well as fame for its attempt to clean it, especially via the pioneering creation of urban parkland [19]. These parks and the events associated with their creation (International Exhibitions of 1888 and 1901, the Scottish Exhibition of National History, Art and Industry in Kelvingrove Park in 1911, the Empire Exhibition in Bellahoustoun Park, south of the river Clyde, in 1938) gave the city its claim to fame as the ‘second city of the empire’ [20]. The urban and industrial growth of Glasgow has been rapid in the 19th Century, drawing people from all over the UK and beyond [21]. However, the success of its industrial base could not withstand the pressures of globalisation, and by the early 20th Century, the city had begun to lose population. Fig. 1 shows the population trend in the city during the last 200 years. Nevertheless, the provision of urban infrastructure and services, much of which remains to this day, means that the land cover changes induced by urbanisation are with us to this day, despite the decline in population. This has implications for the city’s urban climate and we will return to this aspect of our study in the ‘implications’ section of the present paper. 3. Methods In the present study we performed three types of comparisons to elicit the variations in local climate induced by urban growth and decay: (1) historical trend analysis, (2) pairwise comparison of an
Fig. 2. Location of weather stations used by the present study. Key: Paisley e yellow; MIDAS Stations e red; Weather Underground stations e black. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
140
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
Fig. 3. Illustration of land cover characteristics at different Local Climate Zones (LCZ). Source: Stewart and Oke, 2009.
‘urban’ and ‘rural’ weather stations and (3) correlation between local climate and land use/land cover variations in and around Glasgow. According to Koeppen-Geiger’s climate classification, the region is characterised by a Cfb temperate climate type, especially made mild due to stronger maritime influences. With significant precipitation in all seasons (average annual precipitation in Glasgow is 1100 mm), the average maximum temperatures in the warmest months (July and August) remain below 20 C, but with at least four months averaging above 10 C [22]. For the purpose of historical trend analysis we used the data from the Glasgow International Airport weather station, located in Paisley (Lat: 55 520 N, 04 260 W, alt: 19 m asl), available from the UK Meteorological Office (http://www.metoffice.gov.uk). In addition to the running data for the period of 1959e2009 we also compared the two recent ‘climate normals’ used as reference periods during the last 50 years (1961e1990 and 1971e2000). These two periods correspond to the last two climate normals used by the UK Met Office in reporting average climates. Our pairwise comparison utilised data from the Met Office Integrated Data Archive (MIDAS) Surface Weather Stations network, provided by the British Atmospheric Data Centre (BADC) registered at a ‘rural’ and at an ‘urban’ location (http://badc.nerc.ac. uk, last accessed 30 Jun 2011): Rural e Springburn (Lat: 55 530 25.4400 N, Lon: 4 130 3000 W, alt: 107 m asl); ‘urban’ e Glasgow Weather Centre, situated near the Glasgow Central Station (Lat: 55
510 36.3600 N, Lon: 4 130 33.7800 W, alt: 40 m asl) (Fig. 2). Coincidental data for the period of 1974e1985 were used for such comparison. The Springburn station was situated near the Springburn Park, which continues to remain a green area on the northern fringes of the city. The Glasgow Weather Centre was located at the city core, two street blocks westwards from the Central Railway Station, in an area characterised by midrise historical buildings (7e8 storeys). Springburn Station started operation in January 1894 through March 1997 while the Glasgow Weather Centre Station was in operation from January 1974 until December 1985. According to the Met Office Surface Data Users Guide (http://badc.nerc.ac.uk/data/ surface/ukmo_guide.html#2, accessed 1 December 2011), equipment used in both cases regularly underwent a series of calibrations which linked the instrument to a national or international standard instrument. This practice, which for temperature and pressure has been unbroken since 1851, ensures a uniformity of measurement over time. On both stations, maximum and minimum temperature data were provided for a 24 h cycle starting at 9 a.m., using either a liquid-in-glass thermometer or, from the 1980s electrical resistance thermometers (ERT). It should be noted that there is a height difference between stations (67 m), which could partly explain the lower temperatures measured at Springburn. Although this is a limitation of the study, correcting local temperature to account for elevation will introduce errors and the process is not straightforward. The “environmental lapse rate”
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
141
Table 2 Local Climate Zone (LCZ) classification system Source: Based on Stweart and Oke [4]. Local climate zone (LZC)
Compact highrise Open-set highrise Compact midrise Open-set midrise Compact lowrise Open-set lowrise Dispersed lowrise Lightweight lowrise Extensive lowrise Industrial processing
Zone properties
Jsky
H:W
SF
ZH
RC
a
m (J m2 s½ K1)
QF (W m2)
0.25e0.45 0.40e0.70 0.30e0.60 0.80e0.90 0.30e0.50 0.55e0.75 >0.90 0.30e0.50 >0.90 0.70e0.90
>2 0.75e1.25 0.75e1.25 0.20e0.30 1.00e1.50 0.50e0.75 0.10e0.20 1.00e1.50 <0.25 0.2e0.5
>90% 50e75% >90% 30e50% >80% 45e65% 20e30% 70e90% >80% 45e65
>35 m >30 m 15e25 m 10-25 m 3e10 m 3e10 m 3e7 m 2e4 m 3e10 m 5e10 m
8 7e8 6e7 5e6 6 5e6 5e6 4e5 5 5e6
0.12e0.18 0.12e0.20 0.15e0.20 0.15e0.20 0.12e0.20 0.10e0.20 0.10e0.20 0.10e0.20 0.15e0.25 0.12e0.20
1200e1700 1200e1700 1200e2000 800e1500 1200e1500 700e1700 800e2000 600e1000 1200e1500 1500e3000
100e150 20e35 30e40 <10 25e35 10e15 <10 <5 30e50 >200
Jsky ¼ Sky View Factor; H:W ¼ building height to width ratio; SF ¼ building surface fraction; ZH ¼ roughness height; RC ¼ terrain roughness class; m ¼ thermal admittance; QF ¼ anthropogenic heat flux.
1. Brancumhall (55 460 2000 , 4 80 4300 , 173 m amsl): 2-min sampling time, data collected with Oregon Scientific Wireless Weather Station model ‘WMR968’ (data available for 2009e2011) 2. Glasgow Airport (55 520 1100, 4 260 4800 , 8 m amsl): halfhourly data (data for 1997e2011 except 2000), no information is given with regard to equipment used 3. Renfrewshire (55 470 300 , 4 250 1800 , 138 m amsl): 5-min sampling time, data collected with Davis Vantage Pro Weather Station (data for 2003e2011) 4. Wishaw (55 470 900 , 3 550 2200 , 123 m amsl): 2-min sampling time, data collected with Davis Weather Monitor II Weather Station (data for 2003e2011) In order to characterise the land use/land cover patterns around the data stations, we used the ‘Local Climate Zone’ (LCZ) system developed by Stewart and Oke [4] (Fig. 3). LCZs are defined as ‘regions of uniform surface-air temperature distribution at horizontal scales of 102e104 m’ [4]. Their definition is based on characteristic geometry and land cover that generates a unique nearsurface climate under calm, clear skies. These include vegetative fraction, building/tree height and spacing, soil moisture, and anthropogenic heat flux. Typical zone properties are given in Table 2. LCZ has 16 climate zones and the classification system has been validated in Sweden, Japan and Canada [4].
statistically significant, rise may indicate a general warming in the region. Figs. 5e7 show the long term (over 50 years) trends in daily maximum and minimum temperatures. Although the warming rate of daily maxima (Fig. 5) is slightly higher (i.e. higher slope) than that of daily minima (Fig. 6), both trends are very weak. However the seasonally disaggregated trends show significant differences between summer and winter trends (Fig. 7). The temperature rise in winter is more accentuated than during summer, with monthly mean minima showing more subtle changes e suggesting a more frequent occurrence of milder winters and warmer summers. Table 3 shows linear equations (best fit trend lines), corresponding to the different periods evaluated. Results from the trend lines after 50 years can yield as much as a 2.4 K increase in average monthly maximum temperature for winter months e from a theoretical 5.5 C to 7.9 C or, according to actual records, from 5.3 C in 1959 to 7.1 C in 2009. Slope coefficients in the trend lines are more expressive in winter (for the minimum temperatures, 0.0337 against 0.0015 for all months and 0.0233 for summer periods; for the maximum temperatures, 0.0487 against 0.0023 and 0.0296, respectively).Table 3 also presents the difference between the average minimum and maximum
20
16
Temperature degC
(ELR eor the vertical temperature profile), as presented by Oke [23], is not constant and may exhibit temporal variations during the day and due to changing atmospheric stability. The correlation between land cover/land use and air temperature was attempted using data available at the Weather Underground network (http://www.wunderground.com), for four different locations around the city (again, there are height differences between stations, which might have had an effect on the observed temperature differences):
12
8
4. Results 4
4.1. Trend analysis of historical data Fig. 4 shows the monthly average minimum and maximum temperatures corresponding to the two reference periods (1961e1990 and 1971e2000) registered at Paisley. A small increase in maximum temperature especially in the summer months can be seen in Fig. 4. The rise in average annual temperature between both periods reached 0.7 K (p < 0.001) for the maxima and 0.2 K (p < 0.05) for the minima. This small, though
0 Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Month
Paisley 1961-1990 avg max temp
Paisley 1961-1990 avg min temp
Paisley 1971–2000 avg max temp
Paisley 1971–2000 avg min temp
Fig. 4. West of Scotland background temperature from the recent ‘climate normals’ (1961e1990 and 1971e2000) e data from Paisley.
142
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
Fig. 5. Paisley historical data e monthly maximum temperatures (1959e2009).
winter temperature at the start (1959) and end (2009) of the data series as well as those for the starting and ending months; however such differences are to be regarded cautiously, as they don’t account for fluctuations within the full range of measurements. While the trend lines in daily maxima and minima are somewhat weak, the frequency distribution of air temperatures show interesting patterns of change: lower temperatures shifted downwards in the cumulative frequency chart (less frequent) and an increase in the frequency of higher temperatures. Fig. 8 shows the cumulative frequencies for the first and the last decades of the data series e 1960e1969 and 2000e2009.
4.2. Pairwise comparison of ‘urban’ and rural’ data Fig. 9 shows histograms of temperature differences (grouped into 1 C bins) between the ‘urban’ and ‘rural’ sites during a 12-year period from 1974 to 1985 (see Fig. 2 for location of the stations). Data consist of daily minimum and maximum temperatures recorded from October 1974 through December 1985. These differences were found to be statistically significant (p < 0.001). Histograms suggest that minimum temperature differences are more consistently positive and narrowly distributed, i.e. the urban station presents the nocturnal heat island effect more frequently
Fig. 6. Paisley historical data e monthly minimum temperatures (1959e2009).
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
143
Fig. 7. a) Paisley historical data: monthly mean maximum and minimum temperatures e winter periods (1959e2009). b) Paisley historical data: monthly mean maximum and minimum temperatures e summer periods (1959e2009).
Table 3 Trend lines and variations in local air temperature in Paisley (1959e2009). Period
All months
Winter months (JaneFeb) Summer months (JulyeAugust)
Trend line equation
Relative temperature rise in the 50-year period (from trend lines)
Actual temperature increase after 50 years Average minimum
Average minimum
Average maximum
Average minimum
Average maximum
Average maximum
y ¼ 0.0015x þ 5.6196 R2 ¼ 0.0043 y ¼ 0.0337x þ 0.7088 R2 ¼ 0.1188 y ¼ 0.0233x þ 10.984 R2 ¼ 0.161
y ¼ 0.0023x þ 11.968 R2 ¼ 0.0068 y ¼ 0.0487x þ 5.4686 R2 ¼ 0.2768 y ¼ 0.0296x þ 18.395 R2 ¼ 0.1165
0.9
1.4
0.6
1.1
1.7
2.4
1.8
1.9
1.2
1.5
0.7
0.2
144
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
to the mean, for the whole period, encompassing 2649 lines of data, and for selected winter and summer periods of that sample. Average differences are more pronounced for the daily minimum temperatures, which also has the smallest standard deviations. The CVs further corroborate the observation made above regarding both histograms, that there is a higher consistency in the ‘urban’e‘rural’ temperature differences in the daily minima. However, it is for the winter periods that the nocturnal heat island effect is more clearly pronounced. Fig. 10 shows the frequency of nights with UrbaneRural temperature difference were above a set value of 2 K. Data for two summer months and two winter months are shown as columns, whereas the dotted line shows results for all months.
a 100% 90% 80%
Frequency
70% 60% 50% 40% 30% 20% 1960-1969 2000-2009
10%
4.3. Effect of land use/land cover and local climate
0% -2
b
0
2
4
6
8
10
12
14
We selected 261 days of measurements in 2010 for which data from at least 4 of the weather stations surrounding the city of Glasgow were available. Given their relatively consistent distance from the city core, we hypothesised that the differences between the stations were the result of local land cover/land use conditions. We used the local climate zones (LCZ) classification system [4] to categorise the land cover surrounding of each of the selected weather stations. Fig. 11 shows the land cover characteristics of an area approximately 500 m 500 m surrounding each of the weather stations. We used these images to estimate “regions of uniform surface-air temperature distribution at horizontal scales of 102e104 m” [4]. We used the Google Earth images shown in Fig. 11 and street camera views to estimate the land cover class. In case of mixed classes, we used fetch effect likely due to the prevailing winds (south-westerly). LCZ classes for the four stations are:
22
1. Brancumhall: located closer to suburban East Kilbride than to Glasgow City in the county of South Lanarkshire. There is considerable urban development south-west of the weather station, mostly with lowrise buildings. Accordingly, we assigned an LCZ class of OPEN-SET LOWRISE 2. Glasgow Airport: Located at the Glasgow International Airport in Paisley, on the north side of the main airport terminal. LCZ classification: EXTENSIVE LOWRISE/LOW PLANT COVER 3. Renfrewshire: similar to Brancumhall, this location is surrounded by lowrise buildings in Neilston, East Renfrewshire. Although it can be also classified as OPEN-SET LOWRISE, it has substantially less obstructions to the south-west 4. Wishaw: Located close to Wishaw, North Lanarkshire, this site has a suburban aspect and is more vegetated than all the other
Monthly Minimum Temperature deg C (avg) 100% 90% 80%
Frequency
70% 60% 50% 40% 30% 20% 1960-1969 2000-2009
10% 0% 0
2
4
6
8
10
12
14
16
18
20
Monthly Maximum Temperature deg C (avg) Fig. 8. Comparison of cumulative frequencies e 1960e1969 and 2000e2009. (a) monthly minimum temperatures. (b) monthly maximum temperatures.
and consistently. However the maximum temperature difference has a wider frequency distribution and smaller amplitude. It was even negative in some instances (indicating a daytime cool island effect in the city centre). The average daily temperature fluctuation for both sites has an offset of almost one degree in the rural site (6.3 K at the rural site against 5.5 K at the urban site). Table 4 provides average differences, standard deviations and coefficients of variation (CV in %), defined as the ratio of the standard deviation
Fig. 9. Histograms for ‘urban’e‘rural’ temperature differences (1974e1985): left e daily minima; right e daily maxima.
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
145
Table 4 Temperature differences in degrees Celsius between Glasgow Weather Centre and Springburn for the period of 1974e1985. Delta T summer
Delta T winter
UrbaneRural (maxima) Average Standard Deviation Coefficient of Variation %
UrbaneRural (minima)
0.5 2.2
1.4 0.8
478
Delta T all periods
UrbaneRural (maxima)
UrbaneRural (minima)
1.2 1.8
56
UrbaneRural (maxima)
1.8 1.4
152
0.8 2.1
81
stations, having also a public park (Dalziel Park) and a golf course in the vicinity. LCS classification is mixed in this case: OPEN-SET LOWRISE/CLOSE-SET TREES/LOW PLANT COVER.
UrbaneRural (minima) 1.6 1.2
261
74
local differences. Fig. 12 reinforces the usefulness of the LCZ approach, which takes into account intra-urban or, in this case, peri-urban temperature differences. 4.3.2. Energy implications of local climate differences A simpler way to quantify the heating energy implications of local climate differences is to compare the Heating Degree Days (HDD) for the four weather stations (we used an HDD base line of 65 F or 18.3 C).Fig. 14 shows HDD trends sorted by LCZ class while Fig. 15 shows HDDs according to distance from the city centre. Since the data set available was incomplete, we normalised the HDDs per season by dividing the calculated HDDs by the number of days in a season for which data were available. Sorting locations relative to their LCZ classifications gives a clearer pattern than simply sorting them according to the proximity to the city core (Fig. 15). Trend lines in both graphs suggest a slight decrease in heating demand both when there is more urbanization and when the location is closer to the city core.
4.3.1. Temperature profiles Table 5 shows the annual and seasonal differences in Tmax, Tmin and the diurnal temperature range (DTR) for the 4 selected stations. The last row indicates the maximum differences obtained for each variable. As with the urbanerural comparison shown in Section 4.2 temperature differences between stations are more pronounced in the winter period and for Tmin. A further comparison of the departure from the group mean temperature shows the relative behaviour of each location compared to the mean (Table 6). Fig. 12 shows the correspondence between such figures for the daily minimum temperatures (indicator of the urban heat island effect) and the LCZ classification for each location, arranged left to right from the least to the most densely built location. A clear pattern arises suggesting the existence of urban warming in the more built-up areas. This pattern is observable annually as well as in winter and summer periods, although more pronounced in the winter. A comparison to a more simplistic approach, namely the urban heat island is related to the distance from the city centre (Fig. 13) shows that the distance theory is inadequate to explain the
4.3.3. Heat island profile on a clear-sky day Fig. 16 shows the temperature profiles from the four stations on a clear day. The criteria used for choosing the particular date was as follows: the most complete database; lowest average daily cloud cover, lowest average daily wind speed (below 1 m/s) and high daily temperature fluctuation (above 10 K). The date chosen for
% of days with Turb-rur>2K 100 Winter months Summer months All months
Frequency %
80
60
40
Oct-81
Dec-81
Aug-81
Apr-81
Jun-81
Feb-81
Oct-80
Dec-80
Aug-80
Apr-80
Jun-80
Feb-80
Oct-79
Dec-79
Aug-79
Apr-79
Jun-79
Feb-79
Oct-78
Dec-78
Aug-78
Apr-78
Jun-78
Feb-78
Oct-77
Dec-77
Aug-77
Apr-77
Jun-77
Feb-77
Oct-76
Dec-76
Aug-76
Apr-76
Jun-76
Feb-76
Oct-75
Dec-75
Aug-75
Apr-75
Jun-75
Feb-75
Oct-74
0
Dec-74
20
Fig. 10. Frequency patterns for a nocturnal (daily minima) temperature difference between the urban and rural locations Note: Only those days with an urbanerural difference greater 2 K are shown.
146
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
Fig. 11. Land cover characteristics of 500 500 m area surrounding each of the weather from top left, clockwise: a) Brancumhall, b) Glasgow Airport, c) Renfrewshire, d) Wishaw Source: Google Earth.
analysis was 11 April 2010, a day with clear skies and a solar radiation profile closely matching the theoretical maximum. It can be seen from Fig. 16 that a significant difference between locations is more noticeable at night, confirming the nocturnal characteristic of the heat island phenomenon. The most urbanized location (solid line) has the smallest diurnal temperature range and the highest minimum temperature for the day, in striking contrast to the site with the least built-up area (Glasgow Airport, grey dotted line). 5. Discussion Our results from long term temperature records (see Section 4.1) are similar to the climate change reported by the UK Climate Impact Programme (UKCIP’09) [24]. While the UKCIP’09 reported a prominent warming trend in South East England, it also showed some warming in the West of Scotland. Such changes are usually higher in winter than in summer months (increases in average daily maxima: 1.37 K in summer and 1.57 K in winter; increases in average minima: 1.34 K in winter and 1.23 K in summer). Our
Table 5 Annual, winter and summer average maximum (Tmax) and minimum temperatures (Tmin) and corresponding diurnal temperature ranges (DTR). Location
Annual Tmax
Brancumhall Glasgow Airport Renfrewshire Wishaw Maximum difference
10.2 10.8 10.5 10.4 0.6
Tmin 3.8 2.5 3.3 1.9 1.9
Winter DTR 6.4 8.3 7.1 8.5 2.1
Tmax 4.1 4.7 4.1 4.6 0.6
Tmin 1.0 2.9 1.6 3.8 2.8
Summer DTR 5.1 7.6 5.7 8.4 3.3
Tmax 17.9 18.5 18.4 18.1 0.6
Tmin 11.1 10.0 10.4 9.7 1.4
DTR 6.8 8.5 8.0 8.5 1.7
results suggest that urban areas superimpose a local warming effect over this regional trend. Superimposing our long term trend analysis (from section 4.1) on the climate change analysis provided by Barnet et al. [25] for the West, East and North of Scotland shows many similarities (Fig. 17), although air temperatures in our case were generally higher than the regional average. Both the minimum and the maximum air temperatures had a consistent shift of about 1.5 K. Within the period around 1985e1995, minimum temperatures in our case diverged from regional trends, while still maintaining higher values than the general regional trend. With regard to heating degree days (HDD), the regional trends reported by Barnet et al. [25] for a HDD with a base temperature of 60 F (15.5 C) showed a 11.3% decrease in 1961e2004. Our analysis of Paisley data show that HDD dropped from 2198 HDD to 1945 HDD during the same period, which corresponds to a decrease of 11.5%. Thus, the observed HDD decrease may be due to the regional warming. However, looking more closely at absolute values for ambient temperatures near Glasgow, Fig. 18 shows a comparison between Table 6 Annual, winter and summer departures from the group mean maximum (Tmax) and minimum temperatures (Tmin) and corresponding diurnal temperature ranges (DTR). Locations
Brancumhall Glasgow EGPF Renfrewshire Wishaw
Annual
Winter
Summer
Tmax
Tmin
DTR
Tmax
Tmin
DTR
Tmax
Tmin
DTR
0.3 0.3 0.0 0.1
0.9 0.4 0.5 1.0
1.2 0.7 0.5 0.9
0.3 0.3 0.2 0.2
1.3 0.6 0.7 1.5
1.6 0.9 1.0 1.7
0.3 0.3 0.2 0.1
0.8 0.3 0.1 0.6
1.1 0.6 0.1 0.5
1.5
Departure from group mean in deg C
Annual
Winter
Summer
1.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0 EXTENSIVE LOWRISE/LOW OPEN-SET PLANT COVER LOWRISE/CLOSE-SET TREES/LOW PLANT COVER
OPEN-SET LOWRISE
OPEN-SET LOWRISE
LCZ classification Fig. 12. Annual, winter and summer departures from the group mean minimum temperature.
Departure from group mean in deg C
1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5
Winter
Annual
Summer
-2.0 11.3
12.1
13.6
22.4
Distance to city centre (Buchanan Street) in km Fig. 13. Annual, winter and summer departures from the group mean minimum temperatures (Tmin) sorted ordered according to proximity to city core. Annual
Winter
Linear (Annual)
Linear (Winter)
climatic normals for Glasgow Airport and West of Scotland (1971e2000, regional data from [22]). Clearly, local temperature is higher than regional averages. In addition, the number of days with night frost drops significantly, from around 70 to 36 days per annum. This fact points to the possibility of using the heat island in cities like Glasgow as an opportunity to enhance local energy savings and perhaps enhance the efficacy of district heating schemes. Additionally, the pairwise comparison between an ‘urban’ and a ‘rural’ weather stations (Section 4.2) indicates a clear heat island effect. It should be noted that our pairwise comparison (reflecting a 12-year average for the period of 1974e1985) has a weak agreement with the relevant figure for Glasgow in Table 1 (for 2003). Seasonal differences show that in general the nocturnal UHI is stronger in winter than in summer, even though it is more consistent in the summer season (CV 56%). Together with the fact that land cover characteristic appears to have an influence on local climate (Section 4.3), the indications are that the urban features in and around Glasgow are augmenting the slight regional warming in the West of Scotland region. This trend appears to have continued despite the loss of population in the city.
20.0 18.0
Annual
20.0
Winter
Linear (Annual)
Linear (Winter)
16.0
18.0 14.0
16.0 14.0 10.0
12.0
HDD
HDD
12.0
8.0 6.0
10.0 8.0
4.0
6.0
2.0
4.0
0.0 OPEN-SET EXTENSIVE OPEN-SET LOWRISE LOWRISE/CLOSE-SET LOWRISE/LOW PLANT TREES/LOW PLANT COVER COVER
OPEN-SET LOWRISE
LCZ Classification
2.0 0.0 11.3
12.1
13.6
22.4
Distance to city centre (Buchanan Street) in km Fig. 14. Normalized Heating Degree Days for the different locations (sorted by the LCZ class).
Fig. 15. Normalized heating degree days sorted by proximity to city core.
148
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
20 18
Temperature degC
16 14 12 10 8 6 4 2
12:00 AM
11:00 PM
10:00 PM
9:00 PM
8:00 PM
7:00 PM
6:00 PM
5:00 PM
4:00 PM
3:00 PM
2:00 PM
1:00 PM
12:00 PM
11:00 AM
10:00 AM
9:00 AM
8:00 AM
7:00 AM
6:00 AM
5:00 AM
4:00 AM
3:00 AM
2:00 AM
1:00 AM
0
Time of day OPEN-SET LOWRISE/CLOSE-SET TREES/LOW PLANT COVER EXTENSIVE LOWRISE/LOW PLANT COVER OPEN-SET LOWRISE OPEN-SET LOWRISE Fig. 16. Temperature profile for 11 April 2010.
Fig. 17. a) Annual maximum and b) Annual minimum temperature averages for Paisley (thick grey line), plotted against the three climate regions in Scotland (from Barnet et al., 2006).
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
b
25
16 14
20
Number of days
Temperature degC
a
149
12
15
10
10 5 0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
8 6 4 2 0
-5
Month Max Temp [°C] Scotland Max Temp [°C] Paisley
Jan Feb Mar Apr May Jun
Jul Aug Sep Oct Nov Dec
Month Min Temp [°C] Scotland Min Temp [°C] Paisley
Days of Air Frost Scotland
Days of Air Frost Paisley
Fig. 18. a) Ambient temperature averages for Paisley and Scotland from climatic normals 1971e2000 and b) number of days with air frost for Pasiley and Scotland.
6. Planning and energy implications The planning and energy implications of our findings to ‘shrinking cities,’ are intriguing. ‘Shrinking cities’ e a concept initially theorised in the wake of German unification [26] is an increasingly common reality in many parts of the world. Over the last fifty years, 370 cities throughout the world with populations over 100,000 have shrunk by at least 10% [27]. These are more common in the industrial heartlands of the USA (59 cities), Britain (27), Germany (26), Italy (23), Russia (13), South Africa (17) and Japan (12). They are also common in other parts of the world, even as growing cities continue to dominate the discourse. A typical planning approach to this crisis is to reconceptualise decline as shrinkage and to explore creative and innovative ways for cities to successfully shrink. Such approaches have usually taken the form of land for recreation, urban agriculture, green infrastructure, and other non-traditional land uses beneficial to existing residents and attract future development [28]. In their drive towards sustainable and ecologically sound places shrinking cities will need to consider the local climate implications of their current urban trajectories. While population may decline the underlying urban morphology largely remains in place, leading to the continuation of the urban climate anomaly. In the case of a temperate climate city such as Glasgow, the association between urban form and the ‘urban’e‘rural’ temperature differences continue to hold: this aspect of shrinking is beneficial, in that the urban warmth created by a judicious arrangement of land use/land cover (as evidenced by the appropriate LCZ class) could be exploited for energy efficient uses such as district heating and to enhance the feasibility of low carbon options such as district ground source heating or other communal renewable technologies. Our work shows that the UHI itself does not go away, even in shrinking cities, thereby the opportunities to be sustainable and low carbon might still be available. At the same time, the summertime trends suggest that overheating may become a distinct possibility in the future. These realities should inform shrinking cities in their attempt to reinvent themselves in a carbon and energy efficient fashion.
References [1] United Nations, Department of Economic and Social Affairs, Population Division. World urbanization prospects: the 2009 revision; 2010. CD-ROM Edition e Data in digital form (POP/ DB/WUP/Rev.2009). [2] Grimmond CSB, Roth M, Oke TR, Au YC, Best M, Betts R, et al. Climate and more sustainable cities: climate information for improved planning and management of cities (producers/capabilities perspective). Procedia Environ Sci 2010;1:247e74. [3] Mills G, Cleugh H, Emmanuel R, Endlicher W, Erell E, McGranahan G, et al. Climate information for improved planning and management of mega cities (needs perspective). Procedia Environ Sci 2010;1:228e46. [4] Stewart ID, Oke TR. A new classification system for urban climate sites. Bull Am Met Soc 2009;90:922e3.
[5] Fujibe F. Detection of urban warming in recent temperature trends in Japan. Int J Climatol 2009;29:1811e22. [6] Fujibe F. Urban warming in Japanese cities and its relation to climate change monitoring. Int J Climatol 2011;31:162e73. [7] Grimmond CSB. 2011. London’s urban climate: historical and contemporary perspectives, paper presented at the conference City Weathers: Meteorology and Urban Design 1950-2010, June 23, 2011 http://www.sed.manchester.ac. uk/architecture/research/csud/workshop/programme/ [accessed 01.07.11]. [8] Kolokotroni M, Davies M, Croxford B, Bhuiyan S, Mavrogianni A. A validated methodology for the prediction of heating and cooling energy demand for buildings within the Urban Heat Island: case-study of London. Solar Energy 2010;84:2246e55. [9] Kershaw T, Sanderson M, Coley D, Eames M. Estimation of the urban heat island for UK climate change projections. Building Serv Eng Res Technol 2010;31:251e63. [10] Thorsson S, Lindberg F, Björklund J, Holmer B, Rayner D. Potential changes in outdoor thermal comfort conditions in Gothenburg, Sweden due to climate change: the influence of urban geometry. Int J Climatol 2011;31:324e35. [11] United Nations Population Fund (UNFPA). Growing up urban. Supplement to the state of world population 2007. New York: UNFPA; 2007. [12] Vörösmarty CJ, Green P, Salisbury J, Lammers RB. Global water resources: vulnerability from climate change and population growth. Science 2000;289: 284e8. [13] Hansen J, Ruedy R, Sato M, Imhoff M, Lawrence W, Easterling D, et al. A closer look at United States and global surface temperature change. J Geophys Res 2001;106(D20):23947e63. [14] Jin M, Dickinson RE, Zhang D-L. The Footprint of urban areas on global climate as characterized by MODIS. J Clim 2005;18:1551e65. [15] Angel S, Sheppard SC, Civco DL. The dynamics of global urban expansion. Washington, D.C.: Transport and Urban Development Department, the World Bank; 2007. [16] Lamptey B. An analytical framework for estimating the urban effect on climate. Int J Climatol 2010;30:72e88. [17] Parker DE. Urban heat island effects on estimates of observed climate change. WIRES Clim Change 2010;1:123e33. [18] IPCC. Climate change 2007: the physical science basis. In: Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt K, Tignor M, Miller HL, editors. Contribution of working group I to the fourth assessment report of the IPCC. New York: Cambridge University Press; 2007. p. 1056. [19] Maver I. Glasgow’s public parks and the community, 1850e1914: a case study in Scottish civic interventionism. Urban Hist 1998;25:323e47. [20] Mackenzie JM. The second city of the Empire: Glasgow e imperial municipality. In: Driver F, Gilbert D, editors. Imperial cities: landscape, display and identity; 1999. p. 215e37. Manchester. [21] Cameron EA. Glasgow’s going round and round: some recent Scottish urban history. Urban Hist 2003;30:276e87. [22] UK Meteorological Office (UK Met Office). Monthly and annual average climate data at 5 km resolution. Can be accessed at: http://www.metoffice. gov.uk/climate/uk/stationdata/; 2011 [last accessed 29.06.11]. [23] Oke TR. Boundary layer climates. London: Methuen; 1987. [24] Murphy JM, Sexton DMH, Jenkins GJ, Boorman PM, Booth BBB, Brown CC, et al. UK climate projections science report: climate change projections. Exeter: Met Office Hadley Centre, http://www.ukcip.org.uk/ukcp09/; 2009 [accessed 21.07.11]. [25] Barnett C, Hossell J, Perry M, Procter C, Hughes G. Patterns of climate change across Scotland: technical report. 102 p. In: SNIFFER Project CC03. Scotland & Northern Ireland Forum for Environmental Research, http://www.sniffer. org.uk/Webcontrol/Secure/ClientSpecific/ResourceManagement/UploadedFiles/ CC03_Final_report.pdf; 2006 [accessed 07.07.11]. [26] Rieniets T. Shrinking cities: causes and effects of urban population losses in the twentieth century. Nat Cult 2009;4:231e54. [27] Oswalt BP, Rieniets T. Shrinking cities: global study. Can be accessed at, http:// www.shrinkingcities.com/globaler_kontext.0.html?&L¼1; 2007 [last accessed 30.06.11]. [28] Hollander JB, Pallagst K, Schwarz T, Popper FJ. Planning shrinking cities. available at: http://policy.rutgers.edu/faculty/popper/ShrinkingCities.pdf; 2009 [last accessed 3006.11].
Contents lists available at SciVerse ScienceDirect
Building and Environment journal homepage: www.elsevier.com/locate/buildenv
Urban heat island and its impact on climate change resilience in a shrinking city: The case of Glasgow, UK Rohinton Emmanuel*, Eduardo Krüger School of Engineering and Built Environment, Glasgow Caledonian University, 70 Cowcaddens Road, Glasgow G4 0BA, UK
a r t i c l e i n f o
a b s t r a c t
Article history: Received 20 July 2011 Received in revised form 13 December 2011 Accepted 25 January 2012
Given its long urban history of growth and decline, Glasgow, UK, provides a historically significant opportunity to study the local climatic changes brought about by urban variables. This study investigates the changes in air temperature within the central area of Glasgow using three data sources: the UK Meteorological Office historical data for Glasgow (climate normals and running data for a 50-year period), the Weather Underground network; MIDAS Surface Weather Stations network of the British Atmospheric Data Centre (BADC). Three approaches were used to evaluate Glasgow’s local climate change: assessment of mean air temperature increases based on two concurrent climate normals, traditional UHI approach (i.e. differences between a ‘rural’ and an ‘urban’ site) and observed temperatures in locations with different land cover characteristics (using the Local Climate Zone LCZ concept. Planning and building scale implications for other shrinking cities are explored. Ó 2012 Elsevier Ltd. All rights reserved.
Keywords: Urban climate Urban heat island Local climate zone Climate change Shrinking cities
1. Introduction Cities are home to over half of the global population [1]. The current growth trajectories indicate that by 2050, nearly 70% of humanity (6.3 billion out of an estimated global population of 9.1 billion) will live in urban areas [1]. The local, regional and global climate implications of such rapid urban growth are complex: On the one hand, cities are major consumers of energy and materials as well as generate vast amounts of waste. They are also major centres of innovation and finance e which could enable them to be at the forefront of climate mitigation actions. At the same time, the nature of climate change with its long lag-times mean that cities need to prepare themselves now, to act as the first line of defence against catastrophic effects of local and global changes expected in the near future. Given the twin realities of potential innovation in mitigatory action as well as adaptive capacity and need in the face of burgeoning urban population, urban areas are beginning to receive a long overdue attention from climate change scientists and policy makers [2,3]. In these efforts, the role of urban climate change in both contributing to and augmenting global change is a key unknown that needs careful attention. Could there be opportunities to use the mitigatory potential of urban fabric to reduce the effects of urban warming with a view to providing some relief to the wider
* Corresponding author. Tel.: þ44 (0) 141 331 3217. E-mail address: [email protected] (R. Emmanuel). 0360-1323/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.buildenv.2012.01.020
and more extreme weather events spawned by global climate change? A further reality confounding the issue is urban decay. Major centres of urban population have already begun to lose population and even those that are yet growing continue to sprawl, leading to lowering urban densities in many parts of the world. What effects will shrinking cities as well as de-densification have on local climate change? What lessons can we learn from the local climate change trajectories of mature cities to plan better the still growing cities of the world? In this paper, we analyse the trends and local differences in air temperature in and around the mature urban area of Glasgow, UK (55 510 N, 04120 W). Using historical weather data from three different sources we determine the effect of local climate zones [4] on microclimate and postulate what lessons could be learnt by other mature and/or shrinking cities and how the city of Glasgow could benefit from these changes in its quest for a sustainable and low carbon future. 2. Background 2.1. Urban heat island in mature cities Urban warming and its links to regional and global warming have been well documented in several mature cities in Japan, Europe (especially in the UK) North America and Sweden. Fujibe’s work in Japan [5,6] epitomises these efforts. Reviewing data from
138
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
561 meteorological stations established in Japanese cities in the early 20th Century, Fujibe [6] found that there is a warming trend of 0.3e0.4 C/decade even for locations with low population density (<100 people per square kilometre), indicating that the recent temperature increase is largely contributed by background climate change. However, an anomalous warming trend is detected for stations with larger population density (in the order of 100e300/ km2) where the anomalous trend is 0.03e0.05 C/decade [6]. Furthermore the recorded rate of temperature increase tends to be larger at night than during the daytime, although in the case of a megacity (Tokyo) widespread urban warming in the hinterland during afternoons of the warm season is seen as a result of extensive urbanization that enhances daytime surface heating. Among the European mature cities, the heat island effect in London is one of the most well studied (see Ref. [7] for a comprehensive review). On the positive side, heating energy consumption in central London is 65e85% of the heating required for the same building based outside the Urban Heat Island [8]; conversely, cooling energy consumption is 32e42% higher in the city. Given the fuel mix for heating and cooling (gas for heating and coal for electricity, [8]), the carbon implications of UHI to London are negative. An estimate of the heat island effect in major population centres in the UK (based on the UK Met Office historical weather data, see UK Met Office, 2011) was developed by Kershaw et al. [9]. Table 1 shows the seasonal and annual average UHIs for several UK cities. The heat island effect on outdoor thermal comfort at a mature city (Göthenburg, Sweden) was studied by Thorsson et al. [10]. Statistically downscaling climate change projections to the city street-levels, Thorsson et al. [10] showed that urban geometry could cause large intra-urban differences in Mean Radiant Temperature (Tmrt), on hourly, daytime and yearly time scales. In general, open areas are warmer than adjacent narrow street canyons in summer, but cooler in winter. The combination of regional warming coupled with augmentation by urban geometries will triple strong/extreme heat stress (to approx. 20e100 h a year, depending on geometry). Conversely, the number of hours with
Table 1 Seasonal and annual average UHI values ( C) for UK cities. City
Winter (DJF)
Spring (MAM)
Summer (JJA)
Autumn (SON)
Annual
New castle upon-Tyne Portsmouth Central London Liverpool Glasgow Edinburgh London Suburbs Plymouth Sheffield Manchester Bristol Middlesbrough York Cardiff Leeds-Bradford Nottingham Bournemouth Birmingham Coventry Belfast Leicester
1.8 1.3 1.3 1.1 1.0 1.1 0.9 0.9 0.9 0.8 0.6 0.8 0.8 0.5 0.6 0.5 0.6 0.4 0.4 0.4 0.0
1.9 1.73 1.7 1.4 1.5 1.4 1.2 1.2 0.9 1.1 1.0 1.0 0.8 0.9 0.7 0.7 0.8 0.5 0.5 0.5 0.2
2.0 2.0 1.9 1.7 1.5 1.4 1.4 1.2 1.1 1.2 1.2 0.9 1.0 1.1 0.8 0.9 0.8 0.7 0.7 0.3 0.1
1.8 2.0 1.6 1.5 1.2 1.3 1.1 1.1 0.9 0.8 0.9 0.8 0.9 0.6 0.8 0.6 0.5 0.6 0.4 0.3 0.1
1.9 1.8 1.6 1.4 1.3 1.3 1.1 1.1 1.0 0.9 0.9 0.9 0.9 0.8 0.7 0.7 0.7 0.6 0.5 0.4 0.1
Note: UHI values are given as three-month average temperature difference between a city centre weather station and a nearby rural station. Winter ¼ December, January and February (DJF); Spring ¼ March, April and May (MAM); Summer ¼ June, July and August (JJA); Autumn ¼ September, October and November (SON). Source: Based on Kershaw et al., 2010.
strong/extreme cold stress will decrease by 400e450 h. Furthermore, the number of hours with no thermal stress will increase by 40e200 h a year. Considering both winter warming and summer cooling potential of judicious arrangement of urban geometry, Thorsson et al. [10] conclude that a densely built urban structure will mitigate extreme swings in Tmrt and in the generally adopted thermal comfort index ‘physiologically equivalent temperature’ (PET), improving outdoor comfort conditions both in summer and in winter. 2.2. Climate change and urban heat island Given the small fraction of land occupied by cities (all urban sites including green as well as built-up areas cover only 2.8 per cent of the Earth’s land area [11]) it is unlikely that cities have a direct bearing on global climate change. However, cities indirectly drive global climate change on account of their insatiable appetite for energy and material (and associated waste and pollution). Furthermore, an increasing urban population will lead to the expansion of land covered by cities, at which point the direct influence of cities on regional and perhaps global climate may not be insignificant. There is increasing evidence to the scale of urban influence on the global climate to be in the order of El-Nino Southern Oscillation (ENSO) [12]. The significance of the signal from urban climate change on the global climate led Hansen et al. [13] to term it as “urban warming.” A key difficulty in untangling the urban warming from global climate changes is the computational and parametric difficulties associated with representing urban areas in climate models. Yet it is increasingly recognised that climate models provide a good approach for assessing the global consequences of the urban climate modifications [14]. Given the lack of detailed land cover information and computing power, most climate models generally do not include a representation of urban areas, and therefore their climate projections are likely to underestimate the heat island phenomenon [9]. The general consensus appears to be that the UHIs are likely to augment the temperature anomaly arising from global warming [2,9]. A recent study commissioned by the World Bank shows that modern patterns of city growth are increasingly land intensive [15]. Average urban densities have been declining for the past two centuries. As transportation continues to improve, the tendency is for cities to use up more and more land per person [15]. In developing countries, cities of 100,000 or more are expected to triple their built-up land area to 600,000 km2 in the first three decades of this century. Cities in developed countries expand at an even faster rate per resident, despite their smaller population size and lower rate of population growth. They will increase their built-up land area by 2.5 times between 2000 and 2030. At that point, they will occupy some 500,000 km2 [15]. These may have direct consequences to global climate. Even if the urban effects on global climate change remain weak, the influence of UHIs on regional climates is clearer. Lamptey [16] attempted to untangle the sensible and latent heat partitioning associated with different land cover classes in Chester County and surroundings near Philadelphia, Pennsylvania, USA. The urban effect became more important as the fraction of urban land cover to the total increased. When the urban land cover increased from 11% to 19% in 9 years it led to the largest proportionate sensible (21.4 Wm2) and latent (14.2 Wm2) heat fluxes during winter. During summer, urban and vegetation land cover produced the largest proportionate sensible heat (59.2 Wm2) while urban land cover produced the second largest proportionate latent heat flux (39.5 Wm2). The regional climatic implications of these energy partitioning cannot be ignored.
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
139
2.3. Urban evolution of Glasgow
Fig. 1. Glasgow’s population trend during the last 200 years. Source: Population data for 1801e1931 from: www.histpop.org.uk. Data for 1951e2001 from the decennial census records at the National Records of Scotland (there was no census taken in 1941).
Given this increasing body of evidence, the role of urban climate change as a significant part of the human experience of climate change is increasingly being recognised. While pointing out the ‘minor impact on estimates of global trends of land surface-air temperature (LSAT), Parker [17] recognises the combined effect of global and urban warming on human health and welfare. Fujibe [5] showed the differential nature of warming experienced by Japanese cities of different sizes and densities: cities with over 3000 persons/ km2 or urban land cover over 50% of the total land, showed a warming trend of 0.1 C/decade, compared to the centennial global trend estimated by the IPCC [18] of 0.74 C/century for 1906e2005.
From its medieval ecclesiastical origins Glasgow expanded into a major port in the 18th century and with the advent of industrial revolution, added a massive industrial base to its already well developed built fabric. This industrial base was responsible for the intense concentrations of people, wealth, poverty, and eventual decline of the city. The industrial heritage of Glasgow gave it both the notoriety for pollution as well as fame for its attempt to clean it, especially via the pioneering creation of urban parkland [19]. These parks and the events associated with their creation (International Exhibitions of 1888 and 1901, the Scottish Exhibition of National History, Art and Industry in Kelvingrove Park in 1911, the Empire Exhibition in Bellahoustoun Park, south of the river Clyde, in 1938) gave the city its claim to fame as the ‘second city of the empire’ [20]. The urban and industrial growth of Glasgow has been rapid in the 19th Century, drawing people from all over the UK and beyond [21]. However, the success of its industrial base could not withstand the pressures of globalisation, and by the early 20th Century, the city had begun to lose population. Fig. 1 shows the population trend in the city during the last 200 years. Nevertheless, the provision of urban infrastructure and services, much of which remains to this day, means that the land cover changes induced by urbanisation are with us to this day, despite the decline in population. This has implications for the city’s urban climate and we will return to this aspect of our study in the ‘implications’ section of the present paper. 3. Methods In the present study we performed three types of comparisons to elicit the variations in local climate induced by urban growth and decay: (1) historical trend analysis, (2) pairwise comparison of an
Fig. 2. Location of weather stations used by the present study. Key: Paisley e yellow; MIDAS Stations e red; Weather Underground stations e black. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
140
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
Fig. 3. Illustration of land cover characteristics at different Local Climate Zones (LCZ). Source: Stewart and Oke, 2009.
‘urban’ and ‘rural’ weather stations and (3) correlation between local climate and land use/land cover variations in and around Glasgow. According to Koeppen-Geiger’s climate classification, the region is characterised by a Cfb temperate climate type, especially made mild due to stronger maritime influences. With significant precipitation in all seasons (average annual precipitation in Glasgow is 1100 mm), the average maximum temperatures in the warmest months (July and August) remain below 20 C, but with at least four months averaging above 10 C [22]. For the purpose of historical trend analysis we used the data from the Glasgow International Airport weather station, located in Paisley (Lat: 55 520 N, 04 260 W, alt: 19 m asl), available from the UK Meteorological Office (http://www.metoffice.gov.uk). In addition to the running data for the period of 1959e2009 we also compared the two recent ‘climate normals’ used as reference periods during the last 50 years (1961e1990 and 1971e2000). These two periods correspond to the last two climate normals used by the UK Met Office in reporting average climates. Our pairwise comparison utilised data from the Met Office Integrated Data Archive (MIDAS) Surface Weather Stations network, provided by the British Atmospheric Data Centre (BADC) registered at a ‘rural’ and at an ‘urban’ location (http://badc.nerc.ac. uk, last accessed 30 Jun 2011): Rural e Springburn (Lat: 55 530 25.4400 N, Lon: 4 130 3000 W, alt: 107 m asl); ‘urban’ e Glasgow Weather Centre, situated near the Glasgow Central Station (Lat: 55
510 36.3600 N, Lon: 4 130 33.7800 W, alt: 40 m asl) (Fig. 2). Coincidental data for the period of 1974e1985 were used for such comparison. The Springburn station was situated near the Springburn Park, which continues to remain a green area on the northern fringes of the city. The Glasgow Weather Centre was located at the city core, two street blocks westwards from the Central Railway Station, in an area characterised by midrise historical buildings (7e8 storeys). Springburn Station started operation in January 1894 through March 1997 while the Glasgow Weather Centre Station was in operation from January 1974 until December 1985. According to the Met Office Surface Data Users Guide (http://badc.nerc.ac.uk/data/ surface/ukmo_guide.html#2, accessed 1 December 2011), equipment used in both cases regularly underwent a series of calibrations which linked the instrument to a national or international standard instrument. This practice, which for temperature and pressure has been unbroken since 1851, ensures a uniformity of measurement over time. On both stations, maximum and minimum temperature data were provided for a 24 h cycle starting at 9 a.m., using either a liquid-in-glass thermometer or, from the 1980s electrical resistance thermometers (ERT). It should be noted that there is a height difference between stations (67 m), which could partly explain the lower temperatures measured at Springburn. Although this is a limitation of the study, correcting local temperature to account for elevation will introduce errors and the process is not straightforward. The “environmental lapse rate”
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
141
Table 2 Local Climate Zone (LCZ) classification system Source: Based on Stweart and Oke [4]. Local climate zone (LZC)
Compact highrise Open-set highrise Compact midrise Open-set midrise Compact lowrise Open-set lowrise Dispersed lowrise Lightweight lowrise Extensive lowrise Industrial processing
Zone properties
Jsky
H:W
SF
ZH
RC
a
m (J m2 s½ K1)
QF (W m2)
0.25e0.45 0.40e0.70 0.30e0.60 0.80e0.90 0.30e0.50 0.55e0.75 >0.90 0.30e0.50 >0.90 0.70e0.90
>2 0.75e1.25 0.75e1.25 0.20e0.30 1.00e1.50 0.50e0.75 0.10e0.20 1.00e1.50 <0.25 0.2e0.5
>90% 50e75% >90% 30e50% >80% 45e65% 20e30% 70e90% >80% 45e65
>35 m >30 m 15e25 m 10-25 m 3e10 m 3e10 m 3e7 m 2e4 m 3e10 m 5e10 m
8 7e8 6e7 5e6 6 5e6 5e6 4e5 5 5e6
0.12e0.18 0.12e0.20 0.15e0.20 0.15e0.20 0.12e0.20 0.10e0.20 0.10e0.20 0.10e0.20 0.15e0.25 0.12e0.20
1200e1700 1200e1700 1200e2000 800e1500 1200e1500 700e1700 800e2000 600e1000 1200e1500 1500e3000
100e150 20e35 30e40 <10 25e35 10e15 <10 <5 30e50 >200
Jsky ¼ Sky View Factor; H:W ¼ building height to width ratio; SF ¼ building surface fraction; ZH ¼ roughness height; RC ¼ terrain roughness class; m ¼ thermal admittance; QF ¼ anthropogenic heat flux.
1. Brancumhall (55 460 2000 , 4 80 4300 , 173 m amsl): 2-min sampling time, data collected with Oregon Scientific Wireless Weather Station model ‘WMR968’ (data available for 2009e2011) 2. Glasgow Airport (55 520 1100, 4 260 4800 , 8 m amsl): halfhourly data (data for 1997e2011 except 2000), no information is given with regard to equipment used 3. Renfrewshire (55 470 300 , 4 250 1800 , 138 m amsl): 5-min sampling time, data collected with Davis Vantage Pro Weather Station (data for 2003e2011) 4. Wishaw (55 470 900 , 3 550 2200 , 123 m amsl): 2-min sampling time, data collected with Davis Weather Monitor II Weather Station (data for 2003e2011) In order to characterise the land use/land cover patterns around the data stations, we used the ‘Local Climate Zone’ (LCZ) system developed by Stewart and Oke [4] (Fig. 3). LCZs are defined as ‘regions of uniform surface-air temperature distribution at horizontal scales of 102e104 m’ [4]. Their definition is based on characteristic geometry and land cover that generates a unique nearsurface climate under calm, clear skies. These include vegetative fraction, building/tree height and spacing, soil moisture, and anthropogenic heat flux. Typical zone properties are given in Table 2. LCZ has 16 climate zones and the classification system has been validated in Sweden, Japan and Canada [4].
statistically significant, rise may indicate a general warming in the region. Figs. 5e7 show the long term (over 50 years) trends in daily maximum and minimum temperatures. Although the warming rate of daily maxima (Fig. 5) is slightly higher (i.e. higher slope) than that of daily minima (Fig. 6), both trends are very weak. However the seasonally disaggregated trends show significant differences between summer and winter trends (Fig. 7). The temperature rise in winter is more accentuated than during summer, with monthly mean minima showing more subtle changes e suggesting a more frequent occurrence of milder winters and warmer summers. Table 3 shows linear equations (best fit trend lines), corresponding to the different periods evaluated. Results from the trend lines after 50 years can yield as much as a 2.4 K increase in average monthly maximum temperature for winter months e from a theoretical 5.5 C to 7.9 C or, according to actual records, from 5.3 C in 1959 to 7.1 C in 2009. Slope coefficients in the trend lines are more expressive in winter (for the minimum temperatures, 0.0337 against 0.0015 for all months and 0.0233 for summer periods; for the maximum temperatures, 0.0487 against 0.0023 and 0.0296, respectively).Table 3 also presents the difference between the average minimum and maximum
20
16
Temperature degC
(ELR eor the vertical temperature profile), as presented by Oke [23], is not constant and may exhibit temporal variations during the day and due to changing atmospheric stability. The correlation between land cover/land use and air temperature was attempted using data available at the Weather Underground network (http://www.wunderground.com), for four different locations around the city (again, there are height differences between stations, which might have had an effect on the observed temperature differences):
12
8
4. Results 4
4.1. Trend analysis of historical data Fig. 4 shows the monthly average minimum and maximum temperatures corresponding to the two reference periods (1961e1990 and 1971e2000) registered at Paisley. A small increase in maximum temperature especially in the summer months can be seen in Fig. 4. The rise in average annual temperature between both periods reached 0.7 K (p < 0.001) for the maxima and 0.2 K (p < 0.05) for the minima. This small, though
0 Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Month
Paisley 1961-1990 avg max temp
Paisley 1961-1990 avg min temp
Paisley 1971–2000 avg max temp
Paisley 1971–2000 avg min temp
Fig. 4. West of Scotland background temperature from the recent ‘climate normals’ (1961e1990 and 1971e2000) e data from Paisley.
142
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
Fig. 5. Paisley historical data e monthly maximum temperatures (1959e2009).
winter temperature at the start (1959) and end (2009) of the data series as well as those for the starting and ending months; however such differences are to be regarded cautiously, as they don’t account for fluctuations within the full range of measurements. While the trend lines in daily maxima and minima are somewhat weak, the frequency distribution of air temperatures show interesting patterns of change: lower temperatures shifted downwards in the cumulative frequency chart (less frequent) and an increase in the frequency of higher temperatures. Fig. 8 shows the cumulative frequencies for the first and the last decades of the data series e 1960e1969 and 2000e2009.
4.2. Pairwise comparison of ‘urban’ and rural’ data Fig. 9 shows histograms of temperature differences (grouped into 1 C bins) between the ‘urban’ and ‘rural’ sites during a 12-year period from 1974 to 1985 (see Fig. 2 for location of the stations). Data consist of daily minimum and maximum temperatures recorded from October 1974 through December 1985. These differences were found to be statistically significant (p < 0.001). Histograms suggest that minimum temperature differences are more consistently positive and narrowly distributed, i.e. the urban station presents the nocturnal heat island effect more frequently
Fig. 6. Paisley historical data e monthly minimum temperatures (1959e2009).
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
143
Fig. 7. a) Paisley historical data: monthly mean maximum and minimum temperatures e winter periods (1959e2009). b) Paisley historical data: monthly mean maximum and minimum temperatures e summer periods (1959e2009).
Table 3 Trend lines and variations in local air temperature in Paisley (1959e2009). Period
All months
Winter months (JaneFeb) Summer months (JulyeAugust)
Trend line equation
Relative temperature rise in the 50-year period (from trend lines)
Actual temperature increase after 50 years Average minimum
Average minimum
Average maximum
Average minimum
Average maximum
Average maximum
y ¼ 0.0015x þ 5.6196 R2 ¼ 0.0043 y ¼ 0.0337x þ 0.7088 R2 ¼ 0.1188 y ¼ 0.0233x þ 10.984 R2 ¼ 0.161
y ¼ 0.0023x þ 11.968 R2 ¼ 0.0068 y ¼ 0.0487x þ 5.4686 R2 ¼ 0.2768 y ¼ 0.0296x þ 18.395 R2 ¼ 0.1165
0.9
1.4
0.6
1.1
1.7
2.4
1.8
1.9
1.2
1.5
0.7
0.2
144
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
to the mean, for the whole period, encompassing 2649 lines of data, and for selected winter and summer periods of that sample. Average differences are more pronounced for the daily minimum temperatures, which also has the smallest standard deviations. The CVs further corroborate the observation made above regarding both histograms, that there is a higher consistency in the ‘urban’e‘rural’ temperature differences in the daily minima. However, it is for the winter periods that the nocturnal heat island effect is more clearly pronounced. Fig. 10 shows the frequency of nights with UrbaneRural temperature difference were above a set value of 2 K. Data for two summer months and two winter months are shown as columns, whereas the dotted line shows results for all months.
a 100% 90% 80%
Frequency
70% 60% 50% 40% 30% 20% 1960-1969 2000-2009
10%
4.3. Effect of land use/land cover and local climate
0% -2
b
0
2
4
6
8
10
12
14
We selected 261 days of measurements in 2010 for which data from at least 4 of the weather stations surrounding the city of Glasgow were available. Given their relatively consistent distance from the city core, we hypothesised that the differences between the stations were the result of local land cover/land use conditions. We used the local climate zones (LCZ) classification system [4] to categorise the land cover surrounding of each of the selected weather stations. Fig. 11 shows the land cover characteristics of an area approximately 500 m 500 m surrounding each of the weather stations. We used these images to estimate “regions of uniform surface-air temperature distribution at horizontal scales of 102e104 m” [4]. We used the Google Earth images shown in Fig. 11 and street camera views to estimate the land cover class. In case of mixed classes, we used fetch effect likely due to the prevailing winds (south-westerly). LCZ classes for the four stations are:
22
1. Brancumhall: located closer to suburban East Kilbride than to Glasgow City in the county of South Lanarkshire. There is considerable urban development south-west of the weather station, mostly with lowrise buildings. Accordingly, we assigned an LCZ class of OPEN-SET LOWRISE 2. Glasgow Airport: Located at the Glasgow International Airport in Paisley, on the north side of the main airport terminal. LCZ classification: EXTENSIVE LOWRISE/LOW PLANT COVER 3. Renfrewshire: similar to Brancumhall, this location is surrounded by lowrise buildings in Neilston, East Renfrewshire. Although it can be also classified as OPEN-SET LOWRISE, it has substantially less obstructions to the south-west 4. Wishaw: Located close to Wishaw, North Lanarkshire, this site has a suburban aspect and is more vegetated than all the other
Monthly Minimum Temperature deg C (avg) 100% 90% 80%
Frequency
70% 60% 50% 40% 30% 20% 1960-1969 2000-2009
10% 0% 0
2
4
6
8
10
12
14
16
18
20
Monthly Maximum Temperature deg C (avg) Fig. 8. Comparison of cumulative frequencies e 1960e1969 and 2000e2009. (a) monthly minimum temperatures. (b) monthly maximum temperatures.
and consistently. However the maximum temperature difference has a wider frequency distribution and smaller amplitude. It was even negative in some instances (indicating a daytime cool island effect in the city centre). The average daily temperature fluctuation for both sites has an offset of almost one degree in the rural site (6.3 K at the rural site against 5.5 K at the urban site). Table 4 provides average differences, standard deviations and coefficients of variation (CV in %), defined as the ratio of the standard deviation
Fig. 9. Histograms for ‘urban’e‘rural’ temperature differences (1974e1985): left e daily minima; right e daily maxima.
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
145
Table 4 Temperature differences in degrees Celsius between Glasgow Weather Centre and Springburn for the period of 1974e1985. Delta T summer
Delta T winter
UrbaneRural (maxima) Average Standard Deviation Coefficient of Variation %
UrbaneRural (minima)
0.5 2.2
1.4 0.8
478
Delta T all periods
UrbaneRural (maxima)
UrbaneRural (minima)
1.2 1.8
56
UrbaneRural (maxima)
1.8 1.4
152
0.8 2.1
81
stations, having also a public park (Dalziel Park) and a golf course in the vicinity. LCS classification is mixed in this case: OPEN-SET LOWRISE/CLOSE-SET TREES/LOW PLANT COVER.
UrbaneRural (minima) 1.6 1.2
261
74
local differences. Fig. 12 reinforces the usefulness of the LCZ approach, which takes into account intra-urban or, in this case, peri-urban temperature differences. 4.3.2. Energy implications of local climate differences A simpler way to quantify the heating energy implications of local climate differences is to compare the Heating Degree Days (HDD) for the four weather stations (we used an HDD base line of 65 F or 18.3 C).Fig. 14 shows HDD trends sorted by LCZ class while Fig. 15 shows HDDs according to distance from the city centre. Since the data set available was incomplete, we normalised the HDDs per season by dividing the calculated HDDs by the number of days in a season for which data were available. Sorting locations relative to their LCZ classifications gives a clearer pattern than simply sorting them according to the proximity to the city core (Fig. 15). Trend lines in both graphs suggest a slight decrease in heating demand both when there is more urbanization and when the location is closer to the city core.
4.3.1. Temperature profiles Table 5 shows the annual and seasonal differences in Tmax, Tmin and the diurnal temperature range (DTR) for the 4 selected stations. The last row indicates the maximum differences obtained for each variable. As with the urbanerural comparison shown in Section 4.2 temperature differences between stations are more pronounced in the winter period and for Tmin. A further comparison of the departure from the group mean temperature shows the relative behaviour of each location compared to the mean (Table 6). Fig. 12 shows the correspondence between such figures for the daily minimum temperatures (indicator of the urban heat island effect) and the LCZ classification for each location, arranged left to right from the least to the most densely built location. A clear pattern arises suggesting the existence of urban warming in the more built-up areas. This pattern is observable annually as well as in winter and summer periods, although more pronounced in the winter. A comparison to a more simplistic approach, namely the urban heat island is related to the distance from the city centre (Fig. 13) shows that the distance theory is inadequate to explain the
4.3.3. Heat island profile on a clear-sky day Fig. 16 shows the temperature profiles from the four stations on a clear day. The criteria used for choosing the particular date was as follows: the most complete database; lowest average daily cloud cover, lowest average daily wind speed (below 1 m/s) and high daily temperature fluctuation (above 10 K). The date chosen for
% of days with Turb-rur>2K 100 Winter months Summer months All months
Frequency %
80
60
40
Oct-81
Dec-81
Aug-81
Apr-81
Jun-81
Feb-81
Oct-80
Dec-80
Aug-80
Apr-80
Jun-80
Feb-80
Oct-79
Dec-79
Aug-79
Apr-79
Jun-79
Feb-79
Oct-78
Dec-78
Aug-78
Apr-78
Jun-78
Feb-78
Oct-77
Dec-77
Aug-77
Apr-77
Jun-77
Feb-77
Oct-76
Dec-76
Aug-76
Apr-76
Jun-76
Feb-76
Oct-75
Dec-75
Aug-75
Apr-75
Jun-75
Feb-75
Oct-74
0
Dec-74
20
Fig. 10. Frequency patterns for a nocturnal (daily minima) temperature difference between the urban and rural locations Note: Only those days with an urbanerural difference greater 2 K are shown.
146
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
Fig. 11. Land cover characteristics of 500 500 m area surrounding each of the weather from top left, clockwise: a) Brancumhall, b) Glasgow Airport, c) Renfrewshire, d) Wishaw Source: Google Earth.
analysis was 11 April 2010, a day with clear skies and a solar radiation profile closely matching the theoretical maximum. It can be seen from Fig. 16 that a significant difference between locations is more noticeable at night, confirming the nocturnal characteristic of the heat island phenomenon. The most urbanized location (solid line) has the smallest diurnal temperature range and the highest minimum temperature for the day, in striking contrast to the site with the least built-up area (Glasgow Airport, grey dotted line). 5. Discussion Our results from long term temperature records (see Section 4.1) are similar to the climate change reported by the UK Climate Impact Programme (UKCIP’09) [24]. While the UKCIP’09 reported a prominent warming trend in South East England, it also showed some warming in the West of Scotland. Such changes are usually higher in winter than in summer months (increases in average daily maxima: 1.37 K in summer and 1.57 K in winter; increases in average minima: 1.34 K in winter and 1.23 K in summer). Our
Table 5 Annual, winter and summer average maximum (Tmax) and minimum temperatures (Tmin) and corresponding diurnal temperature ranges (DTR). Location
Annual Tmax
Brancumhall Glasgow Airport Renfrewshire Wishaw Maximum difference
10.2 10.8 10.5 10.4 0.6
Tmin 3.8 2.5 3.3 1.9 1.9
Winter DTR 6.4 8.3 7.1 8.5 2.1
Tmax 4.1 4.7 4.1 4.6 0.6
Tmin 1.0 2.9 1.6 3.8 2.8
Summer DTR 5.1 7.6 5.7 8.4 3.3
Tmax 17.9 18.5 18.4 18.1 0.6
Tmin 11.1 10.0 10.4 9.7 1.4
DTR 6.8 8.5 8.0 8.5 1.7
results suggest that urban areas superimpose a local warming effect over this regional trend. Superimposing our long term trend analysis (from section 4.1) on the climate change analysis provided by Barnet et al. [25] for the West, East and North of Scotland shows many similarities (Fig. 17), although air temperatures in our case were generally higher than the regional average. Both the minimum and the maximum air temperatures had a consistent shift of about 1.5 K. Within the period around 1985e1995, minimum temperatures in our case diverged from regional trends, while still maintaining higher values than the general regional trend. With regard to heating degree days (HDD), the regional trends reported by Barnet et al. [25] for a HDD with a base temperature of 60 F (15.5 C) showed a 11.3% decrease in 1961e2004. Our analysis of Paisley data show that HDD dropped from 2198 HDD to 1945 HDD during the same period, which corresponds to a decrease of 11.5%. Thus, the observed HDD decrease may be due to the regional warming. However, looking more closely at absolute values for ambient temperatures near Glasgow, Fig. 18 shows a comparison between Table 6 Annual, winter and summer departures from the group mean maximum (Tmax) and minimum temperatures (Tmin) and corresponding diurnal temperature ranges (DTR). Locations
Brancumhall Glasgow EGPF Renfrewshire Wishaw
Annual
Winter
Summer
Tmax
Tmin
DTR
Tmax
Tmin
DTR
Tmax
Tmin
DTR
0.3 0.3 0.0 0.1
0.9 0.4 0.5 1.0
1.2 0.7 0.5 0.9
0.3 0.3 0.2 0.2
1.3 0.6 0.7 1.5
1.6 0.9 1.0 1.7
0.3 0.3 0.2 0.1
0.8 0.3 0.1 0.6
1.1 0.6 0.1 0.5
1.5
Departure from group mean in deg C
Annual
Winter
Summer
1.0
0.5
0.0
-0.5
-1.0
-1.5
-2.0 EXTENSIVE LOWRISE/LOW OPEN-SET PLANT COVER LOWRISE/CLOSE-SET TREES/LOW PLANT COVER
OPEN-SET LOWRISE
OPEN-SET LOWRISE
LCZ classification Fig. 12. Annual, winter and summer departures from the group mean minimum temperature.
Departure from group mean in deg C
1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5
Winter
Annual
Summer
-2.0 11.3
12.1
13.6
22.4
Distance to city centre (Buchanan Street) in km Fig. 13. Annual, winter and summer departures from the group mean minimum temperatures (Tmin) sorted ordered according to proximity to city core. Annual
Winter
Linear (Annual)
Linear (Winter)
climatic normals for Glasgow Airport and West of Scotland (1971e2000, regional data from [22]). Clearly, local temperature is higher than regional averages. In addition, the number of days with night frost drops significantly, from around 70 to 36 days per annum. This fact points to the possibility of using the heat island in cities like Glasgow as an opportunity to enhance local energy savings and perhaps enhance the efficacy of district heating schemes. Additionally, the pairwise comparison between an ‘urban’ and a ‘rural’ weather stations (Section 4.2) indicates a clear heat island effect. It should be noted that our pairwise comparison (reflecting a 12-year average for the period of 1974e1985) has a weak agreement with the relevant figure for Glasgow in Table 1 (for 2003). Seasonal differences show that in general the nocturnal UHI is stronger in winter than in summer, even though it is more consistent in the summer season (CV 56%). Together with the fact that land cover characteristic appears to have an influence on local climate (Section 4.3), the indications are that the urban features in and around Glasgow are augmenting the slight regional warming in the West of Scotland region. This trend appears to have continued despite the loss of population in the city.
20.0 18.0
Annual
20.0
Winter
Linear (Annual)
Linear (Winter)
16.0
18.0 14.0
16.0 14.0 10.0
12.0
HDD
HDD
12.0
8.0 6.0
10.0 8.0
4.0
6.0
2.0
4.0
0.0 OPEN-SET EXTENSIVE OPEN-SET LOWRISE LOWRISE/CLOSE-SET LOWRISE/LOW PLANT TREES/LOW PLANT COVER COVER
OPEN-SET LOWRISE
LCZ Classification
2.0 0.0 11.3
12.1
13.6
22.4
Distance to city centre (Buchanan Street) in km Fig. 14. Normalized Heating Degree Days for the different locations (sorted by the LCZ class).
Fig. 15. Normalized heating degree days sorted by proximity to city core.
148
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
20 18
Temperature degC
16 14 12 10 8 6 4 2
12:00 AM
11:00 PM
10:00 PM
9:00 PM
8:00 PM
7:00 PM
6:00 PM
5:00 PM
4:00 PM
3:00 PM
2:00 PM
1:00 PM
12:00 PM
11:00 AM
10:00 AM
9:00 AM
8:00 AM
7:00 AM
6:00 AM
5:00 AM
4:00 AM
3:00 AM
2:00 AM
1:00 AM
0
Time of day OPEN-SET LOWRISE/CLOSE-SET TREES/LOW PLANT COVER EXTENSIVE LOWRISE/LOW PLANT COVER OPEN-SET LOWRISE OPEN-SET LOWRISE Fig. 16. Temperature profile for 11 April 2010.
Fig. 17. a) Annual maximum and b) Annual minimum temperature averages for Paisley (thick grey line), plotted against the three climate regions in Scotland (from Barnet et al., 2006).
R. Emmanuel, E. Krüger / Building and Environment 53 (2012) 137e149
b
25
16 14
20
Number of days
Temperature degC
a
149
12
15
10
10 5 0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
8 6 4 2 0
-5
Month Max Temp [°C] Scotland Max Temp [°C] Paisley
Jan Feb Mar Apr May Jun
Jul Aug Sep Oct Nov Dec
Month Min Temp [°C] Scotland Min Temp [°C] Paisley
Days of Air Frost Scotland
Days of Air Frost Paisley
Fig. 18. a) Ambient temperature averages for Paisley and Scotland from climatic normals 1971e2000 and b) number of days with air frost for Pasiley and Scotland.
6. Planning and energy implications The planning and energy implications of our findings to ‘shrinking cities,’ are intriguing. ‘Shrinking cities’ e a concept initially theorised in the wake of German unification [26] is an increasingly common reality in many parts of the world. Over the last fifty years, 370 cities throughout the world with populations over 100,000 have shrunk by at least 10% [27]. These are more common in the industrial heartlands of the USA (59 cities), Britain (27), Germany (26), Italy (23), Russia (13), South Africa (17) and Japan (12). They are also common in other parts of the world, even as growing cities continue to dominate the discourse. A typical planning approach to this crisis is to reconceptualise decline as shrinkage and to explore creative and innovative ways for cities to successfully shrink. Such approaches have usually taken the form of land for recreation, urban agriculture, green infrastructure, and other non-traditional land uses beneficial to existing residents and attract future development [28]. In their drive towards sustainable and ecologically sound places shrinking cities will need to consider the local climate implications of their current urban trajectories. While population may decline the underlying urban morphology largely remains in place, leading to the continuation of the urban climate anomaly. In the case of a temperate climate city such as Glasgow, the association between urban form and the ‘urban’e‘rural’ temperature differences continue to hold: this aspect of shrinking is beneficial, in that the urban warmth created by a judicious arrangement of land use/land cover (as evidenced by the appropriate LCZ class) could be exploited for energy efficient uses such as district heating and to enhance the feasibility of low carbon options such as district ground source heating or other communal renewable technologies. Our work shows that the UHI itself does not go away, even in shrinking cities, thereby the opportunities to be sustainable and low carbon might still be available. At the same time, the summertime trends suggest that overheating may become a distinct possibility in the future. These realities should inform shrinking cities in their attempt to reinvent themselves in a carbon and energy efficient fashion.
References [1] United Nations, Department of Economic and Social Affairs, Population Division. World urbanization prospects: the 2009 revision; 2010. CD-ROM Edition e Data in digital form (POP/ DB/WUP/Rev.2009). [2] Grimmond CSB, Roth M, Oke TR, Au YC, Best M, Betts R, et al. Climate and more sustainable cities: climate information for improved planning and management of cities (producers/capabilities perspective). Procedia Environ Sci 2010;1:247e74. [3] Mills G, Cleugh H, Emmanuel R, Endlicher W, Erell E, McGranahan G, et al. Climate information for improved planning and management of mega cities (needs perspective). Procedia Environ Sci 2010;1:228e46. [4] Stewart ID, Oke TR. A new classification system for urban climate sites. Bull Am Met Soc 2009;90:922e3.
[5] Fujibe F. Detection of urban warming in recent temperature trends in Japan. Int J Climatol 2009;29:1811e22. [6] Fujibe F. Urban warming in Japanese cities and its relation to climate change monitoring. Int J Climatol 2011;31:162e73. [7] Grimmond CSB. 2011. London’s urban climate: historical and contemporary perspectives, paper presented at the conference City Weathers: Meteorology and Urban Design 1950-2010, June 23, 2011 http://www.sed.manchester.ac. uk/architecture/research/csud/workshop/programme/ [accessed 01.07.11]. [8] Kolokotroni M, Davies M, Croxford B, Bhuiyan S, Mavrogianni A. A validated methodology for the prediction of heating and cooling energy demand for buildings within the Urban Heat Island: case-study of London. Solar Energy 2010;84:2246e55. [9] Kershaw T, Sanderson M, Coley D, Eames M. Estimation of the urban heat island for UK climate change projections. Building Serv Eng Res Technol 2010;31:251e63. [10] Thorsson S, Lindberg F, Björklund J, Holmer B, Rayner D. Potential changes in outdoor thermal comfort conditions in Gothenburg, Sweden due to climate change: the influence of urban geometry. Int J Climatol 2011;31:324e35. [11] United Nations Population Fund (UNFPA). Growing up urban. Supplement to the state of world population 2007. New York: UNFPA; 2007. [12] Vörösmarty CJ, Green P, Salisbury J, Lammers RB. Global water resources: vulnerability from climate change and population growth. Science 2000;289: 284e8. [13] Hansen J, Ruedy R, Sato M, Imhoff M, Lawrence W, Easterling D, et al. A closer look at United States and global surface temperature change. J Geophys Res 2001;106(D20):23947e63. [14] Jin M, Dickinson RE, Zhang D-L. The Footprint of urban areas on global climate as characterized by MODIS. J Clim 2005;18:1551e65. [15] Angel S, Sheppard SC, Civco DL. The dynamics of global urban expansion. Washington, D.C.: Transport and Urban Development Department, the World Bank; 2007. [16] Lamptey B. An analytical framework for estimating the urban effect on climate. Int J Climatol 2010;30:72e88. [17] Parker DE. Urban heat island effects on estimates of observed climate change. WIRES Clim Change 2010;1:123e33. [18] IPCC. Climate change 2007: the physical science basis. In: Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt K, Tignor M, Miller HL, editors. Contribution of working group I to the fourth assessment report of the IPCC. New York: Cambridge University Press; 2007. p. 1056. [19] Maver I. Glasgow’s public parks and the community, 1850e1914: a case study in Scottish civic interventionism. Urban Hist 1998;25:323e47. [20] Mackenzie JM. The second city of the Empire: Glasgow e imperial municipality. In: Driver F, Gilbert D, editors. Imperial cities: landscape, display and identity; 1999. p. 215e37. Manchester. [21] Cameron EA. Glasgow’s going round and round: some recent Scottish urban history. Urban Hist 2003;30:276e87. [22] UK Meteorological Office (UK Met Office). Monthly and annual average climate data at 5 km resolution. Can be accessed at: http://www.metoffice. gov.uk/climate/uk/stationdata/; 2011 [last accessed 29.06.11]. [23] Oke TR. Boundary layer climates. London: Methuen; 1987. [24] Murphy JM, Sexton DMH, Jenkins GJ, Boorman PM, Booth BBB, Brown CC, et al. UK climate projections science report: climate change projections. Exeter: Met Office Hadley Centre, http://www.ukcip.org.uk/ukcp09/; 2009 [accessed 21.07.11]. [25] Barnett C, Hossell J, Perry M, Procter C, Hughes G. Patterns of climate change across Scotland: technical report. 102 p. In: SNIFFER Project CC03. Scotland & Northern Ireland Forum for Environmental Research, http://www.sniffer. org.uk/Webcontrol/Secure/ClientSpecific/ResourceManagement/UploadedFiles/ CC03_Final_report.pdf; 2006 [accessed 07.07.11]. [26] Rieniets T. Shrinking cities: causes and effects of urban population losses in the twentieth century. Nat Cult 2009;4:231e54. [27] Oswalt BP, Rieniets T. Shrinking cities: global study. Can be accessed at, http:// www.shrinkingcities.com/globaler_kontext.0.html?&L¼1; 2007 [last accessed 30.06.11]. [28] Hollander JB, Pallagst K, Schwarz T, Popper FJ. Planning shrinking cities. available at: http://policy.rutgers.edu/faculty/popper/ShrinkingCities.pdf; 2009 [last accessed 3006.11].