The inclusion of the urban heat island effect in design weather for net-zero buildings

Abstract

Dry bulb air temperatures are increasing especially in the UK. On top of this there is the urban heat island intensity (UHII), the temperature difference between rural and hotter urban temperatures. This is important as most new and existing buildings are in urban areas subject to the urban heat island (UHI)., In this paper is an analysis of weather data in Manchester city centre is compared to the CIBSE Design Summer Year type 3 (DSY 3) weather data. The latter, based on weather data from the Meteorological Office at a non-urban contains little if any UHII data. The DSY3 is based on a very warm year but the lack of the UHII data severely underestimates city centre night time temperatures and overestimates the winter minima temperatures. This underestimates the heating effect for net-zero design for urban buildings. A simple algorithm, which has been derived from Manchester and London data, is shown to give a useful method of adding in the UHII.

Practical application

Some standard weather files underestimate night-time urban temperatures and this has potential for inaccurate heating and cooling predictions for new net-zero building designs. By comparing Manchester city-centre data with CIBSE DSY3 and applying a simple urban heat island intensity algorithm, designers can more accurately model real urban conditions, improving the reliability of thermal performance assessments and supporting better-informed building design decisions.

Keywords

Urban heat island design summer year manchester climate change

Introduction

CIBSE for its design weather data for designing buildings plant and assessing energy (contained in Guide A and updates¹) uses reliable UK Meteorological Office (Met Office) data. But Met Office sites are in flat grassy areas away from urban areas so that buildings and trees do not influence the measurements. Airports are almost ideal, but most buildings are in urban and city areas. In these urban areas there is often a rise in temperature due to the urban heat island (UHI) effect. The effect is measured by the UHI intensity (UHII) the difference between the urban temperature and the Met Office site closest. This paper examines the temperature data, both rural and city centre for Manchester, UK and compares it to the CIBSE Design Summer Year (DSY) for Manchester. It finds the lack of the UHII data in the DSY is a serious impediment to accurate design. The paper also finds that the UHII is not increasing significantly after the Covid blip even with the rapid change in the construction of tower blocks in Manchester City centre.

The urban heat island intensity (UHII)

The urban heat island² is generated by a number of factors, including the absorption of solar irradiance on urban fabric and its heat storage within the urban fabric as well as the heat generated from human activity in the urban areas, including the outputs from heating and cooling systems.

Urbanisation is increasing in the world.³ The greater the urbanisation the greater the increase in the urban heat island intensity (the difference between urban and the rural temperatures).³ This has been shown to be happening in Manchester City centre in a previous paper.⁴ Hence the concern of the UHII adding to the increasing climate change temperatures of the future.

Figure 1 shows the daytime UHII in the centre of Manchester (China Town). The China Town air temperatures are measured by a temperature sensor, an i button which communicates using the 1-Wire protocol, requiring momentary contact with a reader to exchange data. The i button is hooked into a radiation shield made in the university. The shield is attached to a lamppost about 3m above the pavement. About 60 shields and i buttons were placed around Greater Manchester. Full details are in are in Cheung.^5,6 Figure 2 shows an i button and a shield. The rural air temperature is from the nearest Met Office site, Rostherne, in rural Cheshire.

Figure 1.

China town UHII daytime 2022.

Figure 2.

i button on base plate and hook which is hooked into the radiation shield, right, attached to a lamppost by ties around the aluminium bar, right, and the lamppost.

Measurements of the accuracy of the 60 i buttons used in the project yielded a standard deviation of 0.034°C (at 30°C test temperature, warm test) and 0.021°C (at 5°C test temperature, cold test).⁵ Comparing a representative i button with a calibrated i button (certified to standard EN 13,005 and ISO 17,025) showed a difference of 0.026°C (warm test) and −0.333°C (cold test). The “clock” in the i button did drift over time and each download (about every 3 months) the clock was reset. The Met Office use platinum resistance thermometers for temperature measurement with an accuracy of ±0.1°C.⁷ However, for both the Met Office sensors and the Manchester i buttons, they were placed in a Stevenson screen and radiation shield respectively. The radiation shield performed well against the Stevenson screen.⁵

Referring to Figure 1, the highest frequency of the UHII is 1K with values going up to 4K and down to -1K (the urban area for a very short time actually providing an urban cool island).

Figure 1 is for the daytime (defined below). Figure 3 shows the UHII for China Town for the nighttime (defined below). It is a very different distribution from Figure 1. This is primarily because on clear, cloudless nights there is radiation from the earth to the clear sky through the earth’s infrared (8 μm to 11 μm) atmospheric window.⁸ Without the solar irradiance the atmosphere is more stable and a temperature inversion occurs over the rural site. However, this is far less pronounced over the urban area where the buildings shield the view of the “cold” sky. In fact, in China Town the sky view factor, or the fraction of a hemispherical sky that can be seem from street level is 37%. Added to which the stored heat from the surrounding buildings keeps the air temperature warmer. As a result, the UHII values are higher, the modal value being 1.5 K. The daytime is defined as the when the measured global horizontal irradiance over the hour is ≥50 kJ m⁻²h⁻¹, nighttime as <50 kJ m⁻²h⁻¹. Hence the frequency is related to the 4126 daylight hours and 4634 night time hours. The nighttime modal value in the 1.5 K bin has a frequency of 1,034 hours whereas the daytime modal value in the 1.0 K bin is 600 h.

Figure 3.

China town nighttime UHII 2022.

Design summer years

For design using simulation hourly values of meteorological data is required. With modern highly insulated buildings overheating in summer is now a problem rather than underheating in winter. For summer conditions a near-extreme year is required for adequate design to assess overheating in the UK.⁹ Overheating risk assessments are conducted by using dynamic thermal simulation models running a near-extreme year, the CIBSE Design Summer Year (DSY). The DSY is a historical year, originally selected on the basis of being the year with the third warmest April–September period from a set of 20 years. Originally the DSY for London was 1989 and the weather data used was from the Met Officeffice London Heathrow Airport site. However, a better method of assessing the DSY has now been developed taking account of likely thermal discomfort. The method is fully explained in the CIBSE (2014) TM 49: Design Summer Years for London.⁷ TM 49 now defines three DSYs. One with a moderately warm summer, one a year with a long period of persistent warmth, and a third with a more intense single warm spell. As the method is also used with future climate data from the UKCP09 Climate Change Projections¹⁰ the DSYs are probabilistic DSYs or pDSYs.¹¹ The moderate summer one is referred to pDSY-1, the “long period of persistent warmth” pDSY-2 and the third, pDSY-3.

The method of determining the pDSYs has been subsequently been improved¹² and more modern climate change data based on UKCP18.¹³ New pDSYs have also been determined for other cities around the UK. Although most of them are probabilistic for the future there are those based on historical data, referred to as DSYs. The historical data has been updated¹⁴ to a new baseline, 1994-2023, which has produced new DSYs. For Manchester the new DSYs are: DSY-1 1983, DSY-2 1976 and DSY-3 2022 based on weather data from the Met Office site Hawarden, an airport site near Chester. As the Manchester City centre measurements undertaken by the authors only started in 2011, only the DSY-3 for Manchester can be compared.

Figure 4 compares the Met Office Hawarden for January 2022 with measurements in the centre of Manchester (China Town). Figure 5 shows the two sites for July 2022. The differences, especially at the peaks and troughs, are significant.

Figure 4.

China town and Hawarden January 2022.

Figure 5.

China town, Hawarden July 2022.

Tables 1 and 2 show the yearly differences for 2022 between China Town and the rural sites. Rostherne (Ros) is nearer to Manchester and is the site used by the authors to measure the UHII. Again the difference between China Town from Hawarden is highlighted, it being significantly warmer in China Town during both day and night.

Table 1.

Night time China Town versus Rural sites 2022 (DBT (°C) = air dry bulb temperature).

	China town DBT (^°C)	Hawarden DBT (^°C)	UHII (Ros)	Rostherne DBT (^°C)
Average	10.7	8.8	2.5	8.2
Standard deviation	5.05	5.09	1.84	5.25
Minimum	−4.5	−8.6	−3.4	−11.3
Maximum	31.5	24.6	10.0	26.4

Table 2.

Day time China Town versus Rural sites 2022 (DBT (°C) = air dry bulb temperature).

	China town DBT (^°C)	Hawarden DBT (^°C)	UHII (vs Ros)	Rostherne DBT (^°C)
Average	14.7	13.8	1.0	13.7
Standard deviation	5.6	5.4	1.0	5.7
Minimum	−3.3	−4.7	−3.0	−7.2
Maximum	37.7	36.6	8.1	37.4

London pDSY sites

The first pDSYs were developed for London. To include the UHI the London Weather Centre site’s data was used. This site was based on the rooftop of a building in Holborn. It was not an ideal meteorological site and has been discontinued. However, it probably included a significant UHII and represented an “inner urban” site.⁷ The main site used by CIBSE for London was the Met Office Heathrow site, at the airport. This indeed has its own UHI¹⁵ and the Met Office Gatwick site is used as a rural site.

Manchester pDSY site

The Met Office Manchester site used for determining the pDSYs is at an airport at Hawarden, near Chester. The authors for their UHI work used the Met Office Rostherne site, in the middle of a large field, closer to Manchester city centre. However, the Rostherne site has only been used since 2011 and so does not have the 30-year period needed for analysing near extreme years, whereas Hawarden has been in use for over 30 years. Figure 6 shows that there is little difference between the sites for 2022, although the average monthly minima temperatures for Rostherne are significantly lower than for Hawarden. 2022 is shown as that is the year chosen for the CIBSE DSY-3. The DSY -1 and DSY-2 are years before the Manchester UHI work by the authors started.

Figure 6.

Rostherne and Harwarden average monthly values for 2022, the Manchester DSY - 3

Comparison of the CIBSE manchester DSY - 3 (2022) with manchester city centre data for 2022

Figures 4 and 5 show the differences for January 2022 and July 2022 respectively. Figure 7 shows the average monthly values for 2022 for China Town and Hawarden. It also shows that the minimum average temperatures are lower for Hawarden, whereas the average maximum values are very similar.

Figure 7.

Manchester city centre (China Town) and Harwarden average monthly values for 2022, the Manchester DSY - 3

The consequences for design of not using city centre data for city centre buildings

The Manchester City centre air temperatures are warmer than the rural CIBSE DSY – 3. But does this make much difference to the design of buildings in Manchester? An indication of the error can be addressed by comparing the cooling degrees hours. The cooling degree hours are the number of hours that the temperature is above a base temperature, it being assumed that above this temperature the air conditioning is switched on or the room becomes uncomfortable.

Cooling degree hours comparison

Figure 8 shows the cooling degree hours to the bases of 25°C (clg25) and 22°C (clg22) for China Town, Rostherne (Ros) and Hawarden. They are for complete days rather than separating day and night.

Figure 8.

Cooling degree hours for China Town, Rostherne (Ros) and Hawarden 2022.

The clg25° h for China Town, Rostherne and Hawarden are 607 Kh, 429 Kh and 371 Kh. The Hawarden cooling degree hours are almost 40% lower than those for China Town.

Heating degree hours comparison

Figure 9 shows the heating degree hours. As China Town is warmer the heating degree hours are less than for Hawarden and Rostherne. The heating degree hours base temperature (below which the hours are counted) is taken as 18°C. This gives China Town 52,078 Kh compared to 66,118 Kh for Rostherne and 62,561Kh for Hawarden. This would correspond to a saving in heating energy in the city centre compared to the rural sites. At all sites the heating degree hours are much higher during the nighttime.

Figure 9.

Heating degree hours for China Town, Rostherne (Ros) and Hawarden 2022.

Adding in the approximate urban heat island intensity to the CIBSE DSY – three

Although it has been shown that there is a need for adding in the UHI for most CIBSE UK weather data (London has some UHI in its pDSY) more data can only be found from significant UHI measurements in urban sites. However, as shown in a previous paper¹⁶ there can be differences of up to 0.5 K between Manchester City centre sites. A simple approximate algorithm that works for London and Manchester can be applied¹⁷ to the hourly data.

The model has the simple form:

UHII = w AW + c AC + s AS + k

where AW = (6 – WS), where WS is the wind speed in knots, AC = (8 – cloud cover in oktas), AS is the global short-wave irradiation (KJ m⁻² h⁻¹).

The three coefficients w,c,s and the constant k are now determined by fitting the model to the measured UHII. This was done for Manchester¹⁴ for the year 2014 giving s = 0.001, w = 0.60, c = 0.23 and k = 0.92.

Empirical model parameters (fitted to the data by a least squares error minimisation (e.g. Excel’s Solver)) for London Weather Centre–Heathrow¹⁸ gave s = 0.001 w = 0.326 c = 0.117 k = 0.182. However, there is a significant UHI at Heathrow itself and a better rural site is Bracknell. Table 3 shows that there is a significant similarity between the model parameters for Manchester- Rostherne and London Weather Centre (LWC)-Bracknell.

Table 3.

Empirical model parameters for London Weather Centre - Bracknell and for Manchester - Rostherne.

Parameters	Parameter symbol	LWC – Br average	Manchester – Rostherne, 2014
Short wave irradiation	s	−0.007	−0.001
Wind speed	w	0.67	0.6
Cloud cover	c	0.22	0.23
Constant	k	0.89	0.92

As the cities are quite different (size, population, rivers and green spaces) this bodes well for applying the model to other UK cities. However, as the weather for each year can be quite different (heat waves, rain, storms etc) the model for each year may vary a little. The data for London is rather dated as the London Weather Centre closed

Using the Manchester-Rostherne parameters in Table 3 for the UHII model and applying it to the Rostherne data (wind speed, cloud cover etc), Figure 10 shows the excellent overall correlation (r² = 0.55, t = 104, 0.1% level of significance) between the 2022 Rostherne year modelled UHII and the actual UHII in Manchester City centre.

Figure 10.

The measured Manchester City UHII and the UHII modelled from the Rostherne data, all for 2022.

Design implications

The heating degree hours are lower in the city centre (China Town, Figure 8) indicating a lower heating consumption for the city centre. However, the cooling degree hours are higher indicating a higher cooling consumption.

The design implications for plant sizing can be examined by the design percentiles in Table 4 for 2022.

Table 4.

Design percentiles for 2022 for rural Rostherne and City Centre, China Town.

Rosthern	Rosthern	Rosthern	China town	China town	China town
%	Date	Air temp (°C)	%	Date	Air temp (°C)
0.01	15/12/2022 07:00	−11.3	0.01	15/12/2022 05:00	−4.5
1.00	13/12/2022 21:00	−3.7	1.00	15/12/2022 13:00	0.3
2.50	02/04/2022 05:00	−1.2	2.50	25/01/2022 11:00	2.2
5	18/03/2022 06:00	0.7	5	07/12/2022 13:00	3.7
10	10/02/2022 07:00	3.2	10	01/12/2022 21:00	5.5
50	18/10/2022 20:00	10.7	50	30/09/2022 23:00	12.6
75	04/07/2022 08:00	15	75	08/08/2022 01:00	16.4
90	26/03/2022 13:00	18.7	90	16/08/2022 09:00	19.6
95	21/08/2022 17:00	21.2	95	22/06/2022 12:00	21.7
97.50	13/08/2022 09:00	25.3	97.5	15/08/2022 11:00	23.7

The lower percentiles show that the city centre is indeed warmer than the rural area. At the 5^th percentile and 10^th percentile the city centre heating system are 15% and 10% respectively oversized based on the rural data. For the cooling the 90^th percentile and the 95^th percentile the city centre cooling system are 16% and 6% undersized. It is interesting that the rural and city centre percentiles are often on different days and different times. This is because maximum urban heat island intensity (UHII) is often just before sunrise. It should also be noted that high UHII values often occur on cloud free days producing sunny days. The solar gain to some buildings is considerable.

Conclusion

Comparing the DSY – three for Manchester, which is the year 2022, using the Met Office Hawarden site data, to the data measured in the City Centre by the authors, shows that the urban heat island effect in Manchester makes the city significantly warmer in both the summer and winter. Not taking the UHII into account will produce poor designs for buildings and plant in the city centre. As all the other UK city DSYs do not include data for the UHII, apart from London, similar design errors will occur. There is therefore a need to provide this data either by measurement of by modelling. However, in the meantime the use of the empirical UHII model, as shown in this paper, is a very good approximate solution.

Footnotes

Acknowledgments

There was no external funding for the work outlined in this paper. Thanks are due to the Centre for Environmental Data Analysis (CEDA) with the Met Office MIDAS Open: UK Land Surface Station Data for dry bulb temperatures and solar irradiance data.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Geoffrey Levermore

References

CIBSE Guide

. Environmental design. London: Charted Institution of Building Services Engineers, 2015. updated 2021 (2021) (PDF).

Oke

. Boundary layer climates. Routledge, 2006.

Developments and forecasts on continuing urbanisation. https://knowledge4policy.ec.europa.eu/foresight/topic/continuing-urbanisation/developments-and-forecasts-on-continuing-urbanisation_viewed 24 Feb 2023.

Levermore

Parkinson

. Covid-19 and other influences on the urban heat island of Manchester. CIBSE Technical Symposium, 2020.

Cheung

HKW

Levermore

Watkins

. A low cost, easily fabricated radiation shield for temperature measurements to monitor dry bulb air temperature in built up urban areas. BSERT 2010; 31(Issue 4): 371–380.

Cheung

HKW

. An urban heat island study for building and urban design. Manchester, UK: University of Manchester, 2011. PhD Thesis.

Parker

Horton

. Uncertainties in the central England temperature series 1878-2003 and some changes to the maximum and minimum series. Office 2004; 25(9): 1173–1188.

Stull

. Meteorology today for scientists and engineers. USA: West Publishing Co, 1995.

CIBSE (2014) TM 49: design summer years for London.

10.

Jenkins

. The UKCP09 climate change projections for the UK. London: Exeter: Met Office Hadley Centre, 2009.

11.

Eames

. An update of the UK’s design summer years: probabilistic design summer years for enhanced overheating risk analysis in building design. Building Serv. Eng. Res. Technol 2016; 37(5): 503–522.

12.

Virk

Eames

(2016) CIBSE weather files 2016 release: Technical briefing and testing.

13.

Met Office (2022) Probabilistic projections UK CP18. https://www.metoffice.gov.uk/research/approach/collaboration/ukcp/science/probabilistic-projections viewed 26/10/2024.

14.

Eames

. Private communication. In: addition we just redid this same work with an update in the baseline to 1994-2023. The years became 1983, 1976, and 2022, 2024.

15.

Levermore

Vandaele

Parkinson

. Temperature adjustments for design data for urban air conditioning design. Technical note. Build Serv Eng Res Technol December 2018; 20: 211–218.

16.

Levermore

Parkinson

. The need to include urban heat island data in design weather data for resilient, net-zero buildings. London: CIBSE Symposium 2025. (Paper 17).

17.

Levermore

Parkinson

. An empirical model for the urban heat island intensity for a site in Manchester. Build Serv Eng Res Technol 2017; 38: 21–31.

18.

Levermore

Parkinson

. (2019) The urban heat island of London, an empirical model. Build Serv Eng Res Technol 2019; 40(3): 014362441882287.