This item was submitted to Loughborough’s Institutional Repository
(https://dspace.lboro.ac.uk/) by the author and is made available under the
following Creative Commons Licence conditions.
For the full text of this licence, please go to:
http://creativecommons.org/licenses/by-nc-nd/2.5/
Green Open Access Version
Summertime temperatures and thermal comfort in UK homes
Lomas KJ* and Kane T
School of Civil and Building Engineering, Building Energy Research Group, Loughborough
University, Leicestershire, LE11 3TU, UK.
k.j.lomas@lboro.ac.uk, t.kane@lboro.ac.uk
*Corresponding author.
Total words – 10,900
Abstract
Abstract [213 words]
Internal summertime temperatures measured in 268 homes in the UK city of Leicester are
reported. The hourly data was collected from living rooms and bedrooms during the summer
of 2009, which was generally cool but with a short hot spell. Some household interviews
were conducted. The sample of homes is statistically representative of the socio-technical
characteristics of the city’s housing stock. The data provides insight into the influence of
house construction, energy system usage and occupant characteristics on the incidence of
elevated temperatures and thermal discomfort.
The warmest homes were amongst the 13% that were heated. Significantly more of these
were occupied by those over 70 who are particularly vulnerable to high temperatures. The
national heatwave plan might usefully caution against summertime heating.
Temperatures in the 230 free-running homes were analysed using both static criteria and
criteria associated with the BSEN15251 adaptive thermal comfort model. These indicated
that that flats tended to be significantly warmer than other house types. Solid wall homes and
detached houses tended to be significantly cooler.
It is argued that adaptive criteria provide a valuable and credible framework for assessing
internal temperatures in free-running UK homes. However, the temperatures in the Leicester
homes were much lower than anticipated by the BSEN15251 model. Numerous possible
reasons for this discrepancy are discussed.
Key words: Comfort, UK houses, temperature measurement, summer.
Introduction
Summertime temperatures in homes are of increasing concern, even in the relatively mild
climate of the UK because very high indoor temperatures can be life threatening and are
likely to occur more often as global temperatures rise. Whilst elevated temperatures can be
overcome with air conditioning, this would simply increase electricity use and be, for the UK
at least, a new source of greenhouse gas emissions. Conversely, many UK householders
switch their heating systems off in summer. Thus during cooler weather, internal
temperatures can be low, which, because the fabric of the home will be cool, creates a buffer
against hotter spells of weather. There is therefore an interest in understanding what
summertime temperatures are in UK homes and how these might be influenced by house
type, construction and occupant characteristics.
Heat stress in homes is of particular concern. The European heat wave of 2003, which has
been particularly closely studied, was most intense in the UK in August. It is estimated to
have caused an additional 2045 deaths in the UK (ONS, 2003) with as many as 70,000 excess
deaths between June and September, across Europe as a whole (WHO, 2007). During the
2003 heat wave, Wright et al (2005) measured the internal temperatures in five London flats
Green Open Access Version
and in four homes around Manchester. The heat wave lasted 9 days in Manchester and 12
days in London. In Manchester, the maximum daily mean temperature was 25.4
o
C with an
absolute peak of 32.1
o
C; corresponding values for London were 29.3
o
C
1
and 37.4
o
C. In both
locations, the daily average external temperature was always above 20
o
C, except for one day
in Manchester. In Manchester the living room of one home reached 30
o
C and one of its
bedrooms 36.0
o
C, and a London flat reached 37.9
o
C, which is dangerously high given the
imperative to maintain our core temperature very close to 37
o
C.
Whilst the summer of 2003 was very unusual for our current climate, projections indicate that
similar extreme weather events will take place every two or three years by the 2050s (Mayor
of London, 2008) and by the 2080s, such temperatures would be considered unusually cool
(Eames et al., 2011). There is therefore interest in knowing the extent to which UK homes
should be adapted to withstand higher summertime temperatures and whether adaptation is
necessary for all house types and in all geographical areas. One obvious adaptive measure is
to install air-conditioning, but this would simply increase summertime energy demands and
hinder progress towards a low-carbon future.
Using mortality data for the Greater London region for 1976 to 1996, Hajat et al (2002)
showed that an average daily external temperature over 19
o
C seems to lead to an increase in
heat related deaths. The rate of increased deaths was related to the degree to which the three
day moving average external air temperature exceeded the 97
th
centile value
2
. The use of a
moving average temperature enables hot spells rather than isolated hot days to be identified.
It also corresponds rather well to the way that indoor temperatures change with external
conditions; they tend to be influenced by the external temperature over the recent past rather
than the instantaneous external temperature. The use of a locally defined threshold, i.e. the
97
th
centile value, suggests that deaths due to a heat wave will be fewer in areas that are
generally warm, like the south east of England, than if a heat wave of the same intensity and
duration hit a region that is generally cooler, such as the more northerly areas of England.
Modern, adaptive thermal comfort standards published since the Hajat study, whilst
concerned with general comfort rather than heat stress, are based around a moving average
external temperature and have thresholds that increase as external temperature increases.
They are considered herein as a method for identifying householders at risk of exposure to
high temperatures.
The UK National Health Service has produced a heat wave plan contains advice on coping
with extreme weather events (NHS, 2011). Using data for the hot summer of 2006, it notes
that external temperatures over 25
o
C led to increased mortality. The greatest temperature-
related mortality occurred in the West Midlands, where there were 160 excess deaths between
16
th
and 28
th
July (an increase of 10% above the base level)
3
. Most at risk are: the elderly,
especially, women over 75; those living alone or the socially isolated; those with chronic or
severe illnesses; individuals that cannot take adaptive measures to keep cool, e.g. the bed-
bound, those with Alzheimer’s disease, or babies and the very young; and those living in
urban areas and in south-facing top floor flats. The heatwave plan causes warnings of
possible heatwaves to be issued in a region if there is a 60% or greater risk that forecasted
external temperatures will exceed listed threshold values for 2 or more days. For most of the
country, including the East Midlands, which is where the homes studied in this paper were
located, the listed thresholds are 30
o
C in the day and 15
o
C at night. The heatwave plan
1
In Leicester, the temperatures in August 2003 reached a peak of 37.0
o
C (on August 9
th
) with a maximum daily
mean of 25.2
o
C (on August 6
th
) and a maximum T
rm
value of 21.4
o
C (on August 11
th
).
2
In London the 97
th
centile value of the three day moving average temperature from 1976-96 was 21.5
o
C.
3
The South East and Yorkshire and Humberside experienced the second highest temperature-related death rates
with 110 excess deaths (4% above base level) and 100 deaths (6% above base level), respectively.
Green Open Access Version
provides an example of how internal temperatures can be reduced by insulating a home.
Taking cold drinks, checking temperatures, opening windows at night and moving to a cooler
room are all cited as measures to reduce overheating risk; turning off heating systems is not
explicitly stated
4
.
Modelling studies, for example by Hacker et al. (2008), have shown that thermal mass and
controlled ventilation can much improve summertime thermal comfort. In a thermally
massive home in the London region, bedroom temperatures were predicted to become
excessive in the 2080s, but in a lightweight home overheating was predicted to set in as early
as the 2020s. Similarly, Peacock et al. (2010) have shown that solid masonry wall homes are
more comfortable in summer than thermally lightweight dwellings, which, in the London
region, become uncomfortably hot by the 2030s. The ability to maintain bedrooms at a
comfortable temperature was noted as being of paramount importance in understanding
overheating risk. Mavrogianni et al. (2012) studied the impact of energy efficient
refurbishment on the internal temperatures of homes in London. They noted that retaining
exposed thermal mass and the ability to ventilate effectively would enable mean and peak
internal temperatures to be controlled up to the 2050s, but internal insulation that masked
thermal mass led to increased internal temperatures.
Although modelling studies are extremely useful, it is difficult to capture credibly the full
variability of occupant behaviour and house construction, geometry and ventilation potential.
In contrast, measurement can capture such diversity and, if the study is sufficiently detailed
and large enough, also relationships between those that are vulnerable to extreme
temperatures, such as the elderly, sick and the very young, and the homes in which they live.
There are, however, few large UK studies of summertime temperatures in homes; most large-
scale studies have focused on winter temperatures
5
. The energy follow-up survey
commissioned by the UK Department of Energy and Climate Change in 2010, to supplement
the data from the English housing survey, could help fill this gap.
This paper also contributes to our knowledge of actual summertime temperatures in UK
homes. To the authors’ knowledge it is the first large-scale and detailed study that has been
reported. It presents an analysis of the internal temperatures recorded in 268 homes in the UK
city of Leicester during the summer of 2009. The monitoring period contained a short period
of hot weather followed by two cool months. The spur to this paper was an analysis of
elevated temperatures, which was assessed using established ‘static’ overheating criteria.
Perhaps the more interesting results originate from the analysis of the thermal comfort
conditions during the cool weather, which was assessed using an adaptive model of thermal
comfort. The paper identifies any significant relationships between thermal comfort and
elevated temperatures and house type and age, construction, tenure and occupancy.
Household survey and temperature measurements
Leicester was the case study city chosen by the 4M project consortium that was concerned
with the determination of city carbon footprints (Lomas et al., 2006). Leicester is
geographically central in England and has a clearly identifiable boundary with the
surrounding rural area (Fig. 1). With a resident population of 280,000 in 2007, living in over
4
Hidden in a figure is advice to ‘reduce internal energy and heat’.
5
This is not surprising as in the UK wintertime space heating energy demands are a major source of greenhouse
gas emissions and under heating of homes is a significant health risk. Such studies include, for example, 1600
low income households (Oreszczyn et al., 2006), 427 homes in the CaRB study (Shipworth et al., 2009), 14 low-
energy homes monitored in Milton Keynes (Summerfield et al., 2007) and 25 households in Northern Ireland
(Yohanis et al., 2010).
The most extensive field survey (Hunt & Gidman, 1982) measured spot temperatures in
each room of 100 homes in February and March 1978.
Green Open Access Version
111,000 homes (ONS, 2010), Leicester is the UK’s 15
th
largest city and has households that
cover a wide range of socio-economic categories, from affluent to the most disadvantaged.
The most frequent housing types are semi-detached (37% of the city’s housing stock) and
terraces (35%), which proliferate towards the city centre along with flats (17%) (Table 1).
The detached houses are found primarily in the suburbs (10%) (ONS, 2010). Over the years,
many homes have been made more energy efficient using insulation and modern boilers and
controls.
One aim of 4M was to measure domestic energy use, travel behaviour and garden
management practices. To do this, a face-to-face computerised questionnaire was
administered at 575 homes (i.e. 0.5% of Leicester homes). These were randomly selected
after stratifying by percentage of detached homes and percentage with no dependent children
(Fig. 1), which is important here as the thermal comfort of the elderly is of interest
6
. The
questionnaire was devised by the 4M team and conducted on their behalf by the National
Centre for Social Research (NATCEN) between 17
th
March and 18
th
June 2009. Relevant to
this work, the survey captured the house type
7
, the number of occupants, the age of the oldest
occupant, the age of the house, whether the loft or walls were insulated or not, and the mode
of tenure. The responses of the interviewees were recorded directly onto a laptop and then
downloaded, cleaned and organised in the 4M database. The 4M Living in Leicester (LiL)
survey provides a consistent and comprehensive data set about households, their home energy
demand, travel behaviours and garden management practices. It is the first such data set
collected in the UK and has been exploited for a number of purposes.
As part of the LiL survey, Hobo pendant-type temperature sensors (Fig. 2) were used to
record internal temperatures over an eight-month period beginning on 1
st
July 2009. The
temperature measured approximates to air temperature, but, as the sensors were unshielded,
they will also record a radiant component. The primary purpose was to capture the internal
temperatures during the winter heating season (Kane et al., 2011). The sensors take a spot
measurement of temperature on each hour point. They were calibrated by the manufacturer
and found to be accurate to ±0.4°C (Tempcon Instrumentation Ltd, 2010).
NatCen interviewers asked the occupants to place the sensors in the living room and main
bedroom; 94 households did not, however, want to take the sensors (Fig. 3). Guidance was
provided, which stated that they should be placed away from heat sources and not in direct
sunlight. At the end of the monitoring period, households were asked to return the sensors in
pre-paid envelopes, these arrived back between late March 2010 and August 2011. In all 619
sensors were returned from 321 households
8
(Fig. 3), which represents a household loss rate
of 33%.
Following the surveys, in the winter of 20011/12, face-to-face interviews were conducted
with 20 households, which shed a little more light on their heating practices and how these
relate to their thermal comfort perceptions and lifestyle. Subsequently, web-based imagery
was used to confirm the house type and to determine the external surface areas and the
prevalence of extensions, etc.
Weather measurements
6
The data points bear no direct relationship to the households surveyed but preserve the number and rough
location of those interviewed.
7
Aerial imagery was used to confirm these responses.
8
Some homes were inadvertently not offered sensors by the interviewer.
