Content uploaded by Catharine Ward Thompson
Author content
All content in this area was uploaded by Catharine Ward Thompson
Content may be subject to copyright.
Landscape and Urban Planning 105 (2012) 221–229
Contents lists available at SciVerse ScienceDirect
Landscape and Urban Planning
journal homepage: www.elsevier.com/locate/landurbplan
More green space is linked to less stress in deprived communities:
Evidence from salivary cortisol patterns
Catharine Ward Thompsona,∗, Jenny Roeb,1, Peter Aspinallb,2, Richard Mitchellc,3,
Angela Clowd,4, David Millere,5
aOPENspace Research Centre, University of Edinburgh, Lauriston Place, Edinburgh EH3 9DF, UK
bOPENspace, School of the Built Environment, Heriot-Watt University, Edinburgh EH14 4AS, UK
cCentre for Research on Environment, Society and Health, College of Medical, Veterinary and Life Sciences, University of Glasgow, 1 Lilybank Gardens, Glasgow G12 8RZ, UK
dDepartment of Psychology, University of Westminster, 309 Regent Street, London W1B 2UW, UK
eThe James Hutton Institute, Craigiebuckler, Aberdeen AB15 8QH, UK
article info
Article history:
Received 17 June 2011
Received in revised form 6 December 2011
Accepted 21 December 2011
Available online 20 January 2012
Keywords:
Green space
Stress
Salivary cortisol
Residential environment
Deprivation
abstract
Green space has been associated with a wide range of health benefits, including stress reduction, but
much pertinent evidence has relied on self-reported health indicators or experiments in artificially con-
trolled environmental conditions. Little research has been reported using ecologically valid objective
measures with participants in their everyday, residential settings. This paper describes the results of
an exploratory study (n= 25) to establish whether salivary cortisol can act as a biomarker for variation
in stress levels which may be associated with varying levels of exposure to green spaces, and whether
recruitment and adherence to the required, unsupervised, salivary cortisol sampling protocol within the
domestic setting could be achieved in a highly deprived urban population. Self-reported measures of
stress and general wellbeing were also captured, allowing exploration of relationships between cortisol,
wellbeing and exposure to green space close to home. Results indicate significant relationships between
self-reported stress (P< 0.01), diurnal patterns of cortisol secretion (P< 0.05), and quantity of green space
in the living environment. Regression analysis indicates percentage of green space in the living envi-
ronment is a significant (P< 0.05) and independent predictor of the circadian cortisol cycle, in addition
to self-reported physical activity (P< 0.02). Results also show that compliance with the study protocol
was good. We conclude that salivary cortisol measurement offers considerable potential for exploring
relationships between wellbeing and green space and discuss how this ecologically valid methodology
can be developed to confirm and extend findings in deprived city areas to illuminate why provision of
green space close to home might enhance health.
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
There is an expanding body of research exploring the rela-
tionship between green space and health, from national level
epidemiological studies (de Vries, Verheij, Groenewegen, &
Spreeuwenberg, 2003; Maas, Verheij, Groenewegen, de Vries, &
Spreeuwenberg, 2006; Mitchell & Popham, 2007, 2008; van den
∗Corresponding author. Tel.: +44 131 221 6176/6177; fax: +44 131 221 6157.
E-mail addresses: c.ward-thompson@ed.ac.uk
(C. Ward Thompson), j.roe@hw.ac.uk (J. Roe), p.a.aspinall@hw.ac.uk (P. Aspinall),
Richard.Mitchell@glasgow.ac.uk (R. Mitchell), clowa@westminster.ac.uk (A. Clow),
david.miller@hutton.ac.uk (D. Miller).
1Tel.: +44 131 451 4629.
2Tel.: +44 131 221 6176.
3Tel.: +44 141 330 1663.
4Tel.: +44 207 911 5000x2174.
5Tel.: +44 1224 395000/395276.
Berg, Maas, Verheij, & Groenewegen, 2010) to very localised case
studies (Grahn, Ivarsson, Stigsdotter, & Bengtsson, 2010) and exper-
imental studies (Hartig, Evans, Jamner, Davies, & Gärling, 2003;
Hartig, Mang, & Evans, 1991; Park et al., 2007; Park, Tsunetsugu,
Kasetani, Kagawa, & Miyazaki, 2010; van den Berg & Custers, 2011;
van den Berg, Koole, & van der Wulp, 2003). There is evidence for
a positive relationship between access to green or natural envi-
ronments and people’s perceived overall general health (de Vries
et al., 2003; Maas et al., 2006), mental health (Grahn & Stigsdotter,
2003; Hartig et al., 2003; Maas, Verheij, et al., 2009; Ottosson &
Grahn, 2005), longevity (Takano, Nakamura, & Watanabe, 2002),
physical health (Coombes, Jones, & Hillsdon, 2010; Humpel, Owen,
& Leslie, 2002) and social health (de Vries, 2010; Kim & Kaplan,
2004; Kweon, Sullivan, & Wiley, 1998; Maas, van Dillen, Verheij, &
Groenewegen, 2009; Sullivan, Kuo, & Depooter, 2004). From epi-
demiological studies based in urban settings, these relationships
appear to be stronger among deprived populations (Mitchell &
Popham, 2008).
0169-2046/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.landurbplan.2011.12.015
222 C. Ward Thompson et al. / Landscape and Urban Planning 105 (2012) 221–229
The evidence is particularly strong for positive associations
between experience of natural environments and mental health.
It appears that contact with natural environments promotes
psychological restoration (Kaplan & Kaplan, 1989), improved
mood (Barton & Pretty, 2010; Hartig et al., 2003; Roe & Aspinall,
2011), improved attention (Hartig et al., 2003; Ottosson & Grahn,
2005) and reduced stress and anxiety (Grahn & Stigsdotter,
2003; Hartig et al., 2003; Maas, Verheij, et al., 2009; Ulrich et al.,
1991). Within deprived social housing communities in Chicago,
research has consistently shown the benefit of green space both
to cognitive restoration (Faber Taylor, Kuo, & Sullivan, 2002; Kuo,
2001), self-discipline (Faber Taylor et al., 2002), reduced aggres-
sion (Kuo & Sullivan, 2001a) and reduced crime (Kuo & Sullivan,
2001b).
1.1. Mechanisms by which natural environments might be
associated with stress reduction
Understanding the mechanisms by which natural environments
contribute to stress reduction or restoration is important if this con-
tribution is to be exploited for public health improvement. There
are three candidate behavioural mechanisms which may operate
synergistically, depending on the environment and contact type (de
Vries, 2010). Firstly, many people undertake some form of physical
activity as an inherent part of experiencing natural environments;
walking in a park for example. The positive effects on mood and
stress of physical activity are well established (Barton & Pretty,
2010; Penedo & Dahn, 2005). Secondly, people frequently have
the opportunity for some kind of social contact, however infor-
mal or unplanned, when they experience green space; they may
go with someone, or engage with others while there. Social con-
tact is also known to have positive effects on mood and stress level
(Heinrichs, Baumgartner, Kirschbaum, & Ehlert, 2003). Thirdly, peo-
ple often deliberately seek environments that they find attractive
for relaxing, to allow them to recover from demanding situations
and tasks, and natural environments are frequently sought for this
purpose (Grahn et al., 2010; Hartig, 2007, 2008; Kaplan, 1995;
Kaplan & Kaplan, 1989). In terms of psychological and physiolog-
ical mechanisms, there is evidence of independent responses that
are promoted by the perception of natural environments (Hartig
et al., 1991; Ulrich et al., 1991) and which may contribute to
people’s response to stress and their ability to cope with it (Lee
et al., 2011; McEwen & Stellar, 1993; Park et al., 2007, 2010). The
principal theoretical model for these responses is Ulrich’s psy-
choevolutionary model (Ulrich, 1983; Ulrich et al., 1991), which
proposes a direct impact of perceiving the natural environment on
an individual’s brain and body. This is thought to take place via psy-
choneuroendocrine mechanisms, including the functioning of the
hypothalamic pituitary adrenal (HPA) axis which regulates corti-
sol secretion and whose dysregulation is associated with a range of
disease outcomes (Li, Power, Kelly, Kirschbaum, & Hertzman, 2007;
Tsigos & Chrousos, 2002). If mechanisms such as this are in oper-
ation, we should therefore expect to observe a biological impact
of contact with natural environments. Experimental studies have
confirmed this. Being in or viewing green space has been shown
to reduce physiological measures of stress including blood pres-
sure (Hartig et al., 2003; Ulrich et al., 1991), heart rate (Ulrich et al.,
1991), skin conductance and muscle tension (Ulrich et al., 1991),
although some studies (e.g. Ottosson & Grahn, 2005) have failed to
find measurable physiological differences. In Japan, a study explor-
ing the effect of a green space intervention (Shinrin-yoku – taking
in the forest atmosphere) has shown that forest environments can
promote lower concentrations of cortisol, lower pulse rate, lower
blood pressure, greater parasympathetic nerve activity and lower
sympathetic nerve activity when compared to city environments
(Park et al., 2007, 2010).
1.2. Extending experimental evidence with observational
evidence
However, almost all of the work reported above has either
been carried out in artificially controlled experimental conditions,
whether in the laboratory or in the field, with certain, limited cat-
egories of participant (e.g. university students, hospital patients
with certain morbidities). While experiments that artificially con-
trol exposure to the environment may strengthen the case for a
causal relationship between stress reduction and some form of
visual or embodied access to green space, they are not easily gener-
alised to a population level. As far as we are aware, no research has
attempted to detect a biological impact of exposure to green and
natural environments encountered as part of ‘everyday life’, espe-
cially for deprived populations. The observational study reported
here was specifically designed to extend the evidence provided by
experimental studies, by seeking evidence for the impact of urban
green space on biomarkers of stress in the ‘real world’ of people’s
normal patterns of activity and experience, for those likely to be
experiencing multiple deprivation.
The use of cortisol as a biomarker of stress was selected as it
has been shown to be sensitive to activities in natural environ-
ments in the studies by Lee et al. (2011) and Park et al. (2007,
2010), taking male university students to forest and city environ-
ments away from their usual work or home contexts, and more
recently by van den Berg and Custers (2011), who found simi-
lar effects by assigning allotment gardeners to experimental tasks
before and after time spent gardening or indoor reading. However,
these studies involved controlled tasks and environmental con-
ditions, where salivary cortisol levels were measured before and
after exposure to different, specified environments. By contrast, our
study sought to examine cortisol patterns across the day for peo-
ple in routine contexts and patterns of behaviour in their everyday
environment. We chose diurnal patterns of cortisol as an outcome
measure since they offer an ecologically meaningful measure of
chronic stress: diurnal cortisol patterns reflect everyday circadian
rhythms of health and longer term effects of stressors in the social
and physical environment (Li et al., 2007), rather than responses pre
and post exposure to a particular environment at a single point in
time.
1.3. Patterns of cortisol secretion as a biomarker of chronic stress
Diurnal variation of salivary cortisol was selected as the
biomarker of stress in our study since it is a measure that reflects
everyday physiological functioning of the HPA axis and is sensitive
to a variety of responses to stress (Hsiao et al., 2010; Li et al., 2007).
In addition to its key role in responding to acute stressors, corti-
sol is a vital hormone for orchestrating healthy body functioning
around the 24 h circadian cycle (Nader, Chrousos, & Kino, 2010).
Disrupted patters of cortisol secretion are indicative of circadian
rhythm dysregulation which is associated with poor mental and
physical health (Nader et al., 2010; Wulff, Gatti, Wettstein, & Foster,
2010).
The core characteristic of healthy cortisol secretion is that levels
are carefully regulated by HPA axis to give very different concentra-
tions at different times of the day. The circadian cycle (with levels
changing from a daytime peak that may be as high as 20 nmol/l
shortly after awakening to a low of perhaps 1 nmol/l in the early
hours of sleep, see Edwards, Clow, Evans, & Hucklebridge, 2001)
signals to other body systems when it is night and day. Healthy diur-
nal patterns of cortisol secretion show such a daytime peak shortly
after awakening but a range of conditions are associated with a
flattening of the cortisol circadian rhythm, i.e. a less steep decline
in levels from morning until evening. In some cases, this flatter
slope is associated with increased overall cortisol secretion (a ‘high
C. Ward Thompson et al. / Landscape and Urban Planning 105 (2012) 221–229 223
flat’ slope), for example in normal ageing or in clinical depression
(Deuschle et al., 1997; Weber et al., 2000), whereas in other con-
ditions the flatter slope is associated with overall reduced levels of
cortisol secretion (a ‘low flat’ slope), for example in post-traumatic
stress disorder (PTSD), a combination of PTSD and long term neg-
ative life events, repressive anxiety and chronic fatigue (de Kloet
et al., 2007; Giese-Davis, Sephton, Abercrombie, Duran, & Spiegel,
2004; Jerjes, Cleare, Wessely, Wood, & Taylor, 2005; Li et al., 2007;
Witteveen et al., 2010).
The circadian cycle of cortisol secretion is thus sensitive to the
effects of chronic stress (Meerlo, Sgoifo, & Turek, 2002; Nader et al.,
2010), affording not only a biomarker of chronic stress but also a
mechanism by which stress and health are linked. Disrupted cor-
tisol cycles negatively impact upon a range of other body systems
with health related outcomes (Cacioppo et al., 2002; Kyrou & Tsigos,
2009; Sephton, Sapolsky, Kraemer, & Spiegel, 2000; Sherwood
Brown, Varghese, & McEwen, 2004; Strickland, Morriss, Wearden,
& Deakin, 1998). The circadian cycle, and its variation, thus com-
plicate the use of cortisol as a simple biological marker, but reveal
more about stress than simple cortisol concentrations.
1.4. Aims of this study
Our study asked whether, among residents of a deprived urban
area, the presence of different amounts of green space in the envi-
ronment around people’s homes was associated with:
a. stress as measured by levels and/or patterns of cortisol secretion
over the day; and/or
b. stress and more general wellbeing as measured by self-report
scales.
We also wanted to test:
c. the viability of unsupervised collection of saliva samples, within
the domestic setting of such a population, at the required times
post-awakening; and
d. whether physical activity levels moderate or confound the find-
ings on stress levels.
2. Study design and recruitment of sample
2.1. Study design
The study was located in Dundee, UK. Dundee had a population
of 153,226 in 2001 and contains a number of highly deprived neigh-
bourhoods with varying levels of green space. The design was to
gather data on the diurnal salivary cortisol secretion cycles of peo-
ple likely to face socio-economic adversity (by virtue of both their
individual and neighbourhood characteristics), and to determine if
variation in this cycle, and in overall cortisol levels, exhibited inde-
pendent association with the level of green space, as objectively
measured in their neighbourhoods of residence.
2.2. Choice of sample
In order to maximise exposure to the residential environment
(the literature suggests that those at home are more susceptible
to green space effects (de Vries et al., 2003)), and to target people
likely to face socio-economic adversity, we recruited participants
not in work for any reason (unemployed, on invalidity benefit, car-
ers, etc.). Since cortisol secretion is sensitive to age variations, our
target sample was men and women aged 35–55 years.
2.3. Method of recruitment
People not in work were recruited via centres in Dundee offering
training opportunities for unemployed people, as well as via local
community centres. This recruitment process was carried out city-
wide over a period of 4 weeks in January 2010. In addition to the
age criteria stated above, the following inclusion/exclusion criteria
were applied.
a. People on particular medications were excluded (e.g. steroids)
but more stringent exclusion criteria (e.g. use of anti-
depressants, smoking), although recorded, were waived owing
to the high likelihood of finding such behaviour patterns in the
target sample group.
b. People who had lived for less than 12 months in their neighbour-
hood of residence were excluded.
At the time of recruitment, participants were asked to complete
a short paper-based questionnaire (details below) and were briefed
on the protocol for taking cortisol samples. In particular, great care
was taken to explain to the participants the necessity to collect the
samples at the requested time post awakening, and to inform the
research team if they were unable to comply with the instructions.
3. Outcome measures
3.1. Stress measures
Our primary outcome measures related to salivary cortisol.
Saliva sampling was chosen over other body fluids (e.g. blood and
urine) as it has been shown to accurately represent the biologically
active component of circulating cortisol in the blood, allows for
repeated measurement, is non-invasive, can be self-administered
within the domestic setting and presents the participant no harm
(Kirschbaum & Hellhammer, 2000). Outcome measures were aver-
age daily levels and the diurnal decline, or slope, of cortisol
secretion (nmol/l) derived from multiple saliva sampling across
two consecutive days. The purpose of gathering samples over 2 days
was to test participant adherence to the protocol (i.e. there should
be no significant differences between Day 1 and Day 2 measure-
ments) and enable derivation of a more acute measure of cortisol
secretion and diurnal decline (slope) by using the average of 2 days’
data. This follows standard good practice for measuring circadian
rhythms in salivary cortisol (Edwards et al., 2001).
Our key secondary measure was a self-reported indicator of
stress, the Perceived Stress Scale (Cohen, Kamarck, & Mermelstein,
1983), comprising 10 items (e.g. feeling nervous and stressed; feel-
ing on top of things; being angered because of things outside your
control) measured on a 5-item response from ‘never’ to ‘very often’.
The final score assesses perceived stress over the preceding month
and can range from 0 (minimum level of stress) to 40 (maximum
level of stress).
3.2. Other individual measures
Mental wellbeing and physical activity were measured as
potential explanatory variables. Self-report mental wellbeing was
measured using a shortened version of the Warwick and Edinburgh
Mental Wellbeing Scale (SWEMWBS, Stewart-Brown et al., 2009).
SWEMWBS asks participants how they had felt over the last 4 weeks
in relation to 7 items measuring aspects of mental wellbeing (e.g.
feeling relaxed, feeling useful, etc.), with responses rated on a 5-
item scale from ‘none of the time’ to ‘all of the time’. Final scores
can range from 7 (low wellbeing) to 35 (high wellbeing). Physical
activity was measured using one item asking for the number of days
224 C. Ward Thompson et al. / Landscape and Urban Planning 105 (2012) 221–229
on which physical activity (of sufficient exertion to raise breathing
rate) reached or exceeded 30 min, recalled over the past 4 weeks,
based on 2008 recommendations from the British Heart Foundation
National Centre (described and tested for reliability and validity in
Milton, Bull, & Bauman, 2011).
As an individual measure of deprivation (based on one used
in the annual Scottish Social Attitudes Survey), participants were
asked for self-reported perceptions of coping on current income on
a scale of 1 (living comfortably on current income) to 4 (finding it
very difficult to live on current income). Participants’ age and sex
were also recorded.
3.3. Area-based measures
Socio-economic deprivation based on the Carstairs Index
from the most recent census data available (2001) (Carstairs
& Morris, 1991) were obtained via each participant’s postcode.
This is a widely used and well-validated indicator of area-level
socio-economic deprivation based on prevalence of household
overcrowding, unemployment among men, low social class, and
not having a car. The mean Carstairs score for Dundee is 3.93.
The percentage of a participant’s residential environment that
was green space was measured objectively using data available
at the Centre for Research on Environment Society and Health
(Mitchell, Astell-Burt, & Richardson, 2011; Richardson & Mitchell,
2010), again based on each participant’s postcode. These data are
freely available at the CRESH website (www.cresh.org.uk). The data
capture parks, woodlands, scrub and other natural environments,
but do not include private gardens. Residential environment was
defined by CAS Ward, a geographical unit used in the administra-
tion of the UK’s decennial census. Dundee contains 31 CAS Wards,
with a mean population of 4942 at the 2001 census, and a mean
green space value of 33.89%. Previous epidemiological research has
identified associations between the amount of green space in a
ward of residence, and the health of the residents (Richardson &
Mitchell, 2010).
4. Data collection and profile of participants
4.1. Ethics
This study was carried out in accordance with the British Psycho-
logical Society ‘Ethical Principles for Conducting the Research with
Human Participation’ and with the full ethical approval granted by
the lead researchers’ institutional ethics board.
4.2. Cortisol sampling
Average daily levels and the diurnal decline, or slope, of cortisol
secretion were derived from multiple saliva sampling across two
consecutive days. Data were gathered over the 4-week recruitment
period, with participants using Salivette saliva sampling devices
(Starstedt, Leicester, UK) on two consecutive weekdays. Partici-
pants were instructed to take 4 samples of saliva per day at a time
synchronised to their wake up time: 3, 6, 9 and 12 h after awaken-
ing. They were instructed not to smoke, eat or drink anything but
water 30 min before taking each sample and to keep a log of sam-
ple times. Wake up time was defined as the moment a participant
was first conscious of being awake. To maximise adherence, partici-
pants were sent individualised text prompts, based upon predicted
awakening times, 4 times on each day to remind them to take their
samples. At the end of the day, participants were asked to freeze
their samples (or place in a fridge) in a sealed bag provided. They
were then collected and shipped to the laboratory for assay analysis
within 5 days of collection.
Fig. 1. Distribution of the sample by percentage green space.
4.3. Profile of participants
The number who agreed to participate (n= 25) was 33% of those
approached for recruitment across a 4-week period, 12 males and
13 females. The mean age of the sample was 43.4 years (SD = 8.2)
within an age range of 33–57 years. 72% of the sample was unem-
ployed, with the remainder either in education or looking after
family; only 2 participants reported sickness or disability affecting
their ability to work. 61% of the sample reported finding it difficult
or very difficult to cope on current income levels.
The average Carstairs score for participants’ wards of residence
was 6.09, indicating high levels of deprivation. The percentage of
green space in the residential neighbourhood varied by postcode
and ranged from 14% to 74% (median 28.8%); to illustrate the range
across the sample, the quartiles by percentage green space are
shown in Fig. 1. The figure indicates a greater number of partic-
ipants living in areas of low and high percentage green space, as
compared to medium levels of green space.
The participant numbers for the statistical analyses vary
between 20 and 25 owing to occasional missing data due to non-
compliance in cortisol protocol (2 participants) or insufficient saliva
in some samples (3 participants).
4.4. Stress measures
Cortisol samples were assayed using Enzyme Linked Immuno-
Sorbent Assays (Salimetrics, USA). The mean cortisol level was
calculated from each of the daily measurement times across days
1 and 2; the slope measurement was the difference in mean mea-
sures at 3 h post awakening (sample 1) and 12 h post awakening
(sample 4) across Days 1 and 2. The cortisol mean (nmol/l) for the
sample (n= 20) was 4.58 (SD = 1.67), with the slope measurement
mean at 4.64 (SD =3.64). The self-reported stress (PSS) mean score
(n= 25) was 19.8 (SD = 6.36).
In order to test adherence to the cortisol sampling protocol over
time, we ran a repeated measures ANOVA (analysis of variance)
(n= 20), factoring day (day 1 and day 2), and sample (i.e. the cortisol
mean at 3, 6, 9 and 12 h post-awakening). We found no signifi-
cant main effect between Day 1 and day 2 for cortisol means. We
found a highly significant main effect of time over the day, from
C. Ward Thompson et al. / Landscape and Urban Planning 105 (2012) 221–229 225
3 to 12 h post-awakening (F= 8.77, df =3, P=0.001) indicating, as
expected, that cortisol means varied across the day. Both results
suggest adherence to the protocol and legitimised our strategy of
averaging both cortisol measures (levels and slope) across the 2
sampling days to give the most reliable measures.
4.5. Other health measures
The mean wellbeing (SWEMWBS) score in the sample was 22.72
(SD = 4.83, n= 25). Average levels for physical activity (n= 23) were
below the national recommendations of a minimum of 30 min of
moderate to vigorous activity on at least 5 days per week (Scottish
Executive, 2003), with the majority of the sample exercising on
fewer than 10 days in a 4-week period (mean = 8.65 days, SD = 8.29).
5. Results
5.1. Associations between variables
Relationships between variables were tested using Pearsons
bivariate correlations. Associations between cortisol measures,
self-reported stress, wellbeing and physical activity measures, and
percentage green space in each participant’s residential area, are
shown in Table 1. Higher mean levels of cortisol were associated
with a steeper cortisol slope and with greater wellbeing. Mean
cortisol was not statistically associated with physical activity, self-
reported stress or percentage of green space in the participant’s
home environment (although there were trends suggesting a pos-
itive association with physical activity and a negative association
with self-reported stress).
A steeper cortisol slope (the decline 3–12 h post awakening)
was positively associated with wellbeing, physical activity and per-
centage of green space, and negatively associated with levels of
stress.
There was a significant negative correlation between levels of
self-reported stress (PSS) and wellbeing and between PSS and the
amount of green space in the participant’s ward of residence.
5.2. Interpreting patterns of cortisol levels
To interpret the pattern of cortisol levels in the data, it is impor-
tant to note that a healthy diurnal pattern of cortisol secretion
shows a daytime peak shortly after awakening followed by a steep
decline as the day progresses (as described earlier).
Within our sample, some participants presented with an aver-
age cortisol level and a diurnal slope within the normal range.
However, a proportion had lower and flatter cycles than expected
from comparable studies of residents in non-deprived neighbour-
hoods (see Thorn, Evans, Cannon, Hucklebridge, & Clow, 2011 for
comparison). To illustrate this, we split the sample on the cortisol
slope median (3.92 nmol/l) to give two different groups based on
this measure (n= 20, 10 men and 10 women) (see Fig. 2). This shows
how, in our sample, a higher slope measurement (i.e. a greater
change in levels of cortisol between measurements 3 and 12 h post
Fig. 2. Differences in cortisol slope between sub-groups (n= 20).
awakening) was correlated with a higher mean cortisol measure-
ment overall, while the group with a flatter slope (n= 10, 4 men,
6 women) had lower mean cortisol levels, marked by lower levels
earlier in the diurnal cycle. This latter group’s cortisol pattern shows
a ‘low flat’ slope indicative of exhaustion and dysregulation of the
cortisol secretion system, as described in Section 1.3 (de Kloet et al.,
2007; Giese-Davis et al., 2004; Jerjes et al., 2005; Li et al., 2007;
Witteveen et al., 2010). Using a repeated measures ANOVA, the
between group difference is significant (F= 4.993, df = 17, P= 0.039).
Women had flatter slope profiles than men but the difference was
not statistically significant.
5.3. Predicting cortisol slope
To test the association between percentage green space and
cortisol slope for potential confounding variables, a linear regres-
sion was run to predict cortisol slope, with age, gender, degree
of coping on income level, deprivation (Carstairs score), physical
activity level and percentage green space in a participant’s residen-
tial area as the independent variables. The cortisol slope variable
met Kolmogorov–Smirnov criteria for normality of distribution.
Firstly, the independent variables were entered singly into the
model, using the ‘enter’ method. Of these, only two variables were
significant: percentage green space and physical activity. A second
regression was then run entering these two variables (again using
the ‘enter’ method). The model showed that both variables are sig-
nificant predictors of cortisol slope (see Table 2), which equates
to better wellbeing. For an increase of one day in physical activ-
ity (over 4 weeks) the cortisol slope will increase by 0.2. Green
space exposure is also associated with a steeper cortisol slope –
for every 1% increase in green space, the slope will increase by
Table 1
Relationships between cortisol patterns, health measures, and green space.
Variable Cortisol mean Cortisol slope Stress Wellbeing Physical activity % Green space
Cortisol mean (n= 20) 1
Cortisol slope (n= 20) 0.606** 1
Stress (n= 25) −0.440 −0.561*1
Wellbeing (n= 25) 0.474*0.478*−0.823** 1
Physical activity (n= 23) 0.452 0.576** −0.375 0.426*1
% Green space (n= 25) 0.195 0.489*−0.525** 0.370 0.186 1
*Significant at the P< 0.05 level (two-tailed).
** Significant at the P< 0.01 level (two-tailed).
226 C. Ward Thompson et al. / Landscape and Urban Planning 105 (2012) 221–229
Table 2
Linear regression model predicting cortisol slope (n= 20) over 2 days.
Model Standardised coefficients Sig. 95.0% Confidence interval for B
Beta Lower bound Upper bound
(Constant) 0.930 −2.890 3.142
Physical activity 0.502 0.014 0.050 0.391
% Green space 0.396 0.046 0.001 0.144
Note: the model (SPSS 18 output for coefficients) was significant at P= 0.005 with an R2value of 0.483.
Table 3
Linear regression model predicting self-reported stress (PSS) (n= 25).
Model Standardised coefficients Sig. 95.0% Confidence interval for B
Beta Lower bound Upper bound
(Constant) 0.000 16.544 30.638
% Green space −0.431 0.051 −0.278 0.001
Note: the model (SPSS 18 output for coefficients) was significant at P= 0.022 with an R2value of 0.304.
0.07. The VIF (Variance Inflation Factor) values (at 1.04) indicated
multi-collinearity was not a problem.
5.4. Predicting self-reported stress
A second regression was run in the same way, but using self-
reported stress (PSS) as the dependent variable. PSS data met
Kolmogorov–Smirnov criteria for normality of distribution. The
same 5 independent variables were entered singularly into the
model using the ‘enter’ method. Of these, percentage green space
and gender (reference category male) were significant. When both
variables were entered into the model together, percentage green
space reached marginal significance (at P= 0.051); gender was no
longer significant (see Table 3). Self-reported stress and percentage
green space were negatively associated, so for every 1% increase in
green space, self-reported stress decreased by 0.14 units.
6. Discussion
The main purpose of this study was to test whether the pres-
ence of different amounts of green space in the home environment
was associated with stress as measured objectively by levels and/or
patterns of cortisol secretion over the day, or subjectively by self-
reported measures of stress and general wellbeing. We tested these
questions, and the viability of the methods used, in an exploratory
study in a city in Scotland experiencing higher than national aver-
age levels of deprivation.
Firstly, in relation to percentages of green space in the residen-
tial neighbourhood, we found a relationship with objective markers
of stress as measured by levels and patterns of cortisol secretion.
While the methods used are unable to show a causal relationship,
we found a significant positive correlation between the diurnal
decline in cortisol across the day and percentage of green space
(the greater the slope, the greater the percentage green space), and
this relationship held after adjusting for demographic and socio-
economic variables in a multivariate regression model. The flat
cortisol slope associated with lack of green space in this study is
indicative of dysregulation of HPA axis function which has been
associated with an array of negative health outcomes (Nader et al.,
2010). This is an important finding. It represents the ‘missing link’
between experimental studies which have demonstrated benefi-
cial effects of contact with natural environments on biomarkers
of stress (Hartig et al., 2003; Park et al., 2007, 2010; Ulrich et al.,
1991) but which cannot be easily generalised to a population level,
and population level studies which have demonstrated associations
between green space and health outcomes including cardiovas-
cular mortality rates (Mitchell & Popham, 2008) and psychiatric
morbidity (Maas, Verheij, et al., 2009), but which have lacked infor-
mation on the mechanisms by which the associations are produced.
While further work would be needed to establish the degree to
which variation in green space use and the quality of green space
influence this relationship, the study provides evidence for a salu-
togenic environment–body interaction which is difficult to test for
at a population level but which may lie behind contact with nature
in people’s everyday living environments.
Secondly, we asked if green space has a relationship with stress
and wellbeing as measured by self-report scales. We found an
inverse relationship between percentage of neighbourhood green
space and self-reported stress (PSS) (marginally significant at
P= 0.051), showing that people’s perceived stress levels increased
as the amount of green space in their local environment decreased,
and this relationship again held after adjusting for demographic
and socio-economic variables.
Subsidiary research questions related to participants’ adherence
to the cortisol test protocol and the quality of the cortisol sample
data; and, finally, whether physical activity levels influence any
relationship between measures of stress and of green space lev-
els. Results indicate that our participants largely adhered to the
saliva sampling protocol, with missing data in only a few sam-
ples owing to insufficient saliva. Furthermore, the consistency of
the cortisol data across the 2 study days indicated good levels of
compliance to the strict timing requirements (3-hourly) for saliva
collection. This is a promising outcome, suggesting recruitment
could be effective in a larger sample of the same target popula-
tion. Contributory factors to adherence to the protocol, we believe,
were regular reminders to participants to take samples via mobile
telephone text prompts, and the personal briefing of participants
by researchers in relation to the project requirements.
We found mixed results on the relationship between physical
activity, cortisol measures and subjective stress levels. Regression
analyses showed that level of physical activity was a significant
predictor of cortisol slope but not of self-reported stress. Physical
activity has been shown to reduce stress (Hamer, Stamatakis, &
Steptoe, 2009; Tsatsoulis & Fountoulakis, 2006) but perhaps this
indirect mechanism is less sensitive to self-report than to objective
measures.
In our study, cortisol slope, as a measure of healthy patterns
of daily life, has been shown to be a useful measure in studying
the relationship between green space and stress: it is an effective
biomarker, sensitive to mental wellbeing and self-reported stress,
and showing a positive relationship with levels of green space in
participants’ residential wards.
Daily mean cortisol secretion was less strongly associated
with percentage green space. The strongly positive relationship
C. Ward Thompson et al. / Landscape and Urban Planning 105 (2012) 221–229 227
between diurnal slope and mean cortisol showed that these rela-
tionships were in the opposite direction to that which might have
been predicted, i.e. the lower the mean cortisol levels, the lower the
wellbeing. Low and flat cortisol cycles such as found in this study
are associated with chronic conditions such as post traumatic stress
disorder (PTSD), long term negative life events, repressive anxiety
and chronic fatigue (de Kloet et al., 2007; Giese-Davis et al., 2004;
Jerjes et al., 2005; Witteveen et al., 2010). Our participants were
middle aged and had been residents in the study areas (urban dis-
advantaged neighbourhoods) for many years. Although we did not
measure fatigue, negative life events and trauma in this study, it is
plausible that these participants had been exposed to lifetime social
disadvantage and negative life events which have resulted in HPA
axis dysregulation and the cortisol profiles presented here. Since
our sample were not in work, they might be expected to have had
greater sickness levels than the general population of the same age
living in these locations. Yet most responses to our general health
question showed good levels of health, with only two participants
reporting sickness or disability. Most interesting, for the purpose of
this study, is that those residing within areas of greater percentage
green space appear to have been more resilient to the negative
effects of urban deprivation and the stress-related consequences.
With regard to self-reported measures of stress and wellbeing,
we found several interesting patterns in the data. Firstly, corre-
lations with percentage green space were stronger for the stress
scale (PSS) than for the wellbeing scale (SWEMWBS). Secondly,
percentage green space showed no relationships with physical
activity levels, despite the fact that physical activity was signif-
icantly associated with stress as measured by cortisol slope, but
not by the self-reported measure (PSS). This suggests that green
space is related to subjective stress through mechanisms other than
physical exercise.
7. Limitations
We recognised that January in Scotland was not an ideal time of
year to explore the impact of green space on stress, given the cold
weather and short day length at this time, but the requirement
to undertake and analyse an early survey prior to more extended
project work later in 2010 necessitated this. It may be that asso-
ciations between local green space and wellbeing are stronger in
warmer months with longer daylight hours. However, our results
cannot be attributed to seasonal factors as all participants were
studied within the same 4 weeks, living in the same locality.
There is a limitation in our sample size; it reflects the exploratory
nature of the research to establish whether recruitment and adher-
ence to the study protocol was possible in a deprived population
and a domestic setting. Nonetheless, there was sufficient power in
the sample to demonstrate significant findings in relation to our
outcome measures. Our sample was not sufficient to allow sepa-
rate analysis by gender but the literature suggests such differences
may be likely: future studies should allow for examination of gen-
der differences in the outcome measures. Our sample is based on a
33% response rate from those approached for recruitment; we are
not able to quantify the impact of selection bias in this sample. The
low levels of poor health reported by the sample suggest they might
be healthier than the typical ‘not working’ population, but the low
and flat cortisol cycles observed suggest that the sample did include
those with significant stress problems. In addition, although we
were following recommended practice in cortisol sampling, the 2
days of the sample for each participant may have been atypical of
their everyday lives.
Owing to limitations in UK land use classification systems, our
objective measure of percentage green space was unable to pick
up the finer grain detail of front gardens and street trees. Equally,
we have not included measures of green views from participants’
homes, the quality of green space or their levels of use, which might
contribute explanatory pathways for the findings. Such details sug-
gest important directions for further research.
Perhaps most importantly, our study was cross-sectional and
cannot demonstrate causality. Although it builds on experimen-
tal studies which have, under controlled conditions, proved that
effects similar to those we observed in this study are produced by
contact with green space, residual confounding factors remain a
possibility. It might be that there is some other property of the
greener neighbourhood, or the lives and lifestyles of their resi-
dents, which has produced the impact on cortisol slope although,
in this study, the relationship with green space was independent
of demographic and other SES factors measured.
8. Conclusion
Our study offers new insights into links between green space
and health through an approach that measures the diurnal pat-
tern of salivary cortisol as an objective indicator of stress. This
approach allows us to assess the association between stress and
exposure to green space in a normal, everyday setting, meaningful
in socio-ecological terms, rather than in the artificially controlled
and time-limited exposures of an experiment. Consequently, our
approach offers perhaps greater external validity than previous lab-
oratory and field experiments. We have also shown that reliable
salivary cortisol data can be collected from populations living in
deprived neighbourhoods, across the day, unsupervised and within
the domestic, rather than laboratory, setting.
Using an objective measure of green space has demonstrated
associations with health that are independent of quality of green
space or perceptions of such space by participants. Our study
supports previous epidemiological research in using quantitative
estimates of green space but has added to understanding of the
underlying mechanism by also using an objective measure of stress.
Our findings illustrate one explanatory mechanism behind pos-
itive relationships between living in greener environments and
health: the regulation of the HPA axis as indicated by diurnal corti-
sol patterns, showing that greener environments may offer better
opportunities for moderating or coping with stress. This supports
previous experimental evidence that natural environments might
be associated with stress reduction (Grahn & Stigsdotter, 2003;
Hartig et al., 2003; Maas, Verheij, et al., 2009; Ulrich et al., 1991) and
that such links are particularly relevant to deprived communities
(Faber Taylor et al., 2002; Kuo, 2001; Kuo & Sullivan, 2001a). Our
findings also suggest that this association is not the result of phys-
ical activity per se and point to the likelihood that regular visits to
and/or views of green space lie behind the association.
Barton and Pretty’s (2010) study of exposure to green space and
mental health showed the strongest positive effects on mood and
self-esteem for the shortest duration (5 min) of activity in green
space, irrespective of activity intensity. It may be that a greener liv-
ing environment offers more frequent opportunities to experience
such benefits as people go about their everyday lives. Such findings
suggest that the association between high green space levels and
lower stress found in our study may be the result of many minor
but nonetheless significant episodes of contact with the natural
environment. The need for adequate levels of nearby green space
is an important message for landscape and urban planners when
designing new residential development, renovating existing urban
infrastructure or consulting on land use priorities.
Acknowledgements
The work described in this paper is part of the Green-
Health project funded by the Scottish Government’s Rural and
228 C. Ward Thompson et al. / Landscape and Urban Planning 105 (2012) 221–229
Environment Science and Analytical Services (RESAS) Division and
led by The James Hutton Institute in collaboration with the Univer-
sity of Edinburgh and the University of Glasgow.
References
Barton, J., & Pretty, J. (2010). What is the best dose of nature and green exercise
for improving mental health? A multi-study analysis. Environmental Science and
Technology,44(10), 3947–3955.
Cacioppo, J. T., Kiecolt-Glaser, J. K., Malarkey, W. B., Laskowski, B. F., Rozlog, L. A.,
Poehlmann, K. M., et al. (2002). Autonomic and glucocorticoid associations with
the steady-state expression of latent Epstein–Barr virus. Hormones and Behavior,
42, 32–41.
Carstairs, V., & Morris, R. (1991). Deprivation and Health in Scotland. Aberdeen:
Aberdeen University Press.
Cohen, S., Kamarck, T., & Mermelstein, R. (1983). A global measure of perceived
stress. Journal of Health and Social Behavior,24, 385–396.
Coombes, E., Jones, A. P., & Hillsdon, M. (2010). The relationship of physical activity
and overweight to objectively measured green space accessibility and use. Social
Science & Medicine,70(6), 816–822.
de Kloet, C. S., Vermettena, E., Heijnen, C. J., Geuze, E., Lentjes, E. G. W. M., &
Westenberg, H. G. M. (2007). Enhanced cortisol suppression in response to
dexamethasone administration in traumatized veterans with and without post-
traumatic stress disorder. Psychoneuroendocrinology,32, 215–226.
Deuschle, M., Gotthardt, U., Schweiger, U., Weber, B., Korner, A.,
Schmider, J., et al. (1997). With aging in humans the activity of the
hypothalamus–pituitary–adrenal system increases and its diurnal amplitude
flattens. Life Sciences,61, 2239–2246.
de Vries, S. (2010). Nearby nature and human health: Looking at the mechanisms and
their implications. In C. Ward Thompson, P. Aspinall, & S. Bell (Eds.), Innovative
approaches in researching landscape and health, Open space: People space 2 (pp.
77–96). Abingdon, UK: Routledge.
de Vries, S., Verheij, R. A., Groenewegen, P. P., & Spreeuwenberg, P. (2003). Nat-
ural environments – healthy environments? An exploratory analysis of the
relationship between greenspace and health. Environment and Planning A,35,
1717–1731.
Edwards, S., Clow, A., Evans, P., & Hucklebridge, F. (2001). Exploration of the awak-
ening cortisol response in relation to diurnal cortisol secretory activity. Life
Sciences,68, 2093–2103.
Faber Taylor, A., Kuo, F. E., & Sullivan, W. (2002). Views of nature and self discipline:
Evidence from inner city children. Journal of Environmental Psychology,22, 49–63.
Giese-Davis, J., Sephton, S. E., Abercrombie, H. C., Duran, R. E. F., & Spiegel, D.
(2004). Repression and high anxiety are associated with aberrant diurnal corti-
sol rhythms in women with metastatic breast cancer. Health Psychology,23(6),
645–650.
Grahn, P., Ivarsson, C. T., Stigsdotter, I. K., & Bengtsson, I. (2010). Using affordances
as health promoting tool in a therapeutic gardenin. In C. Ward Thompson, P.
Aspinall, & S. Bell (Eds.), Innovative approaches in researching landscape and
health, Open space: People space 2 (pp. 120–159). Abingdon, UK: Routledge.
Grahn, P., & Stigsdotter, U. A. (2003). Landscape planning and stress. Urban Forestry
& Urban Greening,2(1), 1–18.
Hamer, M., Stamatakis, E., & Steptoe, A. (2009). Dose–response relationship between
physical activity and mental health: The Scottish Health Survey. British Journal
of Sports Medicine,43, 1111–1114.
Hartig, T. (2007). Three steps to understanding restorative environments as health
resources. In C. Ward Thompson, & P. Travlou (Eds.), Open space: People space
(pp. 163–179). Abingdon, UK: Taylor and Francis.
Hartig, T. (2008). Green space, psychological restoration, and health inequality.
Lancet,372(9650), 1614–1615.
Hartig, T., Evans, G. W., Jamner, L. D., Davies, D. S., & Gärling, T. (2003). Tracking
restoration in natural and urban field settings. Journal of Environmental Psychol-
ogy,23, 109–123.
Hartig, T., Mang, M., & Evans, G. W. (1991). Restorative effects of natural environment
experience. Environment and Behavior,23, 3–26.
Heinrichs, M., Baumgartner, T., Kirschbaum, C., & Ehlert, U. (2003). Social support and
oxytocin interact to suppress cortisol and subjective responses to psychosocial
stress. Biological Psychiatry,54(12), 1389–1398.
Hsiao, F. H., Yang, T. T., Ho, R. T., Jow, G. M., Ng, S. M., Chan, C. L., et al. (2010).
The self-perceived symptom distress and health-related conditions associated
with morning to evening diurnal cortisol patterns in outpatients with major
depressive disorder. Psychoneuroendocrinology,35(4), 503–515.
Humpel, N., Owen, N., & Leslie, E. (2002). Environmental factors associated with
adults’ participation in physical activity: A review. American Journal of Preventive
Medicine,22, 188–199.
Jerjes, W. K., Cleare, A. J., Wessely, S., Wood, P. J., & Taylor, N. F. (2005). Diurnal
patterns of salivary cortisol and cortisone output in chronic fatigue syndrome.
Journal of Affective Disorders,87, 299–304.
Kaplan, S. (1995). The restorative benefits of nature: Toward an integrative frame-
work. Journal of Environmental Psychology,15, 169–182.
Kaplan, R., & Kaplan, S. (1989). The experience of nature: A psychological perspective.
New York: Cambridge University Press.
Kim, J., & Kaplan, R. (2004). Physical and psychological factors in sense of community
– New urbanist Kentlands and nearby orchard village. Environment Behavior,36,
313–334.
Kirschbaum, C., & Hellhammer, D. K. (2000). Salivary cortisol. In G. Fink (Ed.), Ency-
clopaedia of stress (pp. 379–383). Oxford: Academic Press.
Kuo, F. E. (2001). Coping with poverty: Impacts of environment and attention in the
inner city. Environment and Behavior,33(1), 5–34.
Kuo, F. E., & Sullivan, W. C. (2001a). Aggression and violence in the inner city:
Effects of environment on mental fatigue. Environment and Behavior,33(4),
543–571.
Kuo, F. E., & Sullivan, W. C. (2001b). Environment and crime in the inner city:
Effects of environment via mental fatigue. Environment and Behavior,33(3),
343–367.
Kweon, B. S., Sullivan, W. C., & Wiley, A. (1998). Green common spaces and the social
integration of inner-city older adults. Environment and Behavior,30, 832–858.
Kyrou, I., & Tsigos, C. (2009). Stress hormones: Physiological stress and regulation
of metabolism. Current Opinion in Pharmacology,9, 787–793.
Lee, J., Park, B.-J., Tsunetsugu, Y., Ohira, T., Kagawa, T., & Miyazaki, Y. (2011). Effect of
forest bathing on physiological and psychological responses in young Japanese
male subjects. Public Health,125, 93–100.
Li, L., Power, C., Kelly, S., Kirschbaum, C., & Hertzman, C. (2007). Life-time socio-
economic position and cortisol patterns in mid-life. Psychoneuroendocrinology,
32, 824–833.
Maas, J., van Dillen, S. M. E., Verheij, R. A., & Groenewegen, P. P. (2009). Social contacts
as a possible mechanism behind the relation between green space and health.
Health and Place,15(2), 586–595.
Maas, J., Verheij, R. A., de Vries, S., Spreeuwenberg, P., Groenewegen, P. P., & Schelle-
vis, G. S. (2009). Morbidity is related to a green living environment. Journal of
Epidemiology and Community Health,63, 967–973.
Maas, J., Verheij, R. A., Groenewegen, P. P., de Vries, S., & Spreeuwenberg, P. (2006).
Green space, urbanity and health: How strong is the relation? Journal of Epidemi-
ology and Community Health,60, 587–592.
McEwen, B. S., & Stellar, E. (1993). Stress and the individual: Mechanisms leading to
disease. Archives of Internal Medicine,153, 2093–2101.
Meerlo, P., Sgoifo, A., & Turek, F. W. (2002). The effects of social defeat and other
stressors on the expression of circadian rhythms. Stress,5(1), 15–22.
Milton, K., Bull, F. C., & Bauman, A. (2011). Reliability and validity testing of a single-
item physical activity measure. British Journal of Sports Medicine,45, 203–208.
Mitchell, R., Astell-Burt, T., & Richardson, R. (2011). A comparison of green space
indicators for epidemiological research. Journal of Epidemiology and Community
Health, doi:10.1136/jech.2010.119172
Mitchell, R., & Popham, F. (2007). Greenspace, urbanity and health: Relationships in
England. Journal of Epidemiology and Community Health,61, 681–683.
Mitchell, R., & Popham, F. (2008). Effect of exposure to natural environment on
health inequalities: An observational population study. The Lancet,372(9650),
1655–1660.
Nader, N., Chrousos, G. P., & Kino, T. (2010). Interactions of the circadian
CLOCK system and the HPA axis. Trends in Endocrinology and Metabolism,21,
277–286.
Ottosson, J., & Grahn, P. (2005). A comparison of leisure time spent in a garden with
leisure time spent indoors: On measures of restoration in residents in geriatric
care. Landscape Research,30(1), 23–55.
Park, B. J., Tsunetsugu, Y., Kasetani, T., Hirano, H., Kagawa, T., Sato, M., et al. (2007).
Physiological effects of Shinrin-yoku (taking in the atmosphere of the forest) –
Using salivary cortisol and cerebral activity as indicators. Journal of Physiological
Anthropology,26(2), 123–128.
Park, B. J., Tsunetsugu, Y., Kasetani, T., Kagawa, T., & Miyazaki, Y. (2010). The phys-
iological effects of Shinrin-yoku (taking in the forest atmosphere or forest
bathing): Evidence from field experiments in 24 forests across Japan. Environ-
mental Health and Preventive Medicine,15, 18–26.
Penedo, F. J., & Dahn, J. R. (2005). Exercise and well-being: A review of mental and
physical health benefits associated with physical activity. Current Opinion in
Psychiatry,18(2), 189–193.
Richardson, E. A., & Mitchell, R. (2010). Gender differences in relationships between
urban green space and health in the United Kingdom. Social Science and Medicine,
71(3), 568–575.
Roe, J., & Aspinall, P. (2011). The restorative benefits of walking in urban and rural
settings in adults with good and poor mental health. Health & Place,17, 103–113.
Scottish Executive. (2003). Let’s make Scotland more active: A strategy for physical
activity. Edinburgh: Physical Activity Task Force.
Sephton, S. E., Sapolsky, R. M., Kraemer, H. C., & Spiegel, D. (2000). Diurnal corti-
sol rhythm as a predictor of breast cancer survival. Journal of National Cancer
Institute,92, 994–1000.
Sherwood Brown, E., Varghese, F. P., & McEwen, B. S. (2004). Association of depres-
sion with medical illness: Does cortisol play a role? Biological Psychiatry,55,
1–9.
Stewart-Brown, S., Tennant, A., Tennant, R., Platt, S., Parkinson, J., & Weich, S. (2009).
Internal construct validity of the Warwick-Edinburgh Mental Well-being Scale
(WEMWBS): A Rasch analysis using data from the Scottish Health Education
Population Survey. Health and Quality of Life Outcomes,7, 15. doi:10.1186/1477-
7525-7-15
Strickland, P., Morriss, R., Wearden, A., & Deakin, B. (1998). A comparison of salivary
cortisol in chronic fatigue syndrome, community depression and health controls.
Journal of Affective Disorders,47, 191–194.
Sullivan, W. C., Kuo, F. E., & Depooter, S. F. (2004). The fruit of urban nature: Vital
neighbourhood spaces. Environment and Behavior,36, 678–700.
Takano, T., Nakamura, K., & Watanabe, M. (2002). Urban residential environments
and senior citizens’ longevity in megacity areas: The importance of walkable
green spaces. Journal of Epidemiological Community Health,56, 913–918.
C. Ward Thompson et al. / Landscape and Urban Planning 105 (2012) 221–229 229
Thorn, L., Evans, P., Cannon, A., Hucklebridge, F., & Clow, A. (2011). Diurnal cortisol
secretion: Seasonal differences in healthy and SAD participants. Psychoneuroen-
docrinology,36, 816–823.
Tsigos, C., & Chrousos, G. P. (2002). Hypothalamic–pituitary–adrenal axis, neuroen-
docrine factors and stress. Journal of Psychosomatic Research,53, 865–871.
Tsatsoulis, A., & Fountoulakis, S. (2006). The protective role of exercise on stress
system dysregulation and comorbidities. Stress, Obesity, and Metabolic Syndrome,
1083, 196–213.
Ulrich, R. S. (1983). Aesthetic and affective responses to natural environment. In
I. Altman, & J. F. Wohlwill (Eds.), Behavior and the natural environment, human
behavior and environment, advances in theory and research (pp. 85–125). New
York: Plenum.
Ulrich, R. S., Simons, R. F., Losito, B. D., Fiorito, E., Miles, M. A., & Zelson, M. (1991).
Stress recovery during exposure to natural and urban environments. Journal of
Environmental Psychology,11, 201–230.
van den Berg, A., & Custers, M. H. G. (2011). Gardening promotes neuroendocrine
and affective restoration from stress. Journal of Health Psychology,16(1), 3–11.
van den Berg, A. E., Koole, S. L., & van der Wulp, N. Y. (2003). Environmental
preference and restoration: (How) are they related? Journal of Environmental
Psychology,23(2), 135–146.
van den Berg, A., Maas, J., Verheij, R. A., & Groenewegen, P. P. (2010). Green space
as a buffer between negative life events and health. Social Science & Medicine,
70(8), 1203–1210.
Weber, B., Lewicka, S., Deuschle, M., Colla, M., Vecsei, P., & Heuser,
I. (2000). Increased diurnal plasma concentrations of cortisone in
depressed patients. Journal of Clinical Endocrinology and Metabolism,85,
1133–1136.
Witteveen, A. B., Huizink, A. C., Slottje, P., Bramsen, I., Smid, T., & van der Ploeg,
H. M. (2010). Associations of cortisol with posttraumatic stress symptoms and
negative life events: A study of police officers and firefighters. Psychoneuroen-
docrinology,35, 1113–1118.
Wulff, K., Gatti, S., Wettstein, J. G., & Foster, R. G. (2010). Sleep and circadian rhythm
disruption in psychiatric and neurodegenerative disease. Nature Reviews Neuro-
science,11, 589–599.
