Article Text

Download PDFPDF

Sex- and age-specific seasonal variations in physical activity among adults
  1. G R McCormack1,
  2. C Friedenreich2,
  3. A Shiell1,
  4. B Giles-Corti3,
  5. P K Doyle-Baker4
  1. 1Population Health Intervention Research Centre, Department of Community Health Sciences, University of Calgary, Alberta, Canada
  2. 2Division of Population Health, Alberta Health Services, Alberta, Canada
  3. 3Centre for the Built Environment and Health, School of Population Health, University of Western Australia, Australia
  4. 4Faculty of Kinesiology, University of Calgary, Alberta, Canada
  1. Correspondence to Dr Gavin McCormack, Population Health Intervention Research Centre, University of Calgary, 3rd Floor TRW Building, 3280 Hospital Drive, NW Calgary, Alberta T2N 4Z6, Canada; gmccorma{at}


Background To examine seasonal variations in self-reported physical activity among an urban population of Calgarian adults.

Method Telephone surveys were conducted with two independent random cross-sectional samples of adults in summer and autumn 2007 (n=2199) and in winter and spring 2008 (n=2223). Participation and duration of walking for recreation (WR), walking for transportation (WT), moderate (MODPA) and vigorous physical activity (VIGPA) undertaken in a usual week were captured. Seasonal comparisons of participation related to these activities and sufficient MODPA (≥210 min/week) and VIGPA (≥90 min/week) physical activity were examined using logistic regression.

Results Compared with winter, participation in WR was significantly (p<0.05) more likely in summer (OR 1.42), autumn (OR 1.35) and spring (OR 1.40), WT was more likely in autumn (OR 1.27), and MODPA was more likely in summer (OR 1.42). Achievement of sufficient MODPA was significantly more likely in summer (OR 1.80), autumn (OR 1.31) and spring (OR 1.24). Although there was no seasonal variation in sufficient VIGPA overall, variations in seasonal pattern among sub-populations were observed. Sex- and age-specific seasonal patterns in physical activity were also found.

Conclusion Measuring physical activity throughout the year, rather than at one time point, would more accurately monitor physical activity and assist in developing seasonally appropriate physical activity interventions. Moreover, in countries that experience extreme weather conditions, creating physical activity-friendly environments that help overcome these conditions might contribute to year-long physical activity participation.

  • Seasonality
  • exercise
  • health
  • surveys
  • environment
  • health behaviour
  • seasonal variation

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Participation in physical activity at moderate to vigorous intensities can provide health-enhancing benefits.1 However, physical activity must be undertaken regularly to maintain its health-protective effects. Markers of cardiovascular health,2 3 functional capacity,4 muscular strength5 and body composition2 are compromised within only a few weeks of ceasing or reducing physical activity. Personal, social and physical environmental factors can interrupt participation in regular physical activity.6 From a public health perspective, advancing knowledge with regard to factors that may interrupt physical activity participation is necessary for intervention planning.

Seasonality can influence physical activity participation.7 Seasonality is the systematic periodic increase or decrease in the prevalence or incidence of a health outcome or behaviour that corresponds to seasons or other calendar periods.8 Weather conditions including precipitation, snow accumulation, temperature and wind, as well as hours of daylight, all vary according to season and affect physical activity participation.9 10 Physical activity levels are found to be lower in winter compared with spring, autumn and summer7 including among Canadian populations.11 Moreover, the types and intensities of physical activities undertaken can also vary according to season.12–15 However, few population studies have considered these relationships among specific sub-groups. The influence of seasonality and weather conditions on physical activity of older16 17 and younger adults15 have been examined in separate studies; however, the differential affect of age on this relationship has not been examined in a single study. Moreover, findings regarding the moderating affect of sex on the association between season and physical activity are mixed. Plasqui and Westerterp15 found a greater reduction in physical activity between summer and winter in men compared with women. Uitenbroek12 also found sex-specific patterns of physical activity throughout the year. Others have found similar seasonal patterns of physical activity regardless of sex.13 18

Reductions in physical activity are known to contribute to adverse acute and chronic health outcomes; thus, advancing knowledge regarding associations between seasonality and physical activity is a relevant public health concern. Understanding how seasonality influences participation in different types of physical activity among specific segments of the population would assist in developing appropriate physical activity interventions within the Canadian context. The aim of the study was to examine seasonal patterns in the prevalence of different types of physical activity including walking for recreation (WR), walking for transportation (WT), moderate-intensity (MODPA) and vigorous-intensity activity (VIGPA) among adults residing in Calgary. The relationship between seasonality and achieving sufficient levels of physical activity is also examined overall and in analyses stratified by sex and age.



Two independent random cross-sectional samples of Calgarian adults were recruited during telephone interviews between August and October 2007 (summer/early autumn) and January and April 2008 (winter/early spring) as part of the Economic Evaluation of using Urban Form to Increase Activity (EcoEUFORIA) project. The aim of EcoEUFORIA is to evaluate the cost-effectiveness of creating walkable neighbourhoods to support physical activity. Publicly listed telephone numbers were sampled, and the final digit of each number was replaced with a random digit to ensure coverage of unlisted numbers. Respondents provided informed verbal consent at the beginning of the telephone interview. Eligible participants were 18 years or older, and in the case of multiple members of one household being eligible, the individual with the most recent birthday was selected. Respondents provided their postal code, and those residing outside the city limits were also excluded. Interviews captured information related to physical activity participation and demographics. Response rates (RR=completed/completed plus refused) for interviews completed during summer/early autumn (n=2199; RR=33.6%) and winter/early spring (n=2223; RR=36.7%) were similar. The Conjoint Health Research Ethics Board at the University of Calgary granted ethics approval for this study.

Seasonal weather characteristics of Calgary, Alberta, Canada

Calgary is located east of the Rocky Mountains, is elevated over 1 km above sea level and has a continental climate. Table 1 presents Calgary weather conditions for the months during which telephone surveys were administered. The average monthly maximum temperature was highest during August (20°C) and September (17.1°C). Total rainfall was also highest during August (54.4 mm) and September (44 mm). January and February were the coldest months with average maximum temperatures of −1.7°C and 1.8°C, respectively. Except for August, some snowfall was recorded for all survey months with the highest total snowfall recorded during April (65.8 cm).19

Table 1

Descriptive statistics of weather conditions for months during which telephone surveys were completed

Data collected

Physical activity items

The frequency and duration of WT, WR, VIGPA (ie, activities that made the respondent breathe harder or puff and pant) and MODPA were collected using a modified version of the Neighbourhood Physical Activity Questionnaire.20 21 The items differentiate between physical activities undertaken within the neighbourhood (ie, a 15-min walk of home) and outside the neighbourhood and have acceptable reliability among adults.20 21

Demographic characteristics

Demographic information included: sex, age, highest level of education completed (elementary/junior high school; senior high school; college/technical school; university undergraduate or university postgraduate), home ownership (owning/buying; renting or other arrangement), number of children <18 years of age at home and country of birth (Canada or other country).

Statistical analysis

Frequencies for the demographic variables were calculated for the whole data collection period (2007 and 2008) and by season. Data collected during January (n=581) and February (n=534) were combined and coded as “winter”, August (n=578) and September (n=739) coded as “summer”, October (n=882) coded as “autumn” and March (n=620) and April (n=488) coded as “spring”. Due to the methodological constraints, data could not be collected during November, December, May, June and July. Pearson's χ2 was used to compare the distribution of demographic characteristics by season. The sample demographic profiles were also compared against recent national22 23 and civic census data.24

Definitions of sufficient physical activity presented in this paper conservatively reflect the Canadian Physical Activity Recommendations that encourage adults to participate in 30–60 min of light to moderate intensity physical activity or 20 min of vigorous-intensity physical activity on most days.25 26 Weekly minutes of WR, WT and MODPA were summed and dichotomised (ie, <210 vs ≥210 min/week) to reflect participation in at least 30 min of moderate-intensity physical activity on all days of the week. Weekly VIGPA minutes were dichotomised (ie, <90 vs ≥90 min/week) to reflect participation in at least 30 min of vigorous-intensity PA on at least 3 days/week. Respondents achieving either sufficient moderate (ie, ≥210 min/week) or sufficient vigorous (ie, ≥90 min/week) physical activity were coded as participating in sufficient total physical activity.

Seasonal comparisons of weekly prevalence of WR, WT, MODPA, VIGPA and sufficient vigorous, sufficient moderate and sufficient total physical activity were undertaken for all respondents and stratified by sex and age. Age was trichotomised to provide approximate representation of younger (18–39 years), middle-aged (40–59 years) and older adults (≥60 years). ORs and 95% CIs (95% CI) were estimated using binary logistic regression to examine the association between season and participation in each of the four activities and in the three sufficient physical activity variables for all respondents and stratified by age group and sex. For logistic regression models, winter was set as the reference group with which summer, autumn and spring physical activity levels were compared. All models were adjusted for demographic characteristics. Data analysis was conducted using SPSS V. 16.


The demographic profiles of the 2007 and 2008 samples were similar (table 2). Both samples consisted mostly of women, 40–59 years old, university-educated participants, home owners, those with no children <18 years of age at home and Canadian born. Demographic profiles for season in which data were collected were also similar; however, a statistically significant difference in the proportion of home owners was found (ie, higher in autumn and lower in spring). Compared to the Calgary population, the study sample included a higher proportion of women, Canadian born, home owners, younger adults, had higher education levels and were more likely to have children <18 years at home.

Table 2

Demographic profiles by year of data collection (2007 and 2008) and season (summer, autumn, winter and spring)

Complete demographic and physical activity data were derived for 4354 respondents (2007 and 2008 surveys combined). Two thirds (66.9%) of respondents participated in WT, while over three quarters (79.9%) participated in WR during a usual week. One half (50.6%) of all respondents participated in MODPA in a usual week, while 61.3% participated in VIGPA. During a usual week, 61.5% of respondents achieved sufficient moderate and 46.1% achieved sufficient vigorous physical activity. Three quarters (75.8%) of all respondents achieved either sufficient moderate or vigorous physical activity during a usual week.

Seasonal variations in recreational and transportation-related walking

Respondents were more likely to report WR during a usual week in summer (OR 1.42; 95% CI 1.10 to 1.83), autumn (OR 1.35; 95% CI 1.12 to 1.64) and spring (OR 1.40; 95% CI 1.14 to 1.73) than in winter. Among women and younger adults, the likelihood of WR was higher during summer, autumn and spring compared with winter (table 3). The likelihood of WR was also higher in spring compared with winter for older adults. Participation in WT was higher during autumn compared with winter for the pooled data (OR 1.27; 95% CI 1.08 to 1.50) and in stratified analysis for women and younger and middle-aged adults.

Table 3

ORs for the association between season and prevalence of physical activity stratified by sex and age group

Seasonal variations in participation in moderate and vigorous-intensity physical activity

Among all respondents, MODPA participation was more likely in summer versus winter (OR 1.27; 95% CI 1.03 to 1.56; table 3). Participation in MODPA was significantly more likely in summer than winter for younger adults when analysed separately. No statistically significant associations were found between season and participation in VIGPA.

Seasonal variations in the achievement of sufficient physical activity

Participants were significantly more likely to achieve at least 210 min of MODPA or 90 min of VIGPA in spring (OR 1.24; 95% CI 1.04 to 1.47), summer (OR 1.80; 95% CI 1.45 to 2.23) and autumn (OR 1.31; 95% CI 1.12 to 1.54) than in winter (table 4). Men and women were both more likely to achieve sufficient levels of physical activity from either moderate or vigorous activity in summer compared with winter, as were younger, middle-aged and older adults. However, men and older adults were also more likely to participate in sufficient total physical activity in autumn than in winter. The likelihood of participating in sufficient MODPA was significantly higher in summer (OR 1.76; 95% CI 1.37 to 2.27), autumn (OR 1.29; 95% CI 1.08 to 1.54) and spring (OR 1.23; 95% CI 1.02 to 1.50) than in winter (table 4). Men and women were more likely to achieve sufficient MODPA in summer and autumn. The likelihood of achieving MODPA was also more likely in spring compared with winter among women. Younger and middle-aged adults were similarly more likely to achieve sufficient MODPA in summer and autumn, while in those aged ≥60 years and over, participation in sufficient MODPA was only significantly more likely during summer but no other season.

Table 4

ORs for the association between season and participation in sufficient physical activity stratified by sex and age group

Inconsistent with findings for all other types of PA, women were significantly less likely to participate in sufficient VIGPA in autumn than in winter (OR 0.82; 95% CI 0.66 to 1.00; table 4). However, men were significantly more likely to participate in sufficient levels of VIGPA during summer than in winter (OR 1.45; 95% CI 1.04 to 2.03). Age-specific variations were also evident. Younger adults were less likely to participate in sufficient vigorous physical activity in autumn compared with winter (OR 0.74; 95% CI 0.57 to 0.96). However, in middle-aged adults, achievement of sufficient VIGPA was more likely in summer than winter (OR 1.40; 95% CI 1.01 to 1.93), and older adults were more likely to participate in sufficient VIGPA during both summer (OR 1.54; 95% CI 1.00 to 2.38) and autumn (OR 1.55; 95% CI 1.03 to 2.34).


Consistent with previous studies, our findings show that in winter, physical activity levels are lower than other seasons.7 11 13 14 27 However, we found that seasonality influences participation in some types of physical activities more than others. Moreover, irrespective of age and sex, the direction of associations between season and physical activity were the same, with only modest differences in the magnitude of these associations observed. The one exception was among women and younger adults, where participation in sufficient vigorous physical activity was less likely in autumn compared with winter.

Participation in irregular physical activity throughout the year due to seasonal changes may reduce the protective health benefits accrued from participation at other times of the year. Since overweight and obesity result from an imbalance between energy expenditure and consumption, it is likely that significant proportions of the population are at increased risk of becoming overweight or obese if they maintain their energy intake but reduce their physical activity in response to weather patterns. This risk is compounded particularly in colder climates given evidence of increased energy consumption during winter compared with summer.28 While seasonal fluctuations in physical activity and dietary intake are likely to be small, the resulting energy imbalance accumulated over several years could increase the risk of becoming overweight or obese.29 30

Pooled results showed significant seasonal differences in the achievement of sufficient moderate but not sufficient vigorous-intensity physical activity. Other studies in Scotland12 and North America18 found seasonal variation in light to moderate, but not vigorous, leisure-time physical activity. However, we found that Calgarian women and younger adults were less likely to achieve sufficient vigorous physical activity in autumn than in winter. Moreover, men and middle-aged adults only showed an increased likelihood of sufficient vigorous-intensity physical activity during summer, while older adults were more likely to achieve this level during autumn and spring. While seasonal weather conditions might explain the irregular patterns in sufficient vigorous-intensity physical activity found among different sex and age groups, the seasonality of sports (ie, winter vs summer sports) or vigorous activities and their popularity among certain populations could also be explanatory factors.13 14 31

Seasonal variations in recreational and transport-related walking were found in this study, reflecting similar findings elsewhere.32 33 These patterns may well be influenced by weather conditions, although it is possible that they reflect events associated with particular calendar periods (eg, school vacations). Nevertheless, it is worth exploring whether policy and built environment interventions could be used to encourage more physical activity, including walking, throughout the year. For example, in cold climate countries with significant snowfall, snow and ice accumulation on sidewalks increases the risk of falls and can make walking more challenging, particularly for those with mobility constraints (eg, physical disabilities, older adults, pushing of strollers).34 35 Thus, it might be important to enforce immediate snow and ice removal from sidewalks that may act as a barrier to walking. The provision of sufficient lighting along sidewalks and in parks to overcome fewer hours of daylight during the winter could also support walking and other physical activity behaviours. More specifically, higher participation in transport-related walking might be encouraged through the provision of weather-proof bus shelters (protecting from rain, snow, heat and wind) and sufficient lighting at bus stops and transit stations. Moreover, increasing the proximity of households and destinations (eg, convenience stores) by building more walkable neighbourhoods could decrease the length of walking trips and reduce exposure to certain weather conditions (eg, rain, snow, extreme heat). Conversely, in hot climates, it may be necessary to increase the presence of shade trees along streetscapes to reduce exposure to the sun.36 37 Reducing the effect of weather-related barriers might assist in maintaining regular year-long participation in both recreational and transport-related walking as well as other forms of physical activity.

Consistent with previous findings, our study confirms that longitudinal monitoring of physical activity should be undertaken at the same time each year to avoid seasonal bias. Measuring physical activity at multiple-time points throughout the year (ie, in each month or season) may produce more accurate prevalence estimates than annual single-time point monitoring. Seasonality should also be considered when designing and evaluating interventions targeting increases in physical activity. Encouraging individuals to participate in alternative physical activities during winter months (ie, winter-specific media campaigns, reducing the cost of indoor facilities) may ensure that regular sufficient physical activity is accrued throughout the year. Moreover, individuals participating in walking and other physical activities in outdoor settings during warmer months could also be encouraged to continue participation in indoor settings during colder months (ie, mall walking, indoor sports).38 39 Nevertheless, more research examining whether or not seasonality influences adherence to physical activity programs and interventions is needed.

Several limitations should be considered when interpreting the findings of this study. A recent survey for Calgary suggests 63.2% of adults accrue sufficient physical activity.40 A higher than expected prevalence of sufficient physical activity in the current study could have resulted from our measure of “usual” physical activity that captured behaviour inside and outside the neighbourhood separately. Similar to other self-report measures, our data are also subject to reporting error and bias. Second, our study did not consider weather conditions that can fluctuate within seasons nor, due to methodological constraints, included data collected for all months of the year. Third, seasonal physical activity patterns observed at the population level from these cross-sectional data may not be observed to the same extent among cohorts of individuals. Furthermore, causality cannot be inferred from cross-sectional studies. Finally, while our results are reflective of findings in Canada11 and the Northern United States,14 17 27 they are likely most generalisable to regions that have similar demographic and socio-cultural profiles and climatic conditions.

The findings of differential influence of season by types of physical activity and population sub-groups (ie, age and sex) are unique contributions to the literature. Moreover, the finding that achievement of sufficient physical activity fluctuates throughout the year and that this may help explain increases in obesity and overweight if compensations in diet intake are not considered is another unique contribution. Taking into consideration seasonal variations of physical activity may allow for more accurate monitoring of physical activity and assist in developing seasonally appropriate physical activity interventions. Moreover, creating physical-activity-friendly environments that incorporate attributes to overcome constraints resulting from weather conditions could also contribute to regular year-long physical activity participation.

What is already known on this subject

  • Physical activity patterns vary according to season.

  • Little is known about the seasonal influences on different types of physical activity for specific populations.

  • Little is known about how levels of sufficient health-benefiting physical activity vary according to season.

What this study adds

  • This study found that the influence of seasons on physical activity is activity specific.

  • Age and sex moderate the relationship between season and participation in specific types of physical activity.

  • Seasonality should be considered when monitoring physical activity and when designing and implementing physical activity interventions.

  • Creating physical-activity-friendly environments that incorporate attributes to overcome weather-related constraints are important for promoting year-long regular physical activity participation.


The Calgary Health Region conducted the telephone interviews and the Canadian Institutes for Health Research (CIHR) funded the study. Career salary support provided by the Alberta Heritage Foundation for Medical Research (AHFMR) and the CIHR to Alan Shiell and by AHFMR to Christine Friedenreich is gratefully acknowledged. Gavin McCormack is supported by an AHFMR Postdoctoral Fellowship. Billie Giles-Corti is supported by an Australian NHMRC/NHF Career Development Award.



  • Funding Canadian Institutes for Health Research (CIHR); Alberta Heritage Foundation for Medical Research (AHFMR); National Health and Medical Research Council (NHMRC); National Heart Foundation (NHF).

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the Conjoint Health Research Ethics Board at the University of Calgary.

  • Provenance and peer review Not commissioned; externally peer reviewed.