Relationship between objective measures of physical activity and weather: a longitudinal study

Abstract

Background: The weather may be a barrier to physical activity but objective assessment of this hypothesis is lacking. Therefore we evaluated the effect of temperature, rain or snow, and wind speed on the daily physical activity of adults.

Methods: This report contains data from 25 males (BMI (mean +/- SD): 28.7 +/- 3.83 kg/m2) and 177 females (BMI: 29.2 +/- 5.92 kg/m2) enrolled in an intervention to increase physical activity. Steps/day of the participants was measured by pedometer. Weather data were obtained from Environment Canada. A total of 8,125 observations were included in a mixed linear model analysis.

Results: Significant weather related variables (at the 5% level) impacting steps/day included: seasonal effects related to the interaction between weekday and month; mean temperature, total rainfall, interactions between gender, BMI and total snow, interactions between maximum wind speed and BMI, and the amount of snow on the ground. The estimated magnitudes for the various effects were modest, ranging from approximately 1% to approximately 20%. Thus for an average individual taking approximately 10,000 steps/day, weather-dependent changes in physical activity could reach 2,000 steps/day.

Conclusion: We conclude that weather had modest effects on physical activity of participants in an intervention to increase their activity. It should be stressed that these effects may be different for less or more motivated people. With this in mind, we suggest that the effect of weather on physical activity in the general population needs to be objectively assessed to better understand the barrier it poses, especially as it relates to outdoor recreation or work activities.

Figure 1

Weekly physical activity patterns as measured by pedometer in (A) February, (B) May, (C) July and (D) December. The pattern shown for February and May was similar to that observed in January, March, April and June (not shown). There were no data obtained for August-November. Steps/day on Sunday were significantly lower than the rest of the week in February and May. In July, there was no effect of weekday on steps/day. In December, steps/day were significantly higher on Saturday. Data are means ± 95% CI.

Figure 2

Effect of weather elements on physical activity determined from multivariable analysis. (A) Effect of total rain, an increase in temperature of 10°C and an increase of 10 cm in accumulated (accum.) snow on the ground on steps/day of an individual with a usual activity of 10,000 steps/day. No interaction with either BMI or gender was detected in multivariable analysis. (B) Percent change in steps/day of males and females, segregated into lean (BMI = 20 kg/m²), obese Class I (BMI = 30 kg/m²), or obese Class II (BMI = 35 kg/m²), in response to 10 cm snowfall in a 24 h period. Data are means ± 95% CI and *p < 0.05 or better.