We investigated the associations between rainfall and objectively measured physical activity in 9-10-yr old British children. We found that overall physical activity, and minutes spent in MVPA decreased with increasing rainfall, while sedentary time increased. However, school break time policies were seen to significantly modify these associations. Contrary to our expectations, being allowed outdoors at break time during wet weather was associated with a decrease in MVPA and an increase in sedentary time on the wettest days. It is possible that when they are outside in wet weather children focus on staying dry rather than playing, and therefore spend their break times standing under shelter. Others have found that children are more active when they are outdoors [22–24], but it appears that this association is dependant on rain conditions, and that children who are allowed outdoors in wet weather exhibit similar activity levels to those kept indoors in a classroom environment.
We found that it was the children who were given the opportunity to be physically active indoors in wet weather who were able to remain most active. Children in this group had significantly higher counts per minute, spent longer in MVPA and were less sedentary than those allowed outdoors, and those kept indoors and not given physical activity opportunities. These differences were most obvious over the lunchtime period, but were also statistically significant over the whole day, where the difference in MVPA on the wettest days between those allowed out and those allowed to be active indoors was over 18 minutes; equivalent to almost a third of a child's recommended daily total of MVPA.
Our findings suggest that a focus on encouraging indoor physical activities in wet weather may help children remain active during school hours, and may also contribute to improved daily physical activity levels on wet days. However, it is unclear what the long term effect of such a policy may be. In adults it has been observed that those who perceive the weather as a barrier are less likely to walk for exercise, and spend less time walking in their home neighbourhoods . In the longer term, a policy of keeping children indoors in wet weather may enhance the perception that the weather is a barrier to physical activity, and this could be detrimental to their physical activity levels as they track into adulthood. It is also noteworthy that six of the seven (86%) schools that allowed indoor physical activities had separate indoor sports halls, compared with 60% of the schools that did not allow indoor PA. This may reflect the fact that schools are limited in their policy choices by their facilities, and where the space is not available for indoor physical activities in wet weather, schools may be advised to focus on encouraging greater activity outdoors, possibly by the provision of adequate wet weather clothing, or sheltered all-weather play spaces. Nevertheless, further work is required to examine how the associations observed here may vary as children age, and also what specific steps may be most effective to engage children in outdoor physical activity in wet weather.
This study has a number of strengths and weaknesses. In terms of strengths, we were able to recruit a large sample of schools and pupils, and objectively measured both physical activity and rainfall. This allowed us to examine in detail activity of differing intensities at specific points in the day, and gave variation in rainfall exposure within subjects. Furthermore our measures of school policy were reported directly by head-teachers rather being taken from secondary sources.
The study's weaknesses include the restricted period over which data were collected; our study was conducted over just one season (mid-April to mid-July), which saw unusual weather conditions across the UK . A six-week dry episode starting in the second week of March included the warmest April on record, and preceded the wettest summer since 1912 . The first three weeks of our study period saw no rainfall, while later months, especially June and July, were considerably wetter, giving a rainfall total for the period double the average . We cannot say whether the associations we observed would usually be seen at this time of year, or at other times of the year, nor begin to disentangle seasonal and weather effects. For example, it may be that the magnitude of wet weather effects is greater in winter when temperatures are lower than during our sampling period. A further limitation is that our sample of children is from a restricted age range and we do not know if associations we observed would be apparent in younger or older groups.
Participants provided up to six days of activity data, but these were over consecutive days (or two separate sets of consecutive days, separated by a weekend), potentially limiting the variability of rainfall each participant was exposed to. While our rainfall data were from an official source, and were available at hourly intervals, they were only from two locations. The mean distance for schools to nearest weather station was 29km (SD:14km), and while we were able to derive daily rain totals with the reasonable confidence that they would be representative of the daily rainfall for each school, the distance between stations and schools meant that we did not attempt to measure whether it was actually raining during the lunchtime period. Nevertheless, we hypothesise that lunchtime bad weather policies at schools may be activated on days when there is rainfall generally observed or forecasted during the day, diminishing the importance of actual conditions during the lunch period.
While accelerometers provide an objective measure of physical activity, they are not without their limitations. They have a limited ability to assess activity while the wearer is cycling , and must be removed altogether during aquatic activities. However, these are unlikely to be activities undertaken by children during their lunch break so their under-measurement is not likely to have impacted our findings. A further limitation of hip worn accelerometers is that they do not capture upper body movements and hence we may have underestimated the activity levels of children who sit but are very active with their arms during periods of the day. As physical activity data were not available at a less than hourly temporal resolution, we used a broad definition of lunchtime (12noon - 2pm) which encompassed the lunch period at all schools. However, this period would also include non-break time, and any activity undertaken then.
The majority of the schools in our study had similar policies for break times in bad weather, namely children were kept indoors and not given the opportunity to be physically active. Despite the relatively small number in the other two policy groups, we were still able to detect significant differences in children's activity levels on wet days. Twenty-seven of our schools reported neither allowing physical activities nor additional sedentary opportunities indoors in bad weather. We do not know what children at these schools are allowed to do at break in bad weather, although their activity levels were not significantly different from those kept in and allowed to do additional sedentary activities (results not presented). The questions in the head teacher questionnaire were framed around 'bad' weather. In a UK summer context this would typically be synonymous with rainfall, but we did not ask schools to actually define what 'bad' weather meant to them. A final limitation is that schools in Norfolk, and consequently our sample, have a low proportion of non-white pupils, which potentially limits the generalisability of our findings to more ethnically diverse populations.