Sunny days, cloudy behavior? Associations between weather patterns and movement behaviors in preschoolers using compositional data analysis

PURPOSE: To investigate associations between daily weather conditions and movement behaviors (moderate-to-vigorous physical activity MVPA, light physical activity LPA, and sedentary behavior SB) in Swedish preschool children, and whether these differed by sex and day type (weekdays vs. weekends). METHODS: A total of 3307 preschool children (4.5±0.9 years; 51.4% girls) wore accelerometers to measure daily MVPA, LPA, and SB during 14 days across two periods. Weather data (temperature, sunshine, daylight duration, precipitation, and snow) were obtained from the Swedish Meteorological and Hydrological Institute. Principal component analysis followed by k-means clustering classified days into weather categories. Compositional data analysis using isometric log-ratio transformation and linear mixed-effects models estimated associations, with predictions back-transformed to minutes/day using bootstrap procedures. RESULTS: Four weather clusters were identified: sunny and warm, cold and cloudy, rainy, and cloudy days. Weather clusters were significantly associated with daily movement composition (p=0.001). Compared with sunny and warm days, cold and cloudy days were associated with -15.1 min MVPA, -32.0 min LPA, and +46.5 min SB. Rainy days showed -11.0 min MVPA, -8.3 min LPA, and +19.4 min SB, while cloudy days showed -8.3 min MVPA, -7.2 min LPA, and +15.6 min SB. Sex interactions were not significant (p>0.05). In contrast, day type significantly modified associations for cold and cloudy and rainy days (p<0.001), with larger reductions in PA on weekends. CONCLUSION: Daily weather conditions were significantly associated with the movement-behavior composition of Swedish preschool children. These findings demonstrate how short-term environmental variability can substantially influence young children's daily activity patterns. Recognizing weather as a dynamic environmental exposure is therefore essential for accurately interpreting movement behavior data and for understanding how environmental conditions shape human activity patterns in early childhood.