Impact of weather, sleep and physical activity on daily affective fluctuations: an ecological momentary assessment study

Weather conditions are increasingly recognised as potential factors contributing to affective states, yet their role in emotional dysregulation remains poorly understood. Earlier research has primarily examined weather effects on mean mood levels, with limited attention to affective instability or the influence of behavioural factors such as physical activity and sleep. The main aim of this study is to examine associations between weather variables (sunshine duration, temperature, rainfall), physical activity, and sleep with (1) momentary positive affect and (2) affective variability in a transdiagnostic clinical sample.

Data used were from the Bergen site of the DynAMoND study, a longitudinal multisite ecological momentary assessment (EMA) project. Participants comprised adults with ADHD, borderline personality disorder (BPD), bipolar disorder (BD), and healthy controls (HC), who completed at least one of five five-day EMA bursts over one year, receiving ten smartphone prompts daily. Momentary affect was assessed on a visual analogue scale (0–100); the valence dimension served as the primary outcome. Meteorological data were obtained via the open-meteo.com API, linked to participants through GPS coordinates. Physical activity was derived as the Euclidean Norm Minus One (ENMO) from wrist accelerometer, in the 60 minutes preceding each prompt, and sleep duration was estimated from actigraphy. Mixed-effects models with random intercepts per participant were estimated using REML. Affective variability was operationalised as absolute residuals from the fully adjusted model, entered as the dependent variable in a second-stage.

Lagged affect was the strongest predictor of momentary affect across all models (β ≈ 0.44–0.45, p < .001), reflecting its robust inertia. Sunshine duration was the only weather variable significantly associated with momentary affect across both weather-only and fully adjusted models, with a consistent positive direction of effect. Daylight duration was significant only when modelled independently of sunshine duration, suggesting that actual solar sunshine length than astronomical season is relevant. Temperature and rainfall were not significantly associated with momentary affect in any model specification. Higher physical activity in the preceding hour was positively associated with momentary affect (β ≈ 18.31–18.41, p = .003) and reduced the sunshine effect, suggesting a potential indirect pathway from light exposure to affect via physical activity behaviour. No weather variable significantly predicted affective variability (all p > .29). Longer sleep duration the preceding night was the sole significant predictor of reduced affective variability (β ≈ −0.233 to −0.237, p < .025), independent of weather, physical activity, and time-of-day effects.

Sunshine duration, but not temperature or rainfall, was associated with a momentary affect in a mixed clinical-control sample. There were no significant effects of weather variables on affective variability, whereas physical activity and sleep emerged as more consistent correlates of both the level and variability of momentary affect level and stability. These findings underscore the importance of distinguishing affect level from affective variability and of incorporating individual-level behavioral context when examining factors that influence affective states.

References

1. Denissen, J.J.A., Butalid, L., Penke, L., van Aken, M.A.G., 2008. The effects of weather on daily mood. Emotion 8 (5), 662–667.
2. Kuppens, P., Allen, N.B., Sheeber, L.B., 2010. Emotional inertia and psychological maladjustment. Psychol. Sci. 21 (7), 984–991.