Background & AimsMany determinants of vitamin D status have been well-described, yet supplementation guidelines largely follow a one-size-for-all model and deficiency remains common. We hypothesised that accounting accurately for ultraviolet-B (UVB) radiation and considering interactions could advance understanding of vitamin D status. MethodsAsian, Black, and White participants from the UK Biobank cohort were included (N=438,978). The Tropospheric Emission Monitoring Internet Service provided UVB data which we linked to participants’ place of residence. UVB dose over 135 days prior to blood draw was weighted and added, yielding cumulative and weighted UVB (CW-D-UVB). The association between 25(OH)D and selected variables was assessed in multivariable linear regression models with and without interactions, stratified by ethnicity. Predictors were ranked using standardised β-coefficients. ResultsMedian 25(OH)D differed by ethnicity (Asian: 25.4 nmol/L (10.2 ng/ml), Black: 30.6 nmol/L (12.2 ng/mL), White: 47.9 nmol/L (19.2 ng/mL), p-value<0.001). CW-D-UVB was strongly associated with 25(OH)D in all ethnicities. It was the most important predictor in White (βAsian=0.15, βBlack=0.20, βWhite=0.35), whereas supplementation was in Asian and Black participants (βAsian=0.30, βBlack=0.24, βWhite=0.21). We identified statistically significant interactions between BMI:supplementation (all), CW-D-UVB:sex (Asian and White), and CW-D-UVB:age (Black and White), and in White population between CW-D-UVB and supplementation, BMI, and cholesterol. ConclusionVitamin D deficiency was widespread, particularly among non-White individuals. UVB was a strong predictor of 25(OH)D and the effect was modified by other factors. Findings suggest that accurately measured ambient-UVB radiation and interactions could improve 25(OH)D prediction models, and support personalised approaches to vitamin D optimisation.