Seasonality, meteorology and ascertainment in paediatric rotavirus and adenovirus detection: a nine-year daily time-series in a semi-arid continental climate

Seasonal and meteorological associations are widely reported for enteric viruses, but evidence is dominated by tropical and subtropical settings and often lacks mutual adjustment of covarying exposures. We examined the environmental correlates of laboratory detection of rotavirus and adenovirus over nine years in a cold semi-arid continental climate. All stool rotavirus and adenovirus antigen tests in children aged 0–17 at a tertiary centre in Konya, Türkiye, 2017–2025 were analysed (20,382 visits, 15,228 children). Daily positive and negative counts were modelled with mutually adjusted binomial models containing calendar-time splines and autocorrelation-consistent standard errors. Because spline flexibility sets the temporal scale of the estimate, models were fitted at one to seven degrees of freedom per year. Rotavirus was detected at 14.93% of tested visits, adenovirus at 2.88%. At the seasonal-contrast specification, higher temperature was associated with lower odds of rotavirus detection (odds ratio 0.642 per 14.7 °C, 95% CI 0.527–0.781), whereas the estimate attenuated to 0.936 (0.761–1.152) under the short-term-deviation specification. Absolute humidity was not associated in the prespecified two-variable model (0.918, 0.826–1.020) and reached 0.817 (0.729–0.916) only in the fully adjusted model. Snowfall was positively associated before adjustment (1.728, 1.432–2.085) but not identified once temperature was accounted for (0.834, 0.706–0.986; q = 0.058). For adenovirus only shortwave radiation retained an association (0.640, 0.480–0.853), which attenuated similarly under finer calendar-time adjustment (0.996, 0.584–1.700). Detection also varied by care setting and day of week. Rotavirus detection was associated with lower temperature at the seasonal scale, with little evidence of an independent short-term association; the two scales are not fully separable here. The absolute humidity finding was model-dependent and should not be interpreted as an independent effect. Adenovirus findings were constrained by group-antigen detection, prolonged faecal shedding and the smaller number of detections. Overall, long-term laboratory surveillance can yield seasonal associations that are robust to analytical choices within a dataset, but their interpretation depends critically on temporal scale, assay provenance and the tested denominator. not applicable.