Football matches as natural experiments: Exploring wastewater quality parameters for dynamic population normalization in sub-sewersheds.

Estimating the contributing de facto population is essential for deriving per-capita rates in wastewater-based epidemiology (WBE), yet remains a major methodological challenge, particularly at the sub-sewershed scale. This pilot study uses sold-out football matches with well-documented, large-scale population fluctuations as a natural experiment to explore wastewater quality parameters as dynamic population normalization proxies in small-scale sub-sewersheds. Hourly wastewater samples (n = 24) were collected during pre-defined match periods from two sub-sewersheds with contrasting characteristics in the Ruhr area, Germany. Total nitrogen (TN), ammonium nitrogen (NH₄-N), chemical oxygen demand, and biochemical oxygen demand after five days were quantified as normalization proxies. Pharmaceutical compounds were used as exemplary human-specific biochemical markers to evaluate the feasibility of dynamic population normalization. Hourly mass loads were calculated using continuous flowmeter data, z-score standardised for temporal comparison, and used to estimate de facto populations from site-specific per-capita emission rates. All four wastewater quality parameters captured match-related population dynamics, with clearer signals in the smaller, stadium-dominated sub-sewershed. TN and NH₄-N provided the most comparable de facto population estimates in the larger sub-sewershed (mean de facto/de jure ratios of 1.04 for both), whereas all parameters substantially over- or underestimated the reference population in the smaller one. Dynamic normalization using TN and NH₄-N indicated that population influx, rather than elevated per-capita consumption, was the primary driver of match-related increases in pharmaceutical loads. Our results suggest that normalization proxy performance is sewershed-dependent, even at smaller spatial scales, and highlights the importance of context-specific proxy selection and subsequent validation in small-scale WBE.