Fig. 3: Modelled contributions of respiratory viruses to pharmaceutical loads.

Left: time series plots of observed baseline-subtracted median pharmaceutical loads (black line) and the modelled contributions, derived from OLS regression coefficients applied to the median viral gene copy loads of SARS-CoV-2, RSV, IAV and IBV. The coloured areas represent the stacked contributions of each virus to the overall baseline-subtracted (bl-subtracted) pharmaceutical load. The baseline loads were determined from location-specific mean values during periods of low respiratory virus exposure (June 2021 and July to August 2023). The OLS coefficient determined for SARS-CoV-2 was applied to data from early 2021 through late October 2022, the period during which only this virus was monitored in wastewater. Right: scatter plots that compare observed versus predicted baseline-subtracted pharmaceutical loads from the OLS analyses. The data points are colour-coded by time periods: November 2022 to February 2024 (blue) and March 2024 to June 2024 (beige, indicating poor model fit with viral aetiologies). Each scatter plot includes the uncentred coefficient of determination (R²) and the number of observations (N). The diagonal black line represents the 1:1 line, indicating perfect agreement between observed and predicted values.