What can CO2 measurements tell us about ventilation and infection risk in classrooms?

Indoor air quality in schools is of critical importance for the health and well-being of pupils and staff. The COVID-19 pandemic highlighted the essential role that ventilation systems play in limiting the spread of airborne diseases and consumer CO2 monitors were deployed in UK classrooms as a cost-effective tool to help manage the ventilation supply. In such settings, which are occupied for long periods by the same group of people, CO2 measurements have also been used to infer the risk of far-field airborne infection. Often, only a single point measurement is available and so it is crucial to understand how exposure might differ depending on the infector location and/or sensor position. This is particularly important in UK classrooms which are typically naturally ventilated, and where spatial variations in the concentration of CO2 and infected breath can be expected.
CO2 data gathered across 5 years prior to the pandemic from 45 classrooms in 11 different schools were first analysed and estimates of the likelihood of airborne infection were calculated. Even without accounting for the variation in disease prevalence, the data highlight significant variation in infection risk between the seasons, with January being nearly twice as risky as July. Results also show that the risk can vary widely between classrooms of the same school despite similar ventilation provisions. The second part of this work provides a detailed examination of the validity of using CO2 as a proxy for far-field exposure. For this, a generic naturally ventilated UK classroom in wintertime was simulated using computational fluid dynamics (CFD). The ratio between actual exposure arising from a single infected individual and proxy exposure calculated from point measurements of CO2 was also analysed. In doing so, the proxy exposure is found to be within a factor of two of the actual far-field exposure. While this factor of two might appear large, it is small relative to the typical uncertainties associated with airborne disease modelling.