Interpreting CO₂ monitoring for assessing airborne transmission risk in indoor environments

Carbon dioxide (CO₂) monitoring has been used for over a century to indicate indoor air quality, and in recent decades to measure ventilation. However, its use for indicating airborne infection risk remains controversial. CO₂ is emitted by everyone and removed almost solely by ventilation, whereas virus-containing aerosols are emitted only by infectors and removed by other additional mechanisms. Accordingly, methods for interpreting CO₂ monitoring data should be developed, and suitable conditions for its use need to be clarified. We propose two CO₂-based surrogates, derived from monitored CO₂ concentrations, to better interpret airborne infection risk. The first, per-person dilution (pCO₂), mirrors CO₂'s emission rate but adopts the removal dynamics of infectious aerosols. The second, room dilution (rCO₂), shares pCO₂'s removal rate but is emitted solely by infected individuals. We define scaling coefficients to estimate these surrogates and compare their temporal profiles with CO₂ concentration excess to assess the impacts of occupant density, ventilation, and non-ventilation removal mechanisms. Using a year-long dataset from two indoor environments, we demonstrate how pCO₂ and rCO₂ effectively indicate steady-state and transient dilution ability, respectively. These surrogates provide a framework for interpreting CO₂ data to assess airborne infection risk under specific conditions.