Toward unraveling airborne pathogen transmission in crowds: Parameter study for an agent-based exposure model

The coronavirus disease 2019 (COVID-19) pandemic has put forth the integration of pathogen transmission into microscopic crowd models to simulate the exposure risk for individuals. However, it is crucial to account for uncertainties in the model input as long as we lack data, particularly regarding airborne transmission of the coronavirus. In this study, we quantify uncertainties in such simulations to increase their reliability and informative value. We consider this an integral aspect of model validation. This study relies on a model originally introduced in a previous contribution. We adapt it to simulate airborne virus transmission in several everyday situations. The locomotion layer of the model captures crowd management strategies, while its pathogen transmission layer returns the virtual persons’ exposures to pathogens. We conduct a global sensitivity analysis to rank uncertain parameters according to their impact on the exposure risk and employ forward propagation techniques to quantify the output uncertainty. The sensitivity analysis reveals that two model parameters related to the extent and spread of aerosols are essential, whereas a parameter describing the decay of pathogens is barely influential. The forward propagation demonstrates how crowd management alleviates the exposure risk in the analyzed situations. Moreover, we identify aerosol spread as a dominant aspect on which research should focus.