Modeling atmospheric aging of small-scale wood combustion emissions: distinguishing causal effects from non-causal associations

Abstract

Small-scale wood combustion is a significant source of particulate emissions. Atmospheric transformation of wood combustion emissions is a complex process involving multiple compounds interacting simultaneously. Thus, an advanced methodology is needed to study the process in order to gain a deeper understanding of the emissions. In this study, we are introducing a methodology for simplifying this complex process by detecting dependencies of observed compounds based on a measured dataset. A statistical model was fitted to describe the evolution of combustion emissions with a system of differential equations derived from the measured data. The performance of the model was evaluated using simulated and measured data showing the transformation process of small-scale wood combustion emissions. The model was able to reproduce the temporal evolution of the variables in reasonable agreement with both simulated and measured data. However, as measured emission data are complex due to multiple simultaneous interacting processes, it was not possible to conclude if all detected relationships between the variables were causal or if the variables were merely co-variant. This study provides a step toward a comprehensive, but simple, model describing the evolution of the total emissions during atmospheric aging in both gas and particle phases.