Large-scale spatiotemporal deep learning predicting urban residential indoor PM_2.5 concentration

Indoor PM_2.5 pollution is one of the leading causes of death and disease worldwide. As monitoring indoor PM_2.5 concentrations on a large scale is challenging, it is urgent to assess population-level exposure and related health risks to develop an easy-to-use and generalized model to predict indoor PM_2.5 concentrations and spatiotemporal variations at the global level. Existing machine learning models of indoor PM_2.5 are prone to deliver single-point predictions, and their input strategies are not widely applicable. Here, we developed a Bayesian neural network (BNN) model for predicting the distribution of daily average urban residential PM_2.5 concentration based on multiple data sources available from nationwide comprehensive sensor-monitoring records in China. The BNN model showed good performance with a 10-fold cross-validation R² of 0.70, mean-absolute-error of 9.45 μg/m³, root-mean-square error of 13.3 μg/m³, and 95 % prediction interval coverage of 85 %. To demonstrate the application process, this model was applied to predict indoor PM_2.5 concentrations on a large spatiotemporal scale. Our modeled population-weighted annual indoor PM_2.5 concentration for China in 2019 was 22.8 μg/m³, far exceeding the WHO standard. The validity of the model at the population level can be further bolstered, making it valuable for assessing and managing indoor air pollution-related health risks.