Black carbon (BC) is a descriptive term that refers to light-absorbing particulate matter (PM) produced by incomplete combustion and is often used as a surrogate for traffic-related air pollution. Exposure to BC has been linked to adverse health effects. Penetration of ambient BC is typically the primary source of indoor BC in the developed world. Other sources of indoor BC include biomass and kerosene stoves, lit candles, and charring food during cooking. Home characteristics can influence the levels of indoor BC. As people spend most of their time indoors, human exposure to BC can be associated to a large extent with indoor environments. At the same time, due to the cost of environmental monitoring, it is often not feasible to directly measure BC inside multiple individual homes in large-scale population-based studies. Thus, a predictive model for indoor BC is needed to support risk assessment in public health. In this study, home characteristics and occupant activities that potentially modify indoor levels of BC were documented in 23 homes, and indoor and outdoor BC concentrations were measured twice. The homes were located in the Cincinnati-Kentucky-Indiana tristate region and measurements occurred from September 2015 through August 2017. A linear mixed-effect model was developed to predict BC concentration in residential environments. The measured outdoor BC concentrations and the documented home characteristics were utilized as predictors of indoor BC concentrations. After the model was developed, a leave-one-out cross-validation algorithm was deployed to assess the predictive accuracy of the output. The following home characteristics and occupant activities significantly modified the concentration of indoor BC: outdoor BC, lit candles and electrostatic or high efficiency particulate air (HEPA) filters in heating, ventilation and air conditioning (HVAC) systems. Predicted indoor BC concentrations explained 78% of the variability in the measured indoor BC concentrations. The data show that outdoor BC combined with home characteristics can be used to predict indoor BC levels with reasonable accuracy.
Keywords: black carbon; estimation; exposure; modeling.