Bayesian Time Series Models of Ultrafine Particle Concentrations

• Activity Number: 306
• Type: Contributed
• Date/Time: Tuesday, August 6, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics and the Environment
• Author(s): Heidi Fischer*+ and Robert E Weiss
• Companies: UCLA and University of California, Los Angeles
• Keywords: Bayesian ; Semi-Parametric ; Time Series ; Ultrafine Particle ; Hetereoscedasticity ; Longitudinal

Abstract

Researchers measured airborne ultrafine particle (UFP) concentrations over time in 102 size-bins ranging from 7 up to 280 nm inside school buses, under multiple scenarios representing common exposure conditions where school buses idle to pick up or drop off school children. Each experiment measured UFP concentrations over the 102 size bins for 15 minutes before the bus engine is turned on, then for 30 minutes as the bus engine idles. We develop Bayesian semi-parametric longitudinal functional models incorporating hetereoscedastic errors to model mean log concentration over time across the range of particle sizes. The model has constant mean log concentration prior to engine-on. Once the engine is turned on, mean log concentrations are assumed to increase quadratically over time, with trends varying smoothly by particle size.