Land use regression modeling for fine particulate matters in Bangkok, Thailand, using time-variant predictors: Effects of seasonal factors, open biomass burning, and traffic-related factors

Abstract

Abstract In recent years, as the level of fine particulate matter (PM2.5) concentration has become more closely monitored in Thailand and its harmful effects on health have been widely recognized by the public, the Thai government has debated various measures to improve air quality. In this paper, the Land Use Regression (LUR) technique was used to model the relationship between the daily PM2.5 concentration and various predictor variables using data from the entire year of 2019. The results confirmed strong seasonal effects on PM2.5 and substantial effects of time-variant predictors, including open biomass burning and meteorological conditions. However, time-invariant variables, including traffic, transportation, and land use characteristics were generally weaker predictors in the LUR models. The results of the model based on data for the entire year showed better statistical fit and robustness than the seasonal models. The relatively low adjusted R2 of the models developed in this study compared with previous LUR studies suggests that more detailed data, especially the traffic volume on roads nearby monitoring sites, might be necessary to improve the model s performance. Finally, the large buffer size of the open biomass burning predictor implied that the measures to reduce PM2.5 by limiting open biomass burning would require international cooperation as some fires within the buffer area occurred in neighboring countries outside the borders of Thailand.