A study of PM in the urban atmosphere with regard to its reduction and an evaluation of its toxicity and health impact

Principal Investigator (Contact person)

Akinori TAKAMI (email: takami[at]nies.go.jp) "[at] is replaced by @"

  • Overall
  • Publications & Link

Overall

Abstract

   In order to understand better the atmospheric behaviors and health effects of atmospheric aerosols (PM2.5), we conducted a project including field measurements, numerical simulations, toxicity assessments, and epidemiological studies of PM2.5 across Japan. A CTM was improved with an enhanced emission inventory, yield measurement data from an atmospheric experiment chamber, and field observations. Toxicity and epidemiological studies were also performed based on the chemical compositions and mass concentrations of the PM.
  The toxicities of the SOAs generated from the precursors, which were emitted from anthropogenic sources (m-xylene, 1,3,5-trimethylbenzene, and 1,3-butadiene), were higher than those generated from natural sources such as α-pinene and isoprene. Field measurements in the Kanto area of Japan were performed and the chemical composition of the PM analyzed. Oxidative stress of airborne PM was found to be lower than the SOA of anthropogenic-origin. Results of atmospheric simulations were evaluated using the observed PM2.5 chemical composition from the Kanto area. The SOA model, which considers the aging reactions of semi-volatile organic compounds, was better able to reproduce the observed organic aerosol concentrations than traditional SOA models. The epidemiological study revealed that sulfate was associated with respiratory mortality, while there were no significant associations of elemental and organic carbon with mortality.

1. Introduction

   The environmental standard of fine PM (PM2.5) was set in September 2009 in Japan. In order to attain the environmental standard, it is necessary to consider the counter measures used to reduce emissions of air pollutants based on the information of the various emission sources. The environmental standard of PM2.5 was determined based on health effects. However, the toxicity of SOAs produced from various volatile organic compounds (VOCs) is not well known. Therefore, it is necessary to assess the toxicity of SOAs. As epidemiological knowledge is largely dependent on foreign results, an epidemiological study has been performed based on domestic data.
   The following studies are planned.
1) The spatiotemporal distribution of PM concentration will be elucidated based on an improved emission inventory, the physico-chemical properties of PM, field observations, and an improved CTM.
2) Toxicity and epidemiological studies of PM will be performed based on PM mass and chemical composition.

2 Method, Results, and Discussion
2.1 Production of SOAs and study of their toxicity

   To study the toxicity of SOAs, we developed a small reaction chamber system for the α-pinene/ozone reaction and photo-oxidation of m-xylene, 1,3,5-trimethylbenzene, 1,3-butadiene, and isoprene in the presence of nitric oxide. After the reaction, each SOA was collected on a TeflonR filter to determine its chemical properties and toxicities related to oxidative stress. The amount of oxidant was determined by the KI method and the consumption of dithiothreitol (DTT), which is related to the oxidative potential of SOAs, was measured. The solution containing the SOAs was exposed to the alveolar epithelial cells and the cytotoxicity assessed. In addition, the gene expression of heme-oxygenase-1 (HO-1) was assessed, which is a marker sensitive to oxidative stress.
   While the amount of oxidant for the α-pinene-SOA was highest, the consumption of DTT was higher in other SOAs; the isoprene-SOA was greatest and m-xylene-SOA, the second highest. The cytotoxicity and HO-1 induction of m-xylene-SOA was higher than any other SOA and the 1,3-butadiene-SOA was the second highest. The toxicities of the SOAs generated from the precursors emitted from anthropogenic sources (m-xylene, 1,3,5-trimethylbenzene, and 1,3-butadiene) were higher than those generated from natural sources such as α-pinene and isoprene.
   To study the effects of aging on the toxicity of the m-xylene-SOA, aging experiments were conducted for 3, 5, and 15 hours. We found that the cytotoxicity was higher for longer duration reactions, and the HO-1 induction reached a maximum after 5 hours, which indicates that toxicity changes by aging and that aging might play an important role in the effect of SOAs on health. These results also indicated that the time scale of the response differed in the two assays, which suggests that toxic reactants in the SOAs degraded with aging.

2.2 Field observation and toxic study of PM in atmosphere

   A field measurement campaign was undertaken in the Kanto area of Japan in summer 2012, from which the chemical composition of PM was analyzed. The results were compared with the CTM to evaluate the model’s performance. It was found that the CTM overestimated nitrate and underestimated organic matter. The oxidative stress of airborne PM, measured using HO-1, was lower by one order of magnitude than the SOA of anthropogenic origin (e.g., SOA produced from m-xylene) in the atmospheric chamber.
   The major species of PM was organic matter and sulfate, and the fraction of organic matter was about 50%. The positive matrix factorization method was applied to evaluate the oxidation of organic matter and it was established that 80% of the organic matter was oxygenated, which is higher than found in previous studies in urban areas. The correlation between the oxidation of organic matter and oxidative stress of HO-1 was found to be low.

2.3 Improvement of the chemical transport model (CTM) and simulation of PM2.5 distribution

   The improvement of SOA models is critical for accurate understanding of the behavior and sources of atmospheric aerosols. In this study, the results of box-model simulations with five SOA models (two yield models, a volatility basis set (VBS) model, a mechanistic model, and a near-explicit model) were compared. The performances of the models were evaluated by comparison of the simulated data with the observed ratios of SOA concentrations to ozone concentrations in Tokyo. All five models showed similar results for the concentrations of gaseous species including ozone, reactive nitrogen, hydroxy radicals, and VOCs. In contrast, the simulated SOA concentrations varied substantially among the five models. The VBS model reproduced the observed [SOA]/[Ox] ratio well, whereas the other four models substantially underestimated the ratio.
   The evaluation of the models capability in simulating the spatiotemporal variations of PM2.5 chemical compositions in Japan has been limited by the lack of observational data. In this study, we used PM2.5 chemical composition data, measured simultaneously over the Kanto area in summer 2013, to evaluate the results of the 3D simulation models. Three sensitivity simulations were compared: one based on an SOA yield model and two based on a VBS model. The SO42- concentration was reproduced well by all the simulations; however, the NO3- concentration was overestimated by the standard simulation, but reproduced better by a model in which the dry-deposition velocities of nitric acid and ammonia were enhanced by a factor of five, as was performed in a previous study. The OA concentration was greatly underestimated by the simulation based on the SOA yield model, but it was reproduced better by the simulations based on the VBS model, because aging reactions were considered. Among the simulated OAs, biogenic SOAs had the highest contributions, followed by anthropogenic SOAs. These contributions should be validated by means of observation-based source contributions of OA in future studies.

2.4 Distribution of PM concentration and assessment of its risk to health

   Seasonal variation and regional heterogeneity have been observed in the estimated effect of PM2.5 mass on mortality, and it is considered that differences in the chemical compositions of PM2.5 might be responsible for this variation. In the epidemiological study conducted in Nagoya from 2003-2007, we combined daily mortality counts with concentration data for PM2.5 mass and its components. Season-specific analysis of PM2.5 mass revealed a stronger association in warmer seasons. Sulfate, whose concentration was higher in warmer seasons, was associated with respiratory mortality, while there was no significant association of mortality with elemental and organic carbon. Significant associations of respiratory mortality were also found for nitrate and ammonium. These findings suggest that some specific components of PM have effects that are more hazardous compared with others.
   Although the uncertainty is large, risk assessments using estimates of the effect of PM components and the distribution of the PM concentration obtained from the CTM are expected to elucidate the component-specific risk to health attributable to PM.

43 Summary for PM reduction

   It is thought that immediate action is not necessary to reduce SOAs in the atmosphere based on the results of the toxicity studies for the SOAs produced in the lab experiments and PM collected in the atmosphere. The distribution of organic matter, which is representative of toxic organic matter, including the SOAs produced from m-xylene and trimethyl-benzene, is very different from the distributions of sulfate and nitrate. However, in order to reduce the toxic SOAs and organic matter in PM, the reduction of anthropogenic VOCs is still important. The method to reduce PM should be considered in terms of emission amount, emission sources, distribution of air pollutants, and population distribution. It is also important to monitor the effects of reductions in emissions, which requires improvements to the emissions inventory, CTM, lab and field experiments, and toxicity and epidemiological studies.

Publications & Link

Publications

This project was summarized at the NIES report (http://www.nies.go.jp/kanko/tokubetu/pdf/sr109.pdf), but in Japanese. It includes the publication list in the end of the report.