Spatiotemporal Variation Of PM ₁₀ Concentration And Its Influencing Factors In The Southern Suburb Of Xi'an

With the development of urbanlization and industrialization in Xi’an, air pollution increasingly deteriorated, particulate matter has been main contaminant. Research on temporal and spatial variations of inhalable particle concentrations of Xi’an is not only to reveal scientific significance of the variation of pollutants but also control atmospheric particulate pollution in Xi’an and improve atmospheric environmental quality, which was important practical significance.This study take sourthen suburb in Xi’an for example, PM10concentrations and meteorological factors was monitored respectively under7-day in June and August. Relationships between temporal and spatial variations of PM10concentrations and meteorological parameters was analysed by correlation coefficient method. Source of pollution of PM10concentrations was analysed by wind rose and reverse trajectories. The effect of PM10concentrations that different models was used to predict was judged by mean absolute difference and root-mean-square error. From this study we can get the following main results:(1) Diurnal variations of PM10concentrations at1～52m in the sourthen suburb of Xi’an was similar, which totally indicated bimodal distribution, reaching peak at8:00～10:00and4:00-6:00respectively. According to the overall trend, it can be divided into three stages. At the first phase (8:00～18:00) which was the lowest concentration stage, the average of PM10concentrations was58μg·m-3, with the range of44～76μg·m-3. at the second phase (18:00～4:00) which was moderate concentration stage, the average of PM10concentrations was61μg·m-3, with the range of47～89μg·m-3. At the third phase (4:00-8:00) which was the highest concentration stage, the average of PM10concentrations was77μg·m-3, with the range of71～83μg·m-3.(2) Relationship between hour average of PM10concentrations and meteorological factors in the sourthen suburb of Xi’an showed that hour average of PM10concentrations and relative humidity were significantly positive correlation, in comparison with the significantly negative correlation between PM10concentrations, mixed layer height, solar radiation intensity, wind speed, temperature and depression of the dew point.(3) Wind rose and reverse trajectory curve of PM10concentrations indicated that higher PM10concentrations was NE45°, SE135°and SW225°, these three wind upwind mainly distributed construction site and transportation corridors. So it can be initially supposed that the main impact of local sources in the area may beconstruction dust and automobile exhaust. Remote transportation sources may be from more developed industrial city or mineral-rich area.(4) Overall, PM10concentrations in the sourthen suburb of Xi’an in June and August within1～52m indicated four distribution types, Ⅰ distribution showed that PM10concentrations constantly decreased as altitude increased. Ⅱ distribution (10:00～12:00and12:00～14:00;16:00～18:00and18:00～20:00) showed that PM10 concentrations indicated bimodal distribution. Ⅲ distribution (14:00-16:00and20:00～2:00;2:00～4:00and6:00～8:00) indicated that PM10concentrations slightly increased then decreased with increasing of altitude. IV distribution (4:00-6:00) suggested that PM10concentrations first decreased then increased with increasing of altitude.(5) In comparison with BP neural networks model and traditional multiple linear regression model to predict the effect of PM10concentrations, BP neural networks model in forcasting PM10concentrations was more accurate under a conditions of moderate pollution.