The effect of meteorological conditions on aerosol size distribution in Istanbul

Kuzu S. L., SARAL A.

Air Quality, Atmosphere and Health, vol.10, no.8, pp.1029-1038, 2017 (SCI-Expanded) identifier

  • Publication Type: Article / Article
  • Volume: 10 Issue: 8
  • Publication Date: 2017
  • Doi Number: 10.1007/s11869-017-0491-y
  • Journal Name: Air Quality, Atmosphere and Health
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.1029-1038
  • Keywords: MLR, Particle size distribution, Pearson’s correlation coefficient, PSCF, Respirable fraction, Thoracic fraction
  • Istanbul Technical University Affiliated: No


Ambient aerosols were sampled by a high-volume cascade impactor in Istanbul, through May 2012 and November 2014. Seventy-eight size-segregated samples were gathered within the period at six different stages. The particles exhibited tri-modal distribution. The peak at <0.49 μm was the most dominant among the others. The average mass median diameter was 1.3 μm. The average total suspended particulate concentration was 75 μg m−3, and PM10, PM4, PM2.5, and PM1 concentrations, derived from log-probability plots, were 62.5, 52.9, 46.9, and 34.2 μg m−3, respectively. Particle concentrations related to meteorological conditions through Pearson’s correlation coefficient. The Pearson’s correlation coefficient was poor in describing the association between coarse particles and meteorological conditions due to the increased urban effect, short-range transportation of marine aerosols, and long-range transportation. Particles >7.2 and 7.2–3 μm had a strong relation, indicating same sources. Increased relative humidity enriched 0.95–1.5-μm particle fraction in winter. Particles between 0.49 and 3 μm were inversely related to ambient temperature. Dilution effect of the wind was significant for PM1.5. Wind acted as a source for larger particles by carrying them from other source regions. Multiple linear regression was applied to particulate matter fractions in order to model the concentrations of each fraction related to meteorological data. In the model, the particle fractions of 1.5–0.95 and 0.95–0.49 μm exhibited the highest prediction performance.