Application of vector autoregressive time series analysis to aerosol studies

Article English OPEN
Hsu, Kuang-Jung (2011)

Interpretation of aerosol processes, such as those related to acid deposition, based on linking laboratory experimental results with atmospheric observations can be difficult, often involving complicated models. In this study, statistical relationships in a time series of aerosol chemical composition measurements are examined to identify different aerosol groups and to estimate their atmospheric lifetimes. Vector autoregressive model (VAR) and vector moving average representations were applied to quantitatively study the relationships among chemical species in aerosols, with emphasis on sulfur oxidation. Data used were collected in Barrow, Alaska from 17 March to 21 April 1986. Results from the statistical analyses indicate: (1) chemical species can roughly be separated into 2 subgroups for either fine or coarse aerosols; and generally, there is no correlation between chemical species in these two subgroups; (2) durations and magnitudes of chemical influences in fine and coarse aerosols are different; (3) the principal species for controlling S variations is Cl or K in coarse particulates and Fe in fine particulates; the reverse relationships are less important; (4) sulfur variations in coarse particles were most likely controlled by the mass transfer process; (5) variations of S-compounds in fine aerosols, dominated by chemical reactions, have estimated lifetimes of 9 to 29 h. Further laboratory and field studies based on these hypotheses are, however, necessary to investigate causality.DOI: 10.1034/j.1600-0889.49.issue3.8.x
Share - Bookmark