Temporal variation of oxides of sulphur and nitrogen in ambient air in eastern Canada: 1979–1994

ABSTRACT The temporal variation of daily ambient concentrations of particulate SO42−, NO3−, NH4+ and gaseous SO2 and HNO3, and total‐N03 (i.e. HNO3 + NO3) in eastern Canada was studied by fitting (using the maximum likelihood statistical technique) a temporal model that included long‐term trend, long‐term cycles, seasonal cycles and an auto‐regressive model. The technique allowed for missing data and used kernel smoothing and spectral analysis to facilitate the selection of the right model. It was first found that the long‐term trends were neither linear nor monotonic. Between the beginning of the 80s and 90s, SO2 ambient air concentration decreased by 21 to 43% at 5 of the 8 sites. During the same period HNO3 and total‐NO3 concentrations increased at 7 of the 8 sites. For the other compounds, no systematic variations were found for that period and any overall increases or decreases were small. Long‐term cycles were found for 70% of the time‐series but their amplitudes were small and they represented only a small fraction of the total variance for each time‐series. Seasonal cycles with large amplitudes were found only for SO2. For the other compounds, statistically significant seasonal cycles were obtained but they had small amplitudes. Auto‐regressive models of order 2 were necessary for most of the time‐series. For some SO2 and NO3− time‐series, auto‐regressive models of order 1 were sufficient. For the other compounds, auto‐regressive models of order 3 were necessary for some of the time‐series. These results indicate that the air masses that arrived at the sites were less homogeneous in the case of SO2 and NO3− than for the other compounds.