Inter-comparison of seasonal variability and nonlinear trend between AERONET aerosol optical depth and PM10 mass concentrations in Hong Kong

Here we used Empirical Mode Decomposition (EMD) method to study seasonal variability and nonlinear trend of corrected AERONET Aerosol Optical Depth (AOD/Hi) and corrected PM₁₀ mass concentrations (PM₁₀×f(RH)) in Hong Kong during 2005–2011. AOD/Hi is highly correlated with PM₁₀×f(RH) in semi-annual and annual time scales (with correlation coefficient 0.67 for semi-annual and 0.79 for annual components, 95% confidence interval). On the semi-annual scale, both AOD/Hi and PM₁₀×f(RH) can capture the two maxima in March and October, respectively, with much stronger amplitude in March probably due to the long-range transport of dust storm. On the annual cycle, the AOD/Hi and PM₁₀×f(RH), which are negatively correlated with the precipitation and solar radiation, vary coherently with the maxima in February. This annual peak occurs about one month earlier than the first peak of the semi-annual variability in March, but with only half amplitude. During 2005–2011, both AOD/Hi and PM₁₀×f(RH) exhibit the pronounced decreasing trend with the mean rate of 14 μg m^?3 per year for PM₁₀×f(RH), which reflects the significant effects of the air pollution control policy in Hong Kong during the past decade. The nonlinear trend analysis indicates that the decreasing of PM₁₀×f(RH) is slower than that of AOD/Hi when the AOD/Hi is less than 0.44 but becomes faster when the AOD/Hi exceeds 0.44. These results illustrate that the AERONET AOD can be used quantitatively to estimate local air-quality variability on the semi-annual, annual, and long-term trend time scales.