Ground-based sky radiometers were used to measure direct solar irradiance and solar aureole radiance for several years at Sapporo, Tsukuba, and Tokyo, Japan. From these measurements, we computed aerosol optical thickness at 0.5 mum, tau(0.5), and the Angstrom exponent, alpha, and volume size distributions within a column. The optical thickness at Sapporo increased markedly over a short period of time following Asian dust events, and a forest fire in Siberia. The columnar volume size distributions observed during the Asian dust events showed a peak radius of 2.0-3.0 mum. Backward trajectory analyses suggest that the particles producing this springtime event originated in the Loess Plateau and Gobi Desert, and reached Sapporo via southern China. The columnar size distribution during the forest fire case showed an increase in the density of particles with a peak radius -0.2 mum. Trajectory analysis clearly linked the atmospheric changes over Sapporo with a forest fire in Siberia. The aerosol optical thickness, tau(0.5), has a clear seasonal cycle at Sapporo, with a vernal maximum and an autumnal minimum. The Angstrom exponent, a, has a clear seasonal cycle at both Tokyo and Tsukuba, where early-winter maxima and springtime minima are observed, but at Sapporo the seasonal cycle is weaker, with a summer maximum and a vernal minimum. Aerosols were classified into four types (Types I-IV) based on tau(0.5), and alpha data observed at the three sites. Aerosols with a tau(0.5) smaller than the total mean of tau(0.5), but greater than or equal to the total mean of 0 (tau(0.5) < (τ) over bar (0.5), alpha greater than or equal to (α) over bar) were classified as Type I; aerosols with tau(0.5) greater than or equal to (τ) over bar (0.5) and alpha greater than or equal to (α) over bar were Type II; those with tau(0.5) < (τ) over bar (0.5) and alpha < (α) over bar were Type III; and those with tau(0.5) greater than or equal to (τ) over bar (0.5), alpha < (α) over bar were Type IV. The most common aerosol type, that is, the background aerosol, was Type I (similar to40%) at all three sites. Type-IV aerosols at all three sites showed the same seasonal cycle (spring maximum), suggesting large-scale phenomena such as Asian dust events may contribute to the production and transport of this aerosol type. The behavior of Type-II aerosols differed at the three sites, indicating that local phenomena are important in the production and transport of Type-II aerosols. The emission of manufactured aerosols and subsequent gas-to-particle processes may contribute to Type-II aerosol formation. Type-III aerosols at Sapporo were characterized by a seasonal cycle opposite to that of Type-IV aerosols, suggesting that large particles have different sources and/or transport processes in spring and autumn. The atmospheric turbidity retrieved from the sky radiometer is compared, beta(SR), at Sapporo with that calculated using direct solar radiation measurements, beta(DSR). Both showed the same seasonal cycle, but beta(SR) was slightly smaller (similar to0.02) than beta(DSR). Sky radiometers can retrieve the optical properties of aerosols under cloudy conditions if there are no clouds around the solar aureole. The seasonal mean of tau(0.5) (or alpha) under cloudy conditions was 1.5 to 1.8 (1.1 to 1.2) times larger (smaller) than under clear sky conditions. In other words, the direct effect of aerosols that exist in clear air between clouds cannot be ignored. We call this the hybrid effect of aerosols.
