The station is equipped with a Cimel Electronique 318

A spectral radiometer. The methods used to measure solar radiation and the instrument description are given, e.g. in Holben et al. (1998) or Smirnov et al. (2003). Only the data for clear sky situations (level 2.0) selleckchem are employed in this study, i.e. the data following automatic cloud-screening and visual correction by the operator. The algorithms for cloud-screening and the retrieval of aerosol properties are given in Dubovik and King, 2000 and Smirnov et al., 2000. The measurement error of the aerosol optical thickness is estimated to be in the range of 0.01–0.02 for λ > 380 nm, and 0.02 for UV (Holben et al., 1998 and Eck et al., 1999). The Gotland AERONET station (57°55′N, 18°57′E) lies in the northern part of the island of Gotland, 50 m inshore (Figure 1). Owing to the location of the island in the central Baltic Sea this station was adopted as being representative of Baltic Sea conditions. The data collected at the Gotland station from 1999 to 2003 comprise about 11200 measurements, which are distributed unevenly over the measurement period. Because of the small amount of data available in winter, the winter periods were www.selleckchem.com/screening/autophagy-signaling-compound-library.html not taken into consideration in this study. A meteorological

dataset from the Fårosund meteorological station (57°55′N, 18°58′E) from 1999–2003 was also used in the present paper. In particular, observations of wind speed and

direction and relative humidity were used. The station is located near the Gotland AERONET station. Meteorological observations were registered every 3 hours. The wavelength dependence of aerosol optical thickness can be expressed using an empirical formula described by Ångström (Weller and Leiterer, 1998, Smirnov et al., 1994, Eck et al., 1999 and Carlund et al., 2005): equation(1) AOT=βλ−α.AOT=βλ−α. The coefficient β characterizes the degree of atmospheric turbidity due to aerosols and equals the aerosol optical thickness for λ = 1 μm. The exponent α(λ1, λ2) (Ångström exponent) determines the Florfenicol slope of spectral AOT(λ) on a log-log scale ( Smirnov et al. 1994), and for the spectral range from λ1 to λ2 it can be expressed as follows: equation(2) α(λ1,λ2)=ln AOT(λ1)−ln AOT(λ2)lnλ1−lnλ2; α(λ1, λ2) as defined in formula (2) is sensitive to errors in AOT(λ) measurements, which are rather high when the aerosol content in the atmosphere is low. To minimize this error individual spectra AOT(λ) were smoothed by fitting a second order polynomial to the original data ( Eck et al., 1999 and O’Neill et al., 2001): equation(3) ln(AOT)=a0+a1lnλ+a2(lnλ)2. The Ångström exponent for the wavelength range λ = 440–870 nm was calculated on the basis of formula (2). The data were additionally examined with respect to their quality.