Since aerosols are an integral part of the Arctic climate system, understanding aerosol radiative properties and the relation of these properties to each other is important for constraining aerosol radiative forcing effects in this remote region where measurements are sparse. In situ measurements of aerosol size distribution, aerosol light scattering and absorption were taken near Eureka (80.05 degrees N, 86.42 degrees W), on Ellesmere Island, in the Canadian High Arctic over three consecutive years to provide insights into radiative properties of Arctic aerosols. During periods of Arctic haze, we find that the single scattering albedo (SSA) at 405 nm is generally higher and more stable than that determined at 870 nm, with values ranging between 0.90-0.99 and 0.79-0.97, respectively. Events with elevated absorption coefficients (B-abs) exhibit generally an absorption Angstrom exponent (AAE) of around 1 suggesting that black carbon (BC) is the dominant absorbing aerosol for the measurement period. AAE values close to 2 occurring with scattering Angstrom exponent (SAE) values near 0 and SAE values below 0 occasionally observed in December indicate a potential contribution from mineral dust aerosols in late fall and early winter. The apparent real and imaginary parts of the complex refractive index at 405 nm have been found to range between 1.6-1.9 and 0.002-0.02, respectively. The low imaginary component indicates very weak intrinsic absorption compared to BC-rich aerosols. Systematic variabilities between different aerosol optical and microphysical properties depend strongly on the given wavelength. SSA at 405 nm shows a strong inverse dependence with B-abs, because B-abs correlates positively with the imaginary component of the refractive index. On the other hand, SSA at 870 nm correlates with scattering coefficient (B-sca) and not with B-abs due to a greater sensitivity to the ambient particle size distribution for 870 nm scattering. Smaller particles with higher SAE that are prevalent during less polluted periods only weakly scatter at 870 nm leading to lower SSA when B-sca is also low. Lastly, FLEXPART back-trajectories show that lower aerosol SSA and higher B-abs correspond to air masses which are more influenced by Eurasian and Alaskan regions, including regions known to have important BC emissions. This work emphasizes the important variability in Arctic aerosol optical properties during winter and spring, which is likely due to changes in source regions.

This work presents a methodology for obtaining vertical profiles of aerosol single scattering properties based on a combination of different measurement techniques. The presented data were obtained under the iAREA (Impact of absorbing aerosols on radiative forcing in the European Arctic) campaigns conducted in Ny-angstrom lesund (Spitsbergen) during the spring seasons of 2015-2017. The retrieval uses in-situ observations of black carbon concentration and absorption coefficient measured by a micro-aethalometer AE-51 mounted onboard a tethered balloon, as well as remote sensing data obtained from sun photometer and lidar measurements. From a combination of the balloon-borne in-situ and the lidar data, we derived profiles of single scattering albedo (SSA) as well as absorption, extinction, and aerosol number concentration. Results have been obtained in an altitude range from about 400 m up to 1600 m a.s.l. and for cases with increased aerosol load during the Arctic haze seasons of 2015 and 2016. The main results consist of the observation of increasing values of equivalent black carbon (EBC) and absorption coefficient with altitude, and the opposite trend for aerosol concentration for particles larger than 0.3 mu m. SSA was retrieved with the use of lidar Raman and Klett algorithms for both 532 and 880 nm wavelengths. In most profiles, SSA shows relatively high temporal and altitude variability. Vertical variability of SSA computed from both methods is consistent; however, some discrepancy is related to Raman retrieval uncertainty and absorption coefficient estimation from AE-51. Typically, very low EBC concentration in Ny-angstrom lesund leads to large error in the absorbing coefficient. However, SSA uncertainty for both Raman and Klett algorithms seems to be reasonable, e.g. SSA of 0.98 and 0.95 relate to an error of +/- 0.01 and +/- 0.025, respectively. (C) 2017 Elsevier Ltd. All rights reserved.