Urban pollution effects on aerosols hygroscopicity and warm clouds droplets activation in Amazon in context of the GoAmazon 2014/5 experiment

The measurements of the Green Ocean Amazon 2014/5 experiment were carried out on the outskirts of Manaus, in the central Amazon region, for two years, with the objective of understanding how the natural aerosol and cloud life cycles would be perturbed by urban emissions. In this context, the present work aimed at studying the aerosol hygroscopicity under polluted condition, comparing it with the pristine environment. Based on these results, we used an adiabatic cloud parcel model to study the impact of Manaus pollution on the first stages warm clouds formation. We observed high concentrations of aerosols coming from Manaus, with average Ncn = 2.425 cm 3 and percentiles 25 % and 75 % of 937 cm 3 and 3.259 cm 3 res- pectively. For the pristine Amazon, typical values would be about Ncn 400 cm 3 . The hygroscopicity of urban pollution particles were notable low, with average t = (0, 09±0, 01) for all diameters investigated, and a high level of heterogeneity was found. On the other hand, natural particles in the Amazon have a hygroscopicity of about t 0,14 and are not as much heterogeneous. We improved and used an adiabatic cloud parcel model to systematically investigate the impact of the Manaus pollution plume on the first stages of warm cloud develop- ment. The model was validated by comparison with four exemplary cases found in the literature, from conceptually similar models, but with different numerical imple- mentations. In our simulations, we considered that the shape of the size distribution could vary with increasing number concentration, as we moved from the clean to the polluted conditions. We also allowed the hygroscopicity to vary with the concentration and the diameter of the aerosol particles. These were done in stages, hence allowing an increasing level of complexity in the representation of the aerosol particles. We observed that the number of activated cloud droplets is as function primarily of the concentration and the vertical velocity. In second place comes the dependence with the shape of the size distribution and, after that, with the hygroscopicity. We showed that simulations that do not consider these other factors will, necessarily, over predict the effect of aerosols on shallow warm clouds. As expected, when we simulated clean conditions changing towards a polluted one, we found an increase in the number of activated droplets and corresponding decrease of effective radius of those droplets, and of the activated fraction. Our results suggest that, under polluted conditions, clouds accumulate liquid water more rapidly during the first stages of its development than under clean conditions. (AU)