Statistical study of day and night hourly patterns of columnar aerosol properties using sun and star photometry

This work focuses on the statistical analysis of day and night hourly pattern of columnar aerosol properties. To that end, we use the large database of star-photometry measurements at the University of Granada station (37.16 degrees N, 3.60 degrees W, 680 m a.s.l; South-East of Spain) for nighttime characterization, and co-located AERONET measurements for the daytime. The aerosol properties studied are the aerosol optical depth (AOD), Angstrom parameter (alpha(440-870)) and aerosol optical depths of fine (AOD(fine)) and coarse mode (AOD(coarse)) through the Spectral Deconvolution Algorithm (SDA). Microphysical properties are calculated by inverting AOD spectra and include the effective radius (r(eff)) and volume concentration (V) of the total size distribution, and also the effective radius of the fine mode (r(fine)). The initial analysis for the different air masses that reach the study area reveals that generally day and night values of AOD and a(440-870) are not different statistically. Nighttime values of AOD(fine), r(eff) and r(fine) do however, present larger values. The influence of North African air-masses is remarkable both during the day and night, with high particle loads and low values of the Angstrom parameters and also with large contribution of coarse particles as AOD(coarse) and r(eff) values are almost the double than for other air masses. The analyses of day-to-night hourly values reveal an increase in AOD, AOD(fine) and AOD(coarse) during the day and a decrease during the night. Such a pattern could be explained by the different emission rates, accumulation, aging and deposition of particles. Changes in particle radius are also observed as part of the day-tonight particle evolution process, being r(fine) variations important mainly at daytime while for r(eff) variations are more important at nighttime. Results of day-to-night evolution were found to be independent of air-mass origin, and seem to be mainly associated with local processes.