Applicability of benchtop multi-wavelength polar photometers to off-line measurements of the Multi-Angle Absorption Photometer (MAAP) samples

In this study, the applicability of a benchtop polar photometer (PP_UniMI) to retrieve multi-wavelength aerosol absorption coefficient data by off-line measurements of the Multi-Angle Absorption Photometer (MAAP) sample spots is presented. MAAP is a widespread single wavelength online instrument providing the aerosol absorption coefficient and the equivalent black carbon concentration. In this work, MAAP samples collected during four different campaigns were analysed off-line with PP_UniMI. First of all, data from PP_UniMI and MAAP were compared to investigate contributions to measurement uncertainties. In particular, the role of the following assumptions performed in the MAAP was investigated: 1) reconstructing angular distribution of light scattered by filter samples from measurements at three fixed angles using analytical functions; 2) setting the fraction of the back-scattered radiation by the blank filter (BM) at a fixed value BM = 0.7; 3) assuming a fixed value for the asymmetry factor g = 0.75. Samples collected at several sites showed an agreement within 5% when data from the two instruments were retrieved using the same approximations (i.e. backscattered radiation from the filter matrix BM set at a fixed value, phase functions reconstructed by analytical functions from measurements at 3 angles, same asymmetry factor) in the data retrieval algorithm. Conversely, larger differences (14% on average) between off-line measurements and averaged MAAP data were obtained when the high-angular resolved information available by PP_UniMI was exploited. By analysing the role of MAAP assumptions for σap retrieval, it resulted that fixing BM = 0.7 was the main responsible for the detected differences. Indeed, high-angular resolved off-line measurements by PP_UniMI allow to directly measure BM, obtaining BM = 0.88 on white spots. It is noteworthy that the observed results were similar at all investigated sites, so they proved to be independent of the aerosol type and can likely be attributed to instrumental effects. Moreover, a sensitivity test was carried out also to check the impact of the fixed value used for the asymmetry factor (set at g = 0.75 in both instruments). Varying g values within the typical range for ambient aerosol (0.50–0.75) the estimate of aerosol absorbance ABS (directly obtained from PP_UniMI measurements and linked to σap) was affected by 8% at most, thus being a minor source of uncertainty in the calculation. The effect of the variability in blank spots used for off-line analyses was also evaluated and resulted in a contribution smaller than 3% to the uncertainty of the methodology employed. Finally, the possibility of exploiting multi-wavelength assessment of absorption coefficients is an added value of PP_UniMI; indeed, it allows to estimate the contribution of different aerosol sources and components to the absorption coefficient using MAAP tapes used in present or past campaigns.