Intercomparison of AOD retrievals from GAW-PFR and SKYNET sun photometer networks and the effect of calibration

Abstract. In this study, we assess the homogeneity of aerosol optical depth (AOD) between the sun photometer networks Global Atmospheric Watch-Precision Filter Radiometer (GAW-PFR) and European Skynet Radiometers network (ESR) at the 2 common wavelengths of their main instruments 500 nm and 870 nm. The main focus of the work is on evaluating the effect of the Improved Langley calibration method, (ILP) used by SKYNET and investigating the factors affecting its performance. We used data from three intercomparison campaigns that took place in the period 2017–2021. Each campaign has two phases in two locations. One is mountainous rural (Davos, Switzerland) and the other urban (Rome, Italy). Our analysis shows that the AOD differences due to post processing and instrument differences are minor. The major factor leading to AOD differences is the calibration method where we found a systematic underestimation of AOD compared to GAW-PFR due to an underestimation in the ILP calibration. The calibration and AOD differences are smaller in Davos where at 870 nm the traceability criteria are satisfied and at 500 nm the median differences are below 0.01. In Rome at 500 nm the AOD median differences per campaign are between 0.015–0.035. Attempting to explain the differences we found no association between the calibration performance and the level or the variability of the aerosol properties. We also conducted a sensitivity study, which shows that part of the difference can be potentially explained by errors in the assumed surface albedo and instrument solid view angle provided as inputs to the ILP code (based on Skyrad pack 4.2). Our findings suggest that the ILP method is mainly sensitive to the measured sky radiance. The calibration underestimation is probably caused by an error on the retrieved scattering aerosol optical depth (sc-AOD) through the sky radiance inversion. Using an alternative retrieval method (Skyrad MRI pack version 2) to derive sc-AOD and repeat ILP calibration, we found no significant differences between the retrieved sc-AOD nor systematic increase of the calibrations. The potential error may be a result of the forward model assumptions, To conclude, calibration of sun photometers on site offers the advantage of avoiding instrument shipments and data gaps. However, ILP shows larger uncertainty and significant systematic difference compared to the traditional Langley calibration performed under low and constant AOD conditions at high altitude sites, due to the uncertainties of the calibration method and the input parameters needed for it. In the following sections we report on results on the AOD retrievals of several instruments in different environments using different principles in their calibration methods. We also perform an investigation to explain the causes of differences.