Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4438856 | Atmospheric Environment | 2012 | 14 Pages |
Accuracy requirements for aerosol optical depth (AOD) in polar regions are much more stringent than those usually encountered in established sun photometer networks, while comparability of data from different archive centres is a further important issue. Therefore, two intercomparison campaigns were held during spring 2006 at Ny-Ålesund (Svalbard) and autumn 2008 at Izaña (Tenerife) within the framework of the IPY POLAR-AOD project, with the participation of various research institutions routinely employing different instrument models at Arctic and Antarctic stations. As reported here, a common algorithm was used for data analysis with the aim of minimizing a large part of the discrepancies affecting the previous studies. During the Ny-Ålesund campaign, spectral values of AOD derived from measurements taken with different instruments were found to agree, presenting at both 500 nm and 870 nm wavelengths average values of root mean square difference (RMSD) and standard deviation of the difference (SDD) equal to 0.003. Correspondingly, the mean bias difference (MBD) varied mainly between −0.003 and +0.003 at 500 nm, and between −0.004 and +0.003 at 870 nm. During the Izaña campaign, which was also intended as an intercalibration opportunity, RMSD and SDD values were estimated to be equal to 0.002 for both channels on average, with MBD ranging between −0.004 and +0.004 at 500 nm and between −0.002 and +0.003 at 870 nm. RMSD and SDD values for Ångström exponent α were estimated equal to 0.06 during the Ny-Ålesund campaign and 0.39 at Izaña. The results confirmed that sun photometry is a valid technique for aerosol monitoring in the pristine atmospheric turbidity conditions usually observed at high latitudes.
► Two sun photometer comparison campaigns were conducted for the POLAR-AOD project. ► AOD mean bias differences were found to be within ±0.004 at all wavelengths. ► Ångström exponent α was estimated to be within 0.06 and 0.39 for the two campaigns.