An "in situ" calibration-correction procedure (KCICLO) based on AOD diurnal cycle:: Comparative results between AERONET and reprocessed (KCICLO method) AOD-alpha data series at El Arenosillo, Spain

A comparative evaluation is carried out for nearly 5 years (February 2000 to May 2004) of data of aerosol optical depth (AOD) measured at the Aerosol Robotic Network (AERONET) site El Arenosillo (Huelva, southwestern Spain). The AERONET database and the reprocessed data set using a new correction procedure, which we call the KCICLO method, are compared with respect to the aerosol local climatology. The cause and necessity of AOD reprocessing were due to the existence of an observed fictitious diurnal cycle (including negative values) because of a deficient calibration as explained in detail in the companion paper (V. E. Cachorro et al., submitted manuscript, 2007). The derived alpha Angstrom coefficient is also compared, as it appears to be an excellent indicator of the AOD data quality, because of its sensitivity to AOD variations and errors. Some illustrative cases show the influence of this fictitious diurnal cycle on the shape and values of diurnal variations of the AOD (or alpha), reaching differences as high as 100%, and the improvement resulting from using the KCICLO method. Absolute and relative differences are evaluated from the overall average of AOD and alpha coefficient of AERONET and KCICLO data series, making an exhaustive analysis for each spectral channel and for every photometer separately. Although great variability is shown for each filter and each photometer, apart from photometer 114 data that did not reach level 2.0, the discrepancy in the AOD local climatology in the four filters varies as a whole from 2.3% to 8.5% (2.4% for alpha coefficient). These values show a considerable reduction because of the compensating effect between the different photometers (positive or negative bias), and several jumps that break the continuity of the data series are observed. When monthly and yearly averages are analyzed, the differences are considerably reduced in such a way that the local climatology is not substantially affected, but we must be cautious with this apparent good result. The comparative results between both data series give a representative pattern of the uncertainties that we have found in establishing a local aerosol climatology.