Comparison of UV irradiances from Aura/Ozone Monitoring Instrument (OMI) with Brewer measurements at El Arenosillo (Spain) - Part 2: Analysis of site aerosol influence
The main objective of this study is to compare the erythemal UV irradiance (UVER) and spectral UV irradiances (at 305, 310 and 324 nm) from the Ozone Monitoring Instrument (OMI) onboard NASA EOS/Aura polar sun-synchronous satellite (launched in July 2004, local equator crossing time 01: 45 p. m.) with ground-based measurements from the Brewer spectrophotometer #150 located at El Arenosillo (South of Spain). The analyzed period comprises more than four years, from October 2004 to December 2008. The effects of several factors (clouds, aerosols and the solar elevation) on OMI-Brewer comparisons were analyzed. The proxies used for each factor were: OMI Lambertian Equivalent Reflectivity (LER) at 360 nm (clouds), the aerosol optical depth (AOD) at 440 nm measured from the ground-based Cimel sun-photometer (http://aeronet.gsfc.nasa.gov), and solar zenith angle (SZA) at OMI overpass time. The comparison for all sky conditions reveals positive biases (OMI higher than Brewer) 12.3% for UVER, 14.2% for UV irradiance at 305 nm, 10.6% for 310 nm and 8.7% for 324 nm. The OMI-Brewer root mean square error (RMSE) is reduced when cloudy cases are removed from the analysis, (e. g., RMSE similar to 20% for all sky conditions and RMSE smaller than 10% for cloud-free conditions). However, the biases remain and even become more significant for the cloud-free cases with respect to all sky conditions. The mentioned overestimation is partially due to aerosol extinction influence. In addition, the differences OMI-Brewer typically decrease with SZA except days with high aerosol loading, when the bias is near constant. The seasonal dependence of the OMI-Brewer difference for cloud-free conditions is driven by aerosol climatology. To account for the aerosol effect, a first evaluation in order to compare with previous TOMS results (Anton et al., 2007) was performed. This comparison shows that the OMI bias is between +14% and +19% for UVER and spectral UV irradiances for moderately-high aerosol load (AOD>0.25). The OMI bias is decreased by a factor of 2 (the typical bias varies from +8% to +12%) under cloud-free and low aerosol load conditions (AOD<0.1). More detailed analysis of absorbing aerosols influence on OMI bias at our station is presented in a companion paper (Cachorro et al., 2010).