Surface ultraviolet solar irradiance from Ozone Monitoring Instrument versus Brewer ground-based measurements at El Arenosillo (South-Western Spain)
Manuel Anton(1), Victoria Cachorro(2), Jose Manuel Vilaplana(3), Carlos Toledano(2), A Arola(4), N Krotkov(5), Antonio Serrano(1), M.L. Cancillo(1) and Benito de la Morena(3)
(1) Universidad de Extremadura, Avd. de Elvas, s/n, 06071 Badajoz, Spain
(2) Universidad de Valladolid, , Valladolid, Spain
(3) INTA, , El Arenosillo, Spain
(4) FMI, , Kuopio, Finland
(5) GEST Center, , Maryland,, United States
The objective of this work is to compare overpass surface ultraviolet (UV) products (new Collection 3), erythemal UV irradiance, and spectral UV irradiances at 305, 310 and 324 nm from EOS-AURA Ozone Monitoring Instrument (OMI), with ground-based measurements from a well-calibrated Brewer spectroradiometer. This instrument is installed at the ESAt-‘‘El Arenosillo’’ (Huelva) station, located in southwest Spain, (37.10º N, 6.73º W). The period of study extends from 2004 to 2008.
Four different atmospheric conditions are tested in order to analyze the effects of clouds and aerosols on the differences between OMI and Brewer UV data: all-skies, cloud-free skies with any aerosol load, as well as with low and high aerosol loads separately. The OMI Lambertian Equivalent Reflectivity (LER) at 360 nm is used as proxy for cloudiness. Thus, cloud-free cases are selected taking days with LER smaller than 10%. In addition, the aerosol optical thickness (AOT) measured with an automatic CIMEL sun photometer located at El Arenosillo is used as proxy for the aerosol load. The AOT (440 nm) equal to 0.25 is considered as the threshold for separating days with low and moderate-high aerosol load.
The results show that the uncertainty of the relative differences between satellite and ground-based UV data increases for high LER values (cloudy days). In this sense, the scatter in OMI-Brewer correlation is notably reduced when only data for cloud-free conditions are selected. In addition, the accuracy of UV irradiance estimated by OMI for cloud-free conditions is best for low aerosol load cases. Thus, the mean absolute bias error when satellite and ground-based UV data are compared is between 8% and 13%. For cases with moderate-high aerosol load, the bias increases considerably (15-20%). This is due to the fact that absorbing aerosols in the boundary layer are currently not included in the operational OMI OMUVB algorithm. Several research groups have recently suggested off-line correction due to aerosol absorption. The problem is that currently it is very difficult to determine the absorption optical thickness. The associated error is high, especially in the UV region where only recently accurate extinction data can be retrieved. In general, atmospheric aerosol absorption is low and frequently masked by other influences. Only under strong events of pollution, biomass-burning or desert dust, these absorbing values can be properly retrieved. This is currently subject of ongoing research.