Search

In the atmosphere, aerosols can originate from numerous sources, leading to the mixing of different particle types. This paper introduces an approach to the partitioning of aerosol mixtures in terms of backscattering coefficients. The method utilizes data collected from the Mie-Raman-fluorescence lidar, with the primary input information being the aerosol backscattering coefficient, particle depolarization ratio (δ), and fluorescence capacity (GF). The fluorescence capacity is defined as the ratio of the fluorescence backscattering coefficient to the particle backscattering coefficient at the laser wavelength. By solving a system of equations that model these three properties (bF, δ and GF), it is possible to characterize a three-component aerosol mixture. Specifically, the paper assesses the contributions of smoke, urban, and dust aerosols to the overall backscattering coefficient at 532 nm. It is important to note that aerosol properties (δ and GF) may exhibit variations even within a specified aerosol type. To estimate the associated uncertainty, we employ the Monte Carlo technique, which assumes that GF and δ are random values uniformly distributed within predefined intervals. In each Monte Carlo run, a solution is obtained. Rather than relying on a singular solution, an average is computed across the whole set of solutions, and their dispersion serves as a metric for method uncertainty. This methodology was tested using observations conducted at the ATOLL observatory, Laboratoire d'Optique Atmosphérique, University of Lille, France. [ABSTRACT FROM AUTHOR]

The objective of the paper is to validate SolaRes (Solar Resource estimate) in clear-sky conditions, and to examine the aerosol influence on the differences between observation and estimate. SolaRes has the ambition to fulfil both research and industrial applications exploiting downwelling solar radiation at surface level. Consistently with solar resource applications, we show the capacity of SolaRes to reproduce the angular behaviour of the angular field, by validating not only global horizontal irradiance (GHI), but also direct normal irradiance (DNI), diffuse horizontal irradiance DifHI), global and diffuse irradiance in tilted plane (GTI, DifTI), and even the circumsolar contributions. Computations are made with the SMART-G radiative transfer code, taking spectral aerosol optical thickness (AOT) data sets as input, which are delivered by the Aerosol Robotic network (AERONET) and the Copernicus Atmospheric Monitoring Service (CAMS). A mixture of two aerosol models is required to compute aerosol optical properties. Measurements for validation are made at two sites in Northern France. Clear-sky is identified by two methods to show its influence: 1) a method reproducing the AOT variability conditions, and 2) a stricter method eliminating some residual cloud influence but also conditions with largest AOT. SolaRes is validated according to comparison scores found in the literature, with the (relative) root mean square difference (RMSD) in GHI as low as 1%, and the mean bias difference (MBD) which could be 0%. Angular behaviour is reproduced with satisfying scores. The circumsolar contribution improves MBD in DNI and DifHI, by 1% and 4% respectively, as well as RMSD by ~0.5%. MBD in DNI is around -1% and RMSD around 2%, and MBD in DifHI is 2% and RMSD around 9%. RMSD and MBD in both DNI and DifHI are larger than in GHI because they are more sensitive to the aerosol and surface properties. DifTI measured in a vertical plane facing South is reproduced with a RMSD of 8%, similar to DifHI. It is suggested a strong influence of reflection by not only ground surface but also surrounding buildings, increasing the albedo from 0.13 to 0.35. The sensitivity study on the aerosol parameterisation shows that the spectral AOT contains enough information for best quality in DNI retrieval. The choice of aerosol models in the parameterisation has an influence in RMSD smaller than 0.7%. Complementary information on angular scattering and aerosol absorption has a significant influence in GHI by reducing RMSD by ~0.5%, and MBD by ~0.8%. The uncertainty on the data source has a significant influence. The CAMS data set increases the RMSD in DNI by 5%, but has has less influence in GHI, by increasing RMSD by ~1% and MBD by ~0.4%. RMSD in GHI still remains slightly smaller than state-of-the-art methods. [ABSTRACT FROM AUTHOR]