The main goal of this paper is to estimate a set of optimal seasonal, daily, and hourly values of atmospheric turbidity and surface radiative parameters Angstrom’s turbidity coefficient (β), Angstrom’s wavelength exponent (α), aerosol single scattering albedo (ω o ), forward scatterance (F c ) and average surface albedo (ρ g ), using the Brute Force multidimensional minimization method to minimize the difference between measured and simulated solar irradiance components, expressed as cost functions. In order to simulate the components of short-wave solar irradiance (direct, diffuse and global) for clear sky conditions, incidents on a horizontal surface in the Metropolitan Area of Rio de Janeiro (MARJ), Brazil (22° 51′ 27″ S, 43° 13′ 58″ W), we use two parameterized broadband solar irradiance models, called CPCR2 and Iqbal C, based on synoptic information. The meteorological variables such as precipitable water (u w ) and ozone concentration (u o ) required by the broadband solar models were obtained from moderate-resolution imaging spectroradiometer (MODIS) sensor on Terra and Aqua NASA platforms. For the implementation and validation processes, we use global and diffuse solar irradiance data measured by the radiometric platform of LabMiM, located in the north area of the MARJ. The data were measured between the years 2010 and 2012 at 1-min intervals. The performance of solar irradiance models using optimal parameters was evaluated with several quantitative statistical indicators and a subset of measured solar irradiance data. Some daily results for Angstrom’s wavelength exponent α were compared with Angstrom’s parameter (440–870 nm) values obtained by aerosol robotic network (AERONET) for 11 days, showing an acceptable level of agreement. Results for Angstrom’s turbidity coefficient β, associated with the amount of aerosols in the atmosphere, show a seasonal pattern according with increased precipitation during summer months (December–February) in the MARJ.
Read full abstract