Improved estimation of diffuse photosynthetically active radiation using two spectral models

The efficiency with which incoming photosynthetically active radiation (PAR) is intercepted by a canopy depends on the proportions of diffuse and direct radiation in the incoming PAR. In the present work, we developed a comparison between two spectral cloudless sky parameterization schemes in terms of their capability to provide accurate estimates of global, diffuse and direct incoming PAR. Both methods provide an estimation of the solar spectral irradiance that can be integrated spectrally within the limits of interest. For this purpose, data recorded at two radiometric stations one located at Granada, an inland location, and the other at Almerı́a, a coastal location, were used. After our study, it appears that the information concerning the aerosol radiative effects is fundamental to obtain a good estimation, especially for the direct and diffuse components. In this sense SMARTS2 model offers increased flexibility concerning the selection of different aerosol models included in the code. The original version of SPCTRAL2 code does not offer this flexibility but in this work a modified version that allows the selection between several aerosol models was considered. Our results show that a rather accurate estimation of the global component can be obtained for any of the aerosol models tested in this work, thus indicating the lower influence of the scattering processes in this term. In our analyses, we considered different aerosol models compatible with the features of each one of the analyzed locations. The a priori choice of the aerosol models dictated by the features of each place provided estimation of the different components of PAR with mean bias deviation (MBD) smaller than the experimental error. This is a clear advantage over the simpler broadband parametric models that need local information of all the relevant aerosol properties and some ad hoc modifications to provide accurate estimates of the PAR components. © 2002 Elsevier Science B.V. All rights reserved.