Direct and diffuse photosynthetically active radiation: measurements and modelling

Knowledge of photosynthetically active radiation is necessary in different applications dealing with plant physiology, biomass production and natural illumination in greenhouses. Nevertheless, due to the absence of widespread measurements of this radiometric flux it is often calculated as a constant ratio of broadband solar radiation. This ratio is affected by many parameters. In our previous work, we analysed the global horizontal component of this flux. This work describes the variations of the direct and diffuse ratios of photosynthetically active radiation to broadband solar radiation in sky conditions. The latter has been characterised by means of variables that could be accessible in common radiometric networks. For this purpose, data recorded at two radiometric stations are used. The first one is located at the University of Almeria, a seashore location (36.83 degrees N, 2.41 degrees W), while the second one is located at Granada (37.18 degrees N, 3.58 degrees W, 660 m a.m.s.l.), an inland location. The database includes hourly values of the relevant variables that cover the years 1993-1994 in Almeria and 1994-1995 in Granada. We have studied the variability of the ratios of photosynthetically active radiation to broadband solar radiation, both for the diffuse and direct components. We have explained this variability through the influence of sky conditions, represented by means of broadband solar radiation dimensionless ratios, solar zenith angle and dewpoint temperature. Additionally we have developed weather dependent functions of this ratio, to obtain an empirical model, which could be applied to estimate the photosynthetically active radiation in those radiometric networks that register solar broadband irradiance. The model validation shows that all the models estimate the experimental values with mean bias deviation (MBD), close to 1-2%. On the other hand the root mean square deviation (RMSD), presents values close to the experimental error. The regression analysis of measured versus estimated values of direct and diffuse photosynthetically active photon flux density, Q(pb) and Q(pd), indicates the absence of given ranges with special tendencies to over or underestimate the fluxes. The use of data sets from two different climatic conditions enables the verification of the local independence of the proposed technique. (C) 1999 Elsevier Science B.V. All rights reserved.