Estimating photosynthetically active radiation using an artificial neural network

Photosynthetically active radiation (PAR) is a portion of solar radiation in the wavelength band of 400-700 nm providing energy for photosynthesis of plants. In this work, we proposed to estimate PAR from atmospheric parameters using an artificial neural network (ANN). The input data of the ANN are solar zenith angle (θz), cloud index derived from MTSAT-1R satellite together with precipitable water from NCEP/NCAR database and aerosol optical depth from AERONET of NASA. The PAR data at 4 stations in Thailand, namely Chiang Mai (18.78°N, 98.98°E), Nakhon Pathom (13.82°N, 100.04°E) and Songkhla (7.20°N 100.60°E) for the years 2008-2010 and Ubon Ratchathani (15.25°N, 104.87°E) for the years 2009-2010 were used to train the ANN. The estimated PAR using ANN was validated at these stations in the year 2011. It was found that the estimated PAR from ANN and those obtained from the measurements were in good agreement, with root mean square difference (RMSD) of 10.2%.