DEVELOPMENT OF A NEURAL-NETWORK MODEL TO PREDICT DAILY SOLAR-RADIATION

Many computer simulation models which predict growth, development, and yields of agronomic and horticultural crops require daily weather dat a as input. One of these input is daily total solar radiation, which i n many cases is not available owing to the high cost and complexity of the instrumentation needed to record it. The aim of this study was to develop a neural network model which can predict solar radiation as a function of readily available weather data and other environmental va riables. Four sites in the southeastern USA, i.e. Tifton, GA, Clayton, NC, Gainesville, FL, and Quincy, FL, were selected because of the exi stence of long-term daily weather data sets which included solar radia tion. A combined total of 23 complete years of weather data sets were available, and these data sets were separated into 11 years for the tr aining data set and 12 years for the testing data set. Daily observed values of minimum and maximum air temperature and precipitation, toget her with daily calculated values for daylength and clear sky radiation , were used as inputs for the neural network model. Day-length and cle ar sky radiation were calculated as a function of latitude, day of yea r, solar angle, and solar constant. An optimum momentum, learning rate , and number of hidden nodes were determined for further use in the de velopment of the neural network model. After model development, the ne ural network model was tested against the independent data set. Root m ean square error varied from 2.92 to 3.64 MJ m(-2) and the coefficient of determination varied from 0.52 to 0.74 for the individual years us ed to test the accuracy of the model. Although this neural network mod el was developed and tested for a limited number of sites, the results suggest that it can be used to estimate daily solar radiation when me asurements of only daily maximum and minimum air temperature and preci pitation are available.