Estimation of sunshine duration with artificial neural networks by using NWCSAF cloud type product

Sunshine duration is a very important parameter in many different fields of applications such as climate, renewable energy, agriculture and health. In this respect the estimation and determination of the temporal and spatial variability of this parameter is critical. Sunshine duration have been measured by heliographs in Turkey. Since these measurements are point-wise, estimations need to be done for other locations where there is no measurement available. The geographical location, total cloudiness and type of the clouds are the most dominant factors affecting the sunshine duration for a given place. The geographical affects are possible to calculate by using the astronomical equations while the cloudiness and cloud type factors are harder to estimate and understand. This is mostly because of the temporal and spatial variations of the cloud parameters and their different effects on the sunshine duration. In this study, sunshine duration for all over the Turkey is estimated with artificial neural networks by using the meteorological station measurements and NWCSAF Cloud Type (PGE 02) product which produced by Meteosat satellite data. Keywords: Sunshine duration, Cloud type, Meteosat, Satellite, Turkey.