Reconstruction of effective cloud field geometry from series of sunshine number

A new method is proposed for extracting the parameters of effective cloud field models from time series of sunshine number (SSN). Data of SSN number and point cloudiness during 2009 and 2010 at Timisoara (Romania, South Eastern Europe; temperate continental climate) are used to illustrate the method. Two procedures of fitting the estimated point cloudiness to the observed point cloudiness data are proposed and tested. Seven simple effective cloud field models are analyzed. All models underestimate the point cloudiness. The MBE ranges between − 0.06 and − 0.23 while RMSE between 0.15 and 0.38, depending on the month and the duration of the SSN data averaging interval. The best model is based on a field of clouds of semicircle form. This agrees with previous results obtained in the semi-arid climate of Great South Plains in US. The dynamics of the effective cloud field is reconstructed during all months of 2010 at Timisoara. The time series of effective cloud fields are dominated by semicircle clouds but short episodes of semielliptic clouds, ellipsoid clouds, truncated cone clouds and cuboidal clouds are included in the series.