SIMPLE MODEL TO GENERATE DAILY AVERAGED POINT CLOUDINESS DATA

Abstract

A new simple model to generate daily averaged point cloudiness values was proposed. The model uses as input the long-term mean value of point cloudiness. The model was tested in two Romanian locations. During the cold season there is a good agreement between the frequency distribution functions (FDF) obtained by using observed and generated data, respectively. When the warm season is considered, the concordance between the FDFs based on synthetic and observed data is slightly worse. The present model generates data whose mean and standard deviation are very close to those of the observed data. The model can be used to synthesize time series in those locations where the long-term mean value of point cloudiness is the only known information about the cloud cover amount. However, if both the long-term mean and standard deviation of point cloudiness are known, one recommends to use first or second order autoregressive (AR) models. If one looks about FDFs based on generated data the present model should be preferred to the usual first and second order AR models.