Analysis and synthesis of seasonality and variability of daily potential evapotranspiration

Daily meteorological data over a period of nine years from Helliniko Station (Athens, Greece) are used to produce daily potential evapotranspiration (PET) series. A multipliative model describes the periodic and stochastic components of the series where Fourier aproximations of two- or three-harmonic periodic components and a first-order autoregressive one are employed, respectively. The assessment of adequacy of the model is based on the properties of (a) the correlogram and (b) the spectrum of a white noise process. The model is further used to generate daily PET values. Generated and historical data show remarkable affinity in average arithmetic mean, coefficient of variation, skewness coefficient and their standard deviations.