LAUSR

Monthly precipitation time series of 46 synoptic stations distributed over diverse climates of Iran were investigated to identify the best‐fitted probability distribution function (PDF) to the monthly precipitation series aggregated at different timescales. The computed skewness and the probability of zero precipitation (PZP) values of the considered precipitation aggregations show that a large proportion of monthly precipitation data in the study period, especially in the warm months of the year, is zero. The skewness of all 12 individual monthly precipitation time series as well as their aggregation at 3‐ and 6‐months timescales were also found very high in many studied stations, particularly in those of the arid and semi‐arid climates of Iran. To identify the best candidate distribution function for computing Standardized Precipitation Index (SPI) in different climatic areas of Iran, the performances of a dozen statistical PDFs in fitting monthly precipitation aggregated at different timescales were statistically evaluated. The results showed that the two‐parameter gamma distribution best fits monthly precipitation time series over most parts of the country while the Pearson Type III (PE3) was identified as the best‐fitted model for 3‐ and 6‐months aggregation periods and general extreme value (GEV) distribution as the best candidate function in fitting precipitation aggregated at 9‐, 12‐, and 24‐month timescales. PE3 was also identified as the second best‐fitted model for both the shorter time scale (1‐month) and the longer timescales (9‐, 12‐, and 24 months). Considering that PE3 ranked either first or second in fitting precipitation time series aggregated at different timescales and in all climatic areas of Iran, it is chosen as the most appropriate distribution function in fitting precipitation aggregated at all considered timescales throughout Iran.