Trend analysis of extreme precipitation events across Iran using percentile indices

Abstract

Abstract Although a number of studies have been conducted on extreme precipitation trends in different parts of the world including Iran, a great number of such studies have reported only the total amount of daily precipitation greater than a certain percentage (e.g., 95%) of the long term data (R95p), ignoring other useful indices. To address this research gap, we used other modified indices, namely R95 tot (fractional contribution of very wet days to annual total amounts), R95 tt (fractional contribution of very wet days to the total annual obtained from fitted gamma probability distribution), and RS95 (same as R95 tt except that it uses Weibull distribution and very wet days defined by 95 percentage of an individual year) by which the spatial and temporal changes of very wet days across Iran was assessed, 1985–2013. In addition, to evaluate the effect of the selected distribution on the results, a new index‐(RS95 gm )—was introduced and reported. This index is similar to RS95, except that it uses gamma distribution instead of Weibull. According to trend analysis of R95p, R95 tot , and R95 tt , reduced frequency of extreme precipitation events was detected in some northwest, west and northeast parts of Iran. On the contrary, RS95 (RS95 gm ) results showed a higher frequency of extreme events across Iran. It was also demonstrated that while R95p, R95 tot , and R95 tt were unequivocally affected by changes in the mean wet‐day/ annual total precipitation, RS95 (RS95 gm ) was more influenced by changes in the distributional shape, showing more stable trends. Although RS95 and RS95 gm were highly correlated with only 19% difference on average, their trend analysis results were not completely consistent (70% agreement). Thus, it may be concluded that any changes in statistical distribution in the calculation of the RS95 would have a considerable effect on whether the obtained trend is significant or not.

Topics & Concepts