Performance evaluation and modification of SPI in drought monitoring of arid and semi arid regions of Iran

Standard precipitation index (SPI) is the most widely drought monitoring index. However, this index only uses the gamma distribution function for fitting precipitation data and does not consider seasonal variations. The purpose of this study was to evaluate the efficiency of SPI in drought monitoring of arid and semi-arid regions of Iran and fix the related problems with this Index. Then the SPI was compared with its modified state (SPImod) over (1956–2010). The results showed that the generalized extreme value distribution function in more than 57% of the cases was the most appropriate probability distribution function of rainfall data. But the default distribution (Gamma) was selected only in 11% of months. Comparison of Kappa index showed that with increasing time window, the agreement between SPImod and SPI indexes increases. The amounts of one-month Kappa for studied stations was Tehran (0.31), Mashhad (0.33), Bushehr (0.32) and Khorram-Abad (0.26), while for nine-month the Kappa increased. Such that in Tehran (0.49) and in Mashhad, Bushehr and Khorramabad, respectively, with values (0.47), (0.56) and (0.45). Also, the results showed that the frequency and displacement of drought classes would be very variable in comparison to these two indices. As the displacement of normal, severe drought and severe, with a total of 259, 147 and 111 events in the time window-three and displacement of moderate drought, normal and moderate wet year, with a total of 68, 54 And 28 events in time window-nine were noticeable.