The SPI value classifications are shown in table 1. This table also gives the corresponding probabilities of occurrence of each severity from the normal probability density function (Rouault et al, 2003).
For this study, SPI calculated both Multi-temporal and Spatial SPI to facilitate the quantitative comparison of drought incidence over the 29 years period and 308 different rain gage locations. 6-month SPI was calculated from total rainy season from May to October and 12 month-SPI was calculated to describe long-term drought. SPI values of equal or less than -1.0 were used to determine drought intensity according to table 1. The frequency of occurrence of drought years was determined by looking at the number of the dry station (SPI < -1) since 1976.
For Spatial SPI, the precipitation means are calculated from the precipitation amounts of every rain gage station each year to compute SPI value. Then, Krigging method was applied to get a spatial pattern of the severity of drought from SPI value.