摘要
One major characteristic of the Standardized Precipitation Index (SPI) is its flexibility to be calculated in a variety of time scales and hence being aware of different types of droughts. However, various time scales may result in confusion of the water resources-researchers, decision makers and users in identifying and specifying drought periods in a certain region. To solve this problem in this article, a multivariate approach has been utilized having the ability to aggregate a variety of the SPI time series into a new time series called the Multivariate Standardized Precipitation Index (MSPI). The MSPI is based on the principal components analysis (PCA) of the SPI time series in a certain location. Having specified the first principal component’s (PC1’s) scores, the MSPI would be simply attained by dividing PC1’s values by the monthly standard deviation. In this article, MSPI’s capability in depicting the variability of the SPI time series (in five ranges of time scales, including 3-, 6-2, 3-2, 12-4, and 24-8?months) was studied at four weather stations representing the four different climates in Iran from the driest to the wettest climates. The results showed that the PC1 is able to justify more than 74?% of the variability for the selected sets of the SPI time scales in the studied climates. Also, it became clear that the drought and wet severity classes monitored by MSPIs matched very satisfyingly to those of the five sets of the SPI time scales. Therefore, in cases where the aggregation time scales for calculating the SPI are not previously known, this study recommends the researchers use the MSPI.