Introduction of new datasets of drought indices based on multivariate methods in semi-arid regions

• Published in: Hydrology Research Vol. 49; no. 1; pp. 266 - 280
• Main Authors: Chitsaz, Nastaran, Hosseini-Moghari, Seyed-Mohammad
• Format: Journal Article
• Language: English
• Published: London IWA Publishing 01.02.2018
• Subjects: Arid zones; Case studies; Climate change; Coefficient of variation; Coefficients; Correlation analysis; Datasets; Decomposition; Drought; Drought index; Hydrologic analysis; Hydrologic drought; Hydrology; Methods; Precipitation; Principal components analysis; Regions; Semi arid areas; Semiarid zones; Singular value decomposition; Standardized precipitation index; Stream discharge; Stream flow; Surface water; Thermal expansion; Time; Variables; Water shortages; Water supply; Iran; China
• ISSN: 0029-1277; 1998-9563; 2224-7955
• DOI: 10.2166/nh.2017.254

Drought is a slow and creeping worldwide phenomenon which has adversely affected arid and semi-arid regions of the world. Drought indices like Streamflow Drought Index (SDI) and Standardized Precipitation Index (SPI) offer quantitative methods for combating probable consequences of drought. In this article, the results of the drought indices trend showed that the case study suffers from hydrological drought more than meteorological drought. The correlation analysis between hydrological and meteorological drought was assessed in monthly and seasonal time scales. To this end, some multivariate techniques were used to summarize the SPI and SDI series of all stations into one new dataset. Three assessment criteria involving higher correlation among drought indices, higher eigenvalue in expansion coefficients, and following fluctuation and variation of original data were used to find the best new datasets and the best multivariate method. Results asserted the superiority of singular value decomposition (SVD) over other multivariate methods. EC1 in the SVD method was able to justify about 80% of the variability in drought indices for monthly time scales, as well as summer and spring for seasonal time series, which followed all fluctuations in original datasets. Therefore, the SVD method is recommended for aggregating drought indices.