Exploring Different Approaches for Computing the Standardized Precipitation Index to Characterize Dry and Wet Periods in the Yarmouk River Basin, Jordan.

This study aimed to investigate the annual precipitation time series in the Yarmouk River Basin, Jordan, and to characterize dry–wet periods using the standardized precipitation index (SPI). Annual precipitation data of 11 stations in the Yarmouk River Basin were analyzed to explore the stationarity, homogeneity, persistence, and trends over the 36-year study period. It was concluded that the annual precipitation data in the Yarmouk River Basin are purely random and homogenous with no autocorrelation or significant trends. Several statistical distributions were investigated for their fit to the annual precipitation data, and it was found that the loglogistic was the most suitable distribution for all stations. The 12-month SPI was computed via three different approaches: the gamma approximation, the Box-Cox transformation, and the selected distribution (loglogistic) method. These three methods did not yield significantly different SPI values for most stations; however, clear differences can be observed in extremely wet or dry conditions. Therefore, the SPI-gamma approach is recommended for characterizing wet conditions, while the Box-Cox and the loglogistic approaches are recommended for drought assessment. It was also found that major drought events (moderate, severe, and extreme) over the study period range from 3 to 8 per station with an average of 5.48 per station and an average relative frequency of 0.152. The driest year for most stations was observed in 1998–1999 while 1991–1992 was the wettest year. [ABSTRACT FROM AUTHOR]