Drought analysis by standardized precipitation index (SPI) and determination of drought occurrence probabilities through using Markov chains in the Eastern Mediterranean region

The aims of this study were to: (1) determine drought indices of Standardized Precipitation Index (SPI) based on the long-term monthly precipitation records of meteorological stations located in the Eastern Mediterranean Region, and (2) figure out drought occurrence probabilities through using Markov chains technique. Totally, data of 58 stations were utilized in the study. Missing values in the precipitation series of meteorological observation stations were firstly estimated by using regression analysis method. SPI series for 1?, 2?, 3?, 6?, 9?, 12?, 24?, and 48?month time scales were, then, established through using the completed precipitation series of each meteorological station considered. A two-state first order Markov Chain Model was used to develop transition probability matrix (TPM) for SPI series of time scales considered. In the long term, the equilibrium probabilities of the chain were obtained by making the TPMs steady-state, i.e. stationary, for each time scales. The equilibrium probabilities were used to evaluate drought characteristics of the region. Research results show that Tufanbeyli is the most critical station for the 48?month time scales with the 70% drought occurrence probability in the long term. However, drought risk was found the lowest in Gözne and Nizip stations for rainfall amounts observed in 1?month time scale. Areal extent of expected drought probabilities revealed that the increase in sequential time scales resulted in an increase at expected drought probabilities, inversely in the probability of rainy periods. Research findings are expected to be used by decision-makers, producers, insurance institutions, and technical staff of agricultural extension services in the region.