Influence of Climatic Variability on Detected Drought Spatio/Temporal Variability and Characteristics by SPI and RDI

In this paper, climatic variability influence on detected drought spatio-temporal variability and characteristics by Standardized Precipitation Index (SPI) and Reconnaissance Drought Index (RDI) are investigated using data from ten synoptic weather stations with various climatic conditions in Iran. Data record lengths of 34-, 44-, 54- and 64-years are used to represent the impact of climatic variability. Long- and short-term temporal trends of generated SPI and RDI series are, respectively, assessed by the Mann–Kendall and Lepage tests. Drought characteristics, i.e., number and maximum duration of severe drought event (SDE) and extreme drought event (EDE), are also investigated. Dispersion from the mean value of computed drought characteristics and the number of short-term temporal trends obtained at different record lengths are used to assess the influence of climatic variability. Dispersions are measured by normalized root mean squared error (NRMSE), with values < 10% indicating no climatic variability influence. According to the Mann–Kendall tests, SPI/RDI results are not generally influenced by climatic variability in detecting (not detecting) long-term temporal trends. The Lepage tests show that increased record length increases the number of short-term temporal trends (change-points), slightly. However, climatic variability influence is not verified by the NRMSE values. The NRMSE values verify that climatic variability influences SPI and RDI results of numbers of SDEs and EDEs in some stations. Except for one station, SPI and RDI results of maximum durations of SDEs and EDEs are not influenced by climatic variability. According to the present study, climatic variability can influence SPI and RDI analysis of drought spatio-temporal variability/characteristics. Emphasizing the challenge in drought investigations as new sources of data become available. More importantly, the efficiency of the artificial neural network (ANN) model is investigated to determine the nonlinear relationship between the inputs and outputs variables.