Analyzing time series of SPI, SPEI and SPTI drought indices by using artificial neural network SOFM method and numerical comparison in chaharmahal va bakhtiari

Analysis and comparison of drought indices are several of the studies needed to monitor and evaluate drought by multiple indicators. Drought evaluation is typically performed using climatic and hydrological parameters. So far, several indices have been introduced to the analysis of this phenomenon. This paper in order to monitor drought in Chaharmahal and Bakhtiari province is compared to the Performance of three drought indicators, SPEI, SPI and SPTI. The basis of the comparison of indices is the use of the SOFM neural network, which by use of the results of the topology of this network, is resulted that the drought indices are in one category or not? Subsequently, the frequency analysis of drought classes and all types of drought analysis have been done. The results show that both numerical and SOFM methods can analyze and evaluate the outputs of drought indices, accurately. Also, according to the obtained results, the most severe period of drought had been occurred during the period between 1988 and 1990 in all three drought indices. On the other hand, the most severe drought was happened in this study in September 2003 based on the SPEI index at Behesht Abad Station with a value of -5.9. From other results of this paper can also be referred to the high sensitivity of the SPEI than two other drought indices in calculations.