Evaluation of effect of Markov order on the accuracy of drought forecasting Based on SPEI index using Markov Chain method

Drought is an inseparable part of climatic fluctuations which caused to lot of damages to different sections annually. Therefore, drought prediction can be effective in reducing the damages of it. Markov Chain method is one of the drought prediction methods. In this study using climatological data of 5 synoptic stations of Iran with different climate condition during 1967 to 2014, effect of Markov Chain order (orders 1, 2 and 3) in drought forecast accuracy (Based on SPEI Index) using Markov Chain method were evaluated. Based on the results, transition probability from all drought classes to class 4 (normal) had the most frequency and to class 1 (extreme wet) had the lowest frequency. According to results, in the conditions that observed and predicted drought classes had not difference, Markov with order 3 has the best results. In the conditions that observed and predicted drought classes had one class difference, in 60% of cases Markov with order 3 and in 40% of cases Markov with order 2 has the best results. In the conditions that observed and predicted drought classes had two or three classes difference, Markov with order 3 has the best results. So it’s recommended, to forecast drought using SPEI index and Markov Chain method, in the first grad Markov with order 3 and in the second grad Markov with order 2 to be used.