Simulation and prediction of drought using Cellular Automata and Markov methods in Najaf Abad plain

The governing factors of drought are non-linearly correlated. Therefore, researcher needs to apply nonlinear methods such as CA to model and predict the drought. CA and its derivatives are among novel methods of drought simulation that rarely used for predicting the drought. While such methods have simple structures; they provide high visual capabilities for drought monitoring. This paper investigates drought in Najaf Abad plain using Markov, CA Markov and Landsat satellite images. At first, satellite image time series of transpiration were classified for 1995, 2008, 2015, and the land zonation of drought condition were estimated. Then, the drought in 2020 was predicted using CA Markov. The Kappa index is 0.63 and the agreement between actual and predicted map (M (m)) is 0.85. Our findings showed that our proposed model can suitably predict the drought. In addition, the drought distribution map showing the possibility of changes in 2020, suggests that if the situation continues and no changes in the type of cultivation and cropping pattern happen, all areas in danger of drought in 2015, will face drought more intensely and more widely, in 2020.