Comparison between Generated Data by Different Markov Chain Methods in the Mola Sany Station of the Karun River in Iran

Abstract

For programming and management of water resources, access to hydrometric data is necessary. Unfortunately, the quantity and quality of hydrometric data are not suitable in developing countries. These data can not show drought periods and wet periods correctly. Prediction of wet periods is an important problem for flood control while forecasting of drought periods is very vital for water supply in critical conditions. In this research, two hydrometric stations on the Karun River were considered (Ahvaz and Mola Sany hydrometric stations). The data of Ahvaz hydrometric stations are more exact than of their Mola Sany hydrometric stations. For generation of synthetic hydrometric data in Mola Sany hydrometric station, two methods were applied: 1- Generation of yearly hydrometric data by yearly Markov chain method and monthly hydrometric data by monthly Markov chain method in Mola Sany hydrometric station 2- Generation of yearly and monthly hydrometric data by multi spatial Markov chain method and generated yearly and monthly hydrometric data by yearly and monthly Markov chain method in Ahvaz hydrometric station By comparison results of two methods, it is observed that multi spatial Markov chain method can produce data of wet and drought periods better than yearly and monthly Markov chain method while the results of yearly and monthly Markov chain method and observed data have more fitness. Two methods can predict the length of time of wet and drought periods exactly too.