An Assessment of the Effect of Land Use Change on the Runoff Using the Markov Chain and Cellular Automata in the Bidgol Watershed, the Province of Fars

The present study was conducted to predict land use change and its impact on runoff of the Bidgol Watershed in the Fars province in 2032. The SWAT hydrological model was first calibrated using the hydrometeorological collected data during the 2004 to 2013 period based on the SUFI-2 algorithm and validated using the 2014 to 2018 data. The values of the statistical parameters of the Nash – Sutcliffe model efficiency coefficient (NS), coefficient of determination (R2), deflection amplification factor (P-factor) and the response modification factor (R-factor) at the calibration stage were 0.74, 0.77, 0.8, 0.83, respectively while for the validation stage, the obtained values were 0.68, 0.66, 0.72 and 1.3, respectively. Then Landsat satellite images of 2004 and 2018 were used and following the necessary steps, the images were classified into six main land use classes. Using the Markov Chain method and Cellular Automata based on the land use maps of 2004 and 2018, the land use forecasting map of 2032 was prepared. The land use maps were imported into the SWAT calibrated model and the impact of land use change from 2018 to 2032 on the basin runoff was predicted. The results showed that the highest change of land use will be related to the conversion of rangelands into agricultural land, and the highest percentage of change will be related to the residential land use. In response to these changes, the annual average runoff value of 2032 represents a 19% decrease as compared to that of 2018.