The aim of this study was to analyze the flow trend and to evaluate the effect of the climate change on the discharge changes using HadCM3, ECHAM4 and neural network in the Gorganroud- Gharehsou watershed basin, Golestan province, Iran. For this purpose, discharge trend analysis was performed using the Mann-Kendall, Sen-slope and Kendall’s τ methods during the past 30 years in this important agriculture region in north of Iran. To evaluate the effect of the climate change on the discharge changes HadCM3 and ECHAM4 models and three scenarios of A2, B1 and A1B were used. The LARS-WG statistical model was applied to downscale the two model data. Besides, the multilayered Perception neural network model was used in order to simulate the discharges. The results indicate that during the past 30 years, precipitation and the minimum temperature have had the more significant effect on the river flows. The results show that during 2011- 2030 and in the A2, B1 and A1B scenarios the average seasonal discharge will be reduced in all seasons. The highest decreasing amount is related to the summer in the two models especially in HadCM3 model. The results reveal that the discharge will increase to 4.26, 3.52 and 3.03 % based on the model HadCM3 and 3.23, 3.20 and 3.71 % based on the model ECHAM4 in scenarios A2, B1 and A1B respectively .
