Estimating the contribution of climate change and human activities on river discharge variations (Case Study: GharehSoo River)

Introduction
Investigating runoff changes in a watershed can help to understand better understand the factors affecting it. Climate changes and human activities in recent years have caused a decrease in runoff in different parts of the globe and have created social and economic problems. In general, the influencing factors on runoff changes can be physical factors (vegetation cover, initial soil moisture, land topography, etc.), climatic factors (precipitation amount, air temperature, earth warming, etc.), and changes caused by human activities (building a dam, building a reservoir, expanding urbanization, illegal withdrawals, etc.). The increase in greenhouse gases and climate change has caused changes in the hydrological cycle and the amount of runoff in watersheds and has increased the frequency of climate extreme events. Also, observations in most regions around the world show, that the hydrological cycle has also been affected by human activities. Human activities, such as agricultural development, urban development, dam construction, and exploitation of reservoirs, have direct and indirect effects on the hydrological cycle, and as a result, the temporal-spatial distribution of water resources has changed. The primary purpose of this study is to determine the contribution of each of these factors to the discharge changes of the GharehSoo River, one of the most important rivers of Ardabil province, using different classical and intelligent methods.
 
Materials and Methods
In this research, some classical and intelligent methods namely, linear regression, bivariate linear regression, double mass curve, and artificial neural network methods are used to determine the contribution of climate changes and human activities on the discharge change of GharehSoo River. First, by using Pettitt's test, the change point of the discharge time series is detected and divided into two natural and changes periods. Then, the contribution of each of these factors is determined using the mentioned methods.
 
Results and Discussion
In general, it can be said that the amount of runoff calculation error is almost the same for all the applied methods, and therefore the methods have relatively similar performance. However, in Samyan station, the two-variable linear regression model shows less error and the single-variable linear regression model shows more error than the other methods. For the Dost-Beiglo station, the two-variable linear regression model shows less error and the artificial neural network model shows more error than the other methods. The reason for the not so small error of the artificial neural network in predicting the runoff can be related to the error in the data used and the relatively short length of the data. In general, the results of different methods in both stations showed that based on the calculation error, the bivariate linear regression method provided better results in modeling the river discharge in both hydrometric stations. The results of this research showed that for the Samyan station, the contribution of the climate change using the linear regression, bivariate regression, double mass curve and artificial neural network is 6.45%, 14.42%, 14.86% and 8.61%; and the contribution of human activities is 93.55%, 85.58%, 85.14% and 91.38%, respectively. For the Dost-Beiglo station, the contribution of climate change for the mentioned methods is 2.1%, 3%, 27% and 0.14%; and the contribution of human activities is 97.9%, 97%, 73% and 99.86% respectively. By comparing the results of Samyan and Dost-Beiglo stations, it can be concluded that the effect of climate change on the discharge of Gharehsoo River at the Samyan station (11.08%) is more than the Dost-Beiglo station (8.06%) and on the contrary, the impact of human activities on the river flow at the Dost-Beiglo station (91.94%) is more than the Samyan station (88.91%), which can be due to the simultaneous effect of the construction of two dams including Yamchi and Sabalan in spstream of the Dost-Beiglo station. Also, as expected, the contribution of climate change (less than 27%) is less than the contribution of human activities (more than 73%) in reducing the flows of Gharehsoo River in both studied stations.
 
Conclusion
In this research, different hydrometeorological data such as precipitation, evaporation and transpiration and monthly discharge from the Samyan and Dost Beiglo stations were used for the statistical period of 1982-2019. First, by using Pettitt's test, it was determined that the river flow rate has changed abruptly since 2016. Therefore, the entire statistical period was divided into two natural and change periods, and then, using the mentioned methods, the contribution of human activities and the contribution of climate change were determined.
Two climatic factors, i.e. decrease in rainfall and increase in evapotranspiration in climate change and carrying out activities such as the construction of Yamchi and Sabalan Dams, development of orchards and agricultural lands as human activities have been effective in reducing the flow of the Gharehsoo River. However, human activities have had a greater impact (over 73 %) than the climate change factor (less than 27 %) in reducing the flow of this river.
Finally, in future studies, it is suggested to use other intelligent and hydrological models of runoff estimation for rivers in the country and to evaluate their efficiency in determining the contribution of climatic and human effects on river flow. Also, other climate variables such as temperature, wind, etc. should be used in determining the contribution of climate change effects.