Investigation of Climate Change impacts on Runoff by Decision tree model (Case Study: Zaringol basin)

In order to investigate the effect of climate change on runoff in Zaringol catchment in Golestan province, the rainfall-runoff relation was modeled using the M5 decision tree model. Moreover, the results of two climate change scenarios -- RCP 4.5 and RCP 8.5 -- under the HADGEM2 general circulation model were used to predict runoff for the near, middle and distant future. Based on Mann-Kendall analysis for 1995-1995, there were noticeable increases only in summer precipitations as well as spring and summer temperatures. However, as temperature increased, a concomitant reduction occurred for most seasonal as well as the average annual discharge values. In addition, Lars provided sufficient accuracy for precipitation and temperature data generation. We found that precipitation changes do not follow a specific increasing or decreasing pattern. In contrast, the air temperature will increase under both climatic scenarios, as in the case of an average of 0.5 to 5 ° C for the annual temperature, which is the largest surge in temperature under the RCP 8.8 scenario in the distant future. The results of M5 and MLR models showed that M5 model can improve the correlation coefficient from 0.7 m^3⁄s to about 0.87 m^3⁄s and the corresponding RMSE value from 0.82 m^3⁄s to 0.59 m^3⁄s According to the M5 model, there will be a runoff reduction in the study area in future periods. The largest decrease will occur in the distant future by approximately 41% in March for the RCP 8.5 scenario.