Abstract: Due to the double impacts of climate changes and human activities in recent years, runoff and sediment transport in the Yellow have been reduced sharply, so it is necessary to make a reliable judgment on the future trends of these changes. Based on the method of feature selection and multiple regression analysis, this paper predicts the future runoff and sediment conditions of the river, using the sediment transport and runoff data measured at the Tongguan station, the mesoscale data from the European medium-range weather forecast center, and the global climate model data of CO2 emission concentrations along different climate change paths. The results show that under the scenario RCP26 in the next 10, 20 and 50 years, the runoff at the Tongguan station will be 2.34×1010 m3, 2.28×1010 m3 and 2.2×1010 m3 respectively, increased by 2.3%, -0.5% and -4.0% in comparison with the average of 2.29×1010 m3 over 2000-2016; the sediment discharge will be 2.68×108 t, 3.44×108 t and 5.72×108 t, increased by 7.9%, 38.4% and 130.2% respectively against the 2000-2016 average of 2.48×108 t. This indicates the less water and more sand feature of the Yellow basin will remain in place for quite a long period in the future.

Key words: Yellow River, runoff and sediment prediction, multiple regression analysis, data-driven