Quantifying the impact of climate change and human activities on runoff changes is beneficial for developing sustainable water-management strategies within the local ecosystem. Machine-learning models were widely used in scientific research; yet, whether it is applicable for quantifying the contribution of climate change and human activities to runoff changes is not well understood. To provide a new pathway, we quantified the contribution of climate change and human activities to runoff changes using a machine-learning method (random forest model) in two semi-humid basins in this study. Results show that the random forest model provides good performances for runoff simulation; the contributions of climate change and human activities to runoff changes from 1982 to 2014 were found between 6–9% and 91–94% in the Zijinguan basin, and 31–44% and 56–69% in the Daomaguan basin, respectively. Furthermore, the model performances were also compared with those of well-known elasticity-based and double-mass curve methods, and the results of these models are approximate in the investigated basins, which implies that the random forest model has the potential for runoff simulation and for quantifying the impact of climate change and human activities on runoff changes. This study provides a new methodology for studying the impact of climate change and human activities on runoff changes, and the limited numbers of parameters make this methodology important for further applications to other basins elsewhere. Nevertheless, the physical interpretation should be made with caution and more comprehensive comparison work must be performed to assess the model’s applicability.

Efforts to slow down groundwater depletion requires reliable estimates of the impacts of irrigation on shallow groundwater levels. However, spatial assessment of the effects of irrigation schedule on groundwater level change is still lacking. Here, we present a linked APSIM-MODDLOW model framework to reveal irrigation-groundwater linkages in the plain area of Ziyahe River Basin by considering irrigation schedules, crop yield, ETa and leakage. The linked APSIM-MODLFOW model effectively simulates grain yield and regional groundwater level fluctuations in the winter wheat-summer maize cropping system of the NCP. We found that the simulated grain yield and evapotranspiration increase rapidly with the increase of irrigation frequency, whereas the yield exhibits a gradual increase when irrigation exceeds three times (70 mm each time). The simulated regional average groundwater levels exhibit an increase under rain-fed and single-irrigation scenarios-with irrigation applied at the jointing stage of wheat-whereas a decline is observed under alternative irrigation regimes. These findings indicate that regional groundwater can be replenished through recharge when irrigation schedules range from single-irrigation (applied at the jointing stage of wheat) to double-irrigation (applied at both the jointing stage and the flowering and grain-filling stage of wheat) The unconfined aquifer maintains extraction-recharge equilibrium under the irrigation amount 70 – 140 mm (1-irrigation to 2-irrigations). Spatial distribution of groundwater flow field revealed that to prevent a decline in regional groundwater levels and reduce grain yield loss, the existing 3-irrigations schedule should be maintained in the southern region of study area and Shijin Irrigation District, while a 1-irrigatoin schedule should be implemented in the northern region. These findings can serve as a reference for regional irrigation and groundwater resource evaluation and management. The linked model framework has important implications for other studies where groundwater variations are highly impacted by agriculture irrigation.