Objective The sediment in the Yellow River primarily originates from the loess hilly-gully area and the loess tableland-gully area, where significant differences exist in sediment reduction mechanisms and effectiveness. Currently, the lack of applicable sediment yield models and future scenario assessment methods for the loess tableland-gully area hinders the precise management and effectiveness evaluation of soil and water loss in this area.

Methods The 20 hydrological station control units concentrated on the loess tableland-gully area were selected as the study area. Based on remote sensing images and DEM data, the tablelands were identified using the human-computer interactive method in ArcGIS. The fragmentation index (M_eff) was introduced to evaluate the degree of tableland fragmentation, and the relationship between M_eff and sediment yield was analyzed. Based on the comparative relationship between the actual sediment yield and the sediment yield calculated by the model for the loess hilly-gully area, M_eff was incorporated to construct a sediment yield model for the loess tableland-gully area. Future sediment yield changes were then predicted and the sediment yield situation was evaluated using future management scenarios.

Results 1) The total number of tablelands in the loess tableland-gully area was 3 524, covering an area of about 8 120.26 km². Tablelands were classified based on the M_eff value. Those of M_eff ≥ 100 km² were considered as intact tablelands, those of 50 km² ≤ M_eff < 100 km² were considered as slightly fragmented tablelands, and those of M_eff < 50 km² were considered as fragmented tablelands. 2) M_eff was highly negatively correlated with the actual measured sediment yield and the sediment yield calculated using sediment yield model for the loess hilly-gully area, with a coefficient of determination R² of 0.96. Based on this, a sediment yield model for the loess tableland-gully area was established. Model predictions indicated that for tablelands with M_eff ≥ 50 km², sediment yields in 2025, 2030, 2035, and 2050 will decrease by 8.38%, 14.66%, 19.85%, and 31.44%, respectively, compared to 2018. However, if the same scale terraced fields and forests and grasslands were deployed in the loess hilly-gully area, the reduction in sediment yield will be as high as 13.20%, 22.30%, 29.65%, and 45.08%, respectively, indicating that the sediment reduction effectiveness of terraced fields and vegetation measures in the tableland area will be significantly lower than that in the hilly area.

Conclusions The sediment yield situation in the loess tableland-gully area can be effectively assessed using a model incorporating M_eff, though the sediment reduction potential of terraced fields and vegetation measures remains limited. Future management should shift toward tableland protection, controlling concentrated runoff into gullies, and efficient utilization of rainfall and flood resources.