Study region: Inland river basins in arid northwestern China (specifically the Tarim and Heihe Basins), characterized by complex rain-snow-ice mixed recharge systems and high sensitivity to climate warming. Study focus: Using monthly data (1961-2020) from six representative basins, we assess how rain, snow, and ice jointly reshape streamflow seasonality. We quantify the sensitivity of streamflow to precipitation, snowmelt, and glacial runoff to reveal the mechanisms driving seasonal redistribution in these multi-source basins. New hydrological insights for the region: Global warming is reshaping hydrological regimes toward seasonality flattening. Specifically, the streamflow ratio in late spring and early summer decreases by 3.46%, while autumn and winter ratios increase by 30.55%, narrowing the seasonal difference. Accelerated warming appears to precipitate the earlier onset of spring snowmelt. Conversely, the notable rise in autumn and winter streamflow is consistent with increases in precipitation (autumn: 33.2%, winter: 32.89%) and an extended glacial ablation period. Sensitivity analysis shows that streamflow is more sensitive to snowmelt (0.41) than to precipitation (0.09). Crucially, increased snowmelt exerts a buffering effect, reducing streamflow sensitivity to precipitation. Additionally, streamflow sensitivity to glacier melt intensifies with elevation. Advancing beyond single-driver frameworks, these findings demonstrate that the buffering role of snowmelt and the elevation-dependent intensification of glacier sensitivity are key mechanisms governing how multi-source recharge systems respond to warming, providing a transferable diagnostic framework for similarly complex basins.