Spatiotemporal Dynamics and Environmental Associations of Vegetation Carbon Sinks in the Middle and Lower Yellow River Basin, China

Vegetation carbon sinks are an important component of the terrestrial carbon cycle, and net ecosystem productivity (NEP) is widely used to indicate ecosystem carbon-sink strength. However, the long-term dynamics and environmental associations of vegetation carbon sinks in the middle and lower Yellow River Basin remain insufficiently understood. Based on remote-sensing net primary productivity (NPP) and an empirical heterotrophic-respiration model, annual NEP was estimated for 2001–2024. Theil–Sen trend analysis, the Mann–Kendall test, coefficient of variation, optimal-parameter geographical detector (OPGD), and regression residual analysis were applied. Basin-mean annual NEP increased from approximately 214 to 425 g C m−2 yr−1, with significantly increasing areas accounting for 89.93% of the study area. Areas with NEP above 300 g C m−2 yr−1 expanded from 18.51% to 78.86%. Precipitation and solar radiation showed the highest explanatory power for NEP spatial differentiation, with mean q values of 0.538 and 0.490, and their interaction reached 0.722. Comparison with a published NEP product, parameter-sensitivity analysis, and an NPP-based robustness test supported the main temporal patterns and factor rankings. The residual-derived non-climatic component exceeded 40% of the combined component-trend magnitude across approximately 94% of the study area, but should not be interpreted as a direct measure of human activities. These findings support regional carbon-sink monitoring, water-constrained ecological restoration, and land-use management.