A remote sensing evidence on the marginality, stagementation and spatiotemporal heterogeneity of vegetation evolution characteristics in the Yinshan Mountains, China: Based on PKU GIMMS NDVI (1984–2022)

• Analysis of NDVI and climate factors dynamics on multiple spatiotemporal scales. • NDVI, is “stable-plunge-rise” temporally; declines southeast to northwest spatially. • Significant positive spatial clustering between NDVI and precipitation. • Similar phase patterns observed between vegetation growth and human activities. This study focuses on a typical Ecological Vulnerable Area (EVA)-Yinshan mountains (YSMs), using PKU GIMMS NDVI as a vegetation growth indicator, to explore the spatiotemporal heterogeneity and driving mechanisms of vegetation in in desertified steppe ecosystems (DSEs) of arid and semi-arid regions. The Theil-Sen (T-S) Median trend analysis & Mann-Kendall (M-K) trend test, Hurst index, correlation analysis, partial correlation analysis, residual analysis and bivariate spatial autocorrelation (Bi-SA) analysis were used for quantitative analysis. Findings revealed: (1) Vegetation distribution had spatial heterogeneity, with a stepwise distribution from southeast to northwest; (2) Vegetation growth had temporal stagementation, with “stable-plunge-rise” pattern; (3) Climate trend displayed “warmer and wetter”, with growth rates of 0.045 °C/a and 0.558 mm/a, respectively; (4) Vegetation response to precipitation was sensitive with strong spatial correlation ( Moran’s I = 0.88 , P < 0.01 ); (5) Influence of human activities on the growth of vegetation was weak. The study emphasizes that the overall ecological quality of the YSMs has long been low. In recent decades, despite some improvements, the ecosystem remains extremely sensitive, fragile and volatile. Consequently, there is an urgent need to strengthen eco-environmental protection and management.