Ecological restoration largely alleviates livestock grazing pressure in a montane grassland

Abstract

Over the past few decades, grassland productivity in mountainous regions has been increasingly affected by livestock grazing and land use/cover changes. Human appropriation of net primary production (HANPP) is a valuable metric for assessing human impacts on grasslands. To date, knowledge of the patterns and drivers of HANPP in montane grasslands remains limited. In this study, we developed a 1 km × 1 km resolution HANPP dataset for the grasslands of the Qilian Mountains from 2000 to 2020 using two process-based models. We aimed to examine the spatial distribution characteristics and temporal trends of HANPP and explore the effects of climatic conditions, rural populations, cultivated areas, and livestock numbers on the interannual variability of HANPP. The results showed that HANPP decreased significantly with increasing elevation due to the corresponding decrease in population density. During the study period, HANPP showed a widespread decline, which was significant in 15.4% of the grassland area, mainly in alpine meadows. Climatic factors and human activities explained 60.8% and 39.2% of the temporal variability in HANPP, respectively. In particular, precipitation was negatively correlated with HANPP, whereas air temperature and number of livestock were positively correlated with HANPP based on correlation analysis. By providing a refined spatial assessment of human impacts on montane grasslands, this study suggests that ecological restoration projects, especially the Grain to Green and Grazing Withdrawal Programs, can reduce human pressures on grassland productivity by controlling population density and grazing intensity. Furthermore, the adoption of enclosed paddocks and supplemental feeding may mitigate montane grassland degradation near villages. This study elucidates the primary factors influencing the spatiotemporal patterns of HANPP in an area with typical montane grasslands, thereby supporting research-based decision making.