Abstract:

The aim of this research was gain deeper insights into the spatio-temporal evolution of land use/cover change （LUCC） and its driving mechanisms in Qilian Mountain National Park， an alpine ecosystem. Based on multi-temporal remote sensing data， this study constructed maps showing where land use/cover has remained constant， expanded， and contracted， and systematically analyzed the evolution of LUCC and its driving factors during 1990-2022. The results show that： 1） The LUCC trajectory maps effectively captured the consistency among typical time nodes （1995， 1997， 2004， 2008， and 2019） and regional ecological projects. 2） The LUCC trajectory maps revealed differences in LUCC across periods. Specifically， the constancy map indicates that grassland （>48.00%） and bare land （>25.00%） are primarily distributed in patches around the Daxue and Zewujie peaks. The expansion map shows that grassland expansion has been concentrated around the Daxue and Lenglongling peaks， while bare land expansion has mainly occurred around the Qilian and Daxue peaks， although the magnitude of expansion has been lower for bare land than for grassland （by 2.75% overall）. Forest， glacier/snow， and water expansion have been relatively minor （<3.00%）. The contraction map shows that grassland （>1.40%） and bare land （≥1.09%） areas have decreased more than other land types （>0.01%）， with areas of contraction patchily distributed between the Qilian and Lenglongling peaks and scattered across the entire Qilian range. 3） The LUCC trajectory maps further reveal a southeast-northwest gradient in land use intensity， which has been mainly affected by natural factors such as potential evapotranspiration， elevation， and human footprint （>0.09）， whereas the effects of economic development and policy interventions have become increasingly evident over time. The results of this study demonstrate a novel trajectory-based approach for dynamic LUCC identification that distinguishes stable areas from changing areas and quantifies their dynamics， thereby improving the precision and depth of land-system dynamics representation. The findings provide a scientific reference and decision-making support for alpine ecosystem management.

Key words: temporal segmentation, change atlas, integrated land use degree, geo-detector, Qilian Mountain National Park