Surface soil phytolith assemblages across an altitudinal transect in the Qilian Mountains of Northwestern China, and their implications for palaeoecologic analysis in arid alpine regions.

Reconstructing changes in altitudinal vegetation belts is crucial for comprehending vegetation response to past environmental changes in arid alpine regions, however, proxy data are limited. In this study, we report 51 surface soil phytolith samples from various vegetation types across an altitudinal transect in the arid alpine region of the Qilian Mountains, to elucidate the relationships between vegetation types and phytolith assemblages. Our results show that phytolith morphotypes in surface soil samples could be categorized into 16 classes, with Gobbet, C renate , R ondel , E longate entire, and A cute bulbosus being the predominant classes. The characteristics of surface soil phytolith assemblages vary across an altitudinal transect, reflecting differences in local vegetation. Discriminant analysis identifies Cubic, E longate entire , C renate , Gobbet, and R ondel as discriminant variables that correlated with five vegetation types (desert vegetation, mountain steppe vegetation, mountain coniferous forest vegetation, subalpine shrub meadow vegetation, and alpine meadow vegetation). We establish five discriminant functions based on the relationships between vegetation types and phytolith assemblages, demonstrating a high accuracy of our method for distinguishing between different vegetation types. These discriminant functions are applicable for distinguishing between diverse vegetation types in the mountains of arid inland regions of Asia. They are also effective in identifying coniferous forest vegetation in mid-latitude mountains. Therefore, these discriminant functions hold significant potential for reconstructing changes in altitudinal vegetation belts in mountainous regions and determining the vegetative types recorded by terminal lakes in arid regions. Our study introduces a new perspective on palaeoecologic analysis in arid alpine regions, advancing our understanding of ecosystem dynamics in response to environmental fluctuations. • Phytolith assemblages can reveal vegetation types across an altitudinal transect. • Five discriminant functions are established for arid mountain vegetation. • The study shows potential for palaeoecologic analysis in arid alpine regions. [ABSTRACT FROM AUTHOR]