Migration of vegetation boundary between alpine steppe and meadow on a century-scale across the Tibetan Plateau

• Precipitation determines the vegetation distribution patterns of alpine meadow and steppe on Tibetan Plateau, with the highest relative importance of 41.26%. • Random Forest could be employed to classify the boundary of alpine meadow and steppe on Tibetan Plateau. • Vegetation boundary of alpine meadow and steppe may transfer from northeast to southwest under different greenhouse gas emissions scenarios in the future. The distribution of alpine vegetation is highly sensitive to climate change, which attracted the attention of climate scientists as well as ecologists. However, the dominant factors of vegetation distribution showed greatly spatiotemporal variation, especially the alpine grasslands on the Tibetan Plateau, which is known as an amplifier of climate warming. In this study, to identify the dominant factors of vegetation distribution, we verified the reliability and accuracy of the classification of alpine steppe and alpine meadow by Random Forest and tried to provide a new reference for the classification schemes. Dataset collected from the field investigations (200 sites) and previous publications(200 sites) were used in this study for model calibration and validation. Then climate models of CMIP6 were used to forecast the underlying transfers and changes of vegetation boundary in the future. Our results revealed that precipitation may be the dominant driver in the alpine grassland distribution, with the highest relative importance of 41.26%. Through different climate scenarios from 2000 to 2100, the vegetation boundary showed a shift from northeast to southwest. Within the SSP5-8.5 scenario it moves from 94.50° to 93.49° and shifted from north to south by 0.34°. Our results demonstrated the significant role of precipitation in the alpine grassland distribution and revealed the migration of the alpine grassland ecosystem under different climate change scenarios. Our work may be useful to deal with global climate change, early warning, and grassland protection, as well as adaptive management of grassland.