The vegetation dynamic (e.g., community productivity) is an important index used to evaluate the ecosystem function of grassland ecosystem. However, the critical factors that affect vegetation biomass are disputed continuously, and most of the debates focus on mean annual precipitation (MAP) or temperature (MAT). This article integrated these two factors, used the aridity index (AI) to describe the dynamics of MAP and MAT, and tested the hypothesis that vegetation traits are influenced primarily by the AI. We sampled 275 plots at 55 sites (five plots at each site, including alpine steppe and meadow) across an alpine grassland of the northern Tibet Plateau, used correlation analysis and redundancy analysis (RDA) to explore which key factors determine the biomass dynamic, and explained the mechanism by which they affect the vegetation biomass in different vegetation types via structural equation modelling (SEM). The results supported our hypothesis, in all of the environmental factors collected, the AI made the greatest contribution to biomass variations in RDA , and the correlation between the AI and biomass was the largest (R = 0.85, p < 0.05). The final SEM also validated our hypothesis that the AI explained 79.3% and 84.4% of the biomass variations in the alpine steppe and the meadow, respectively. Furthermore, we found that the soils with higher carbon to nitrogen ratio and soil total nitrogen had larger biomass, whereas soil organic carbon had a negative effect on biomass in alpine steppe; however, opposite effects of soil factors on biomass were observed in an alpine meadow. The findings demonstrated that the AI was the most critical factor affecting biomass in the alpine grasslands, and different reaction mechanisms of biomass response to the AI existed in the alpine steppe and alpine meadow.
Keywords: Aboveground biomass; Alpine grassland; Aridity index; Coverage; Northern Tibet Plateau; Species richness.