Precipitation Dominates the Relative Contributions of Climate Factors to Grasslands Spring Phenology on the Tibetan Plateau

Temperature and precipitation are the primary regulators of vegetation phenology in temperate zones. However, the relative contributions of each factor and their underlying combined effect on vegetation phenology are much less clear, especially for the grassland of the Tibetan Plateau To quantify the contribution of each factor and the potential interactions, we conducted redundancy analysis for grasslands spring phenology on the Tibetan Plateau during 2000-2017. Generally, the individual contribution of temperature and precipitation to grasslands spring phenology (the start of growing season (SOS)) was lower, despite a higher correlation coefficient, which further implied that these factors interact to affect the SOS. The contributions of temperature and precipitation to the grasslands spring phenology varied across space on the Tibetan Plateau, and these spatial heterogeneities can be mainly explained by the spatial gradient of long-term average precipitation during spring over 2000-2017. Specifically, the SOS for meadow was dominated by the mean temperature in spring (T-spring) in the eastern wetter ecoregion, with an individual contribution of 24.16% (p < 0.05), while it was strongly negatively correlated with the accumulated precipitation in spring (P-spring) in the western drier ecoregion. Spatially, a 10 mm increase in long-term average precipitation in spring resulted in an increase in the contribution of T-spring of 2.0% (p < 0.1) for meadow, while it caused a decrease in the contribution of P-spring of -0.3% (p < 0.05). Similarly, a higher contribution of P-spring for steppe was found in drier ecoregions. A spatial decrease in precipitation of 10 mm increased the contribution of P-spring of 1.4% (p < 0.05). Considering these impacts of precipitation on the relative contribution of warming and precipitation to the SOS, projected climate change would have a stronger impact on advancing SOS in a relatively moist environment compared to that of drier areas. Hence, these quantitative interactions and contributions must be included in current ecosystem models, mostly driven by indicators with the direct and the overall effect in response to projected climate warming.