Variation of vegetation autumn phenology and its climatic drivers in temperate grasslands of China

Understanding the variation of autumn phenology and its climatic drivers is important for predicting terrestrial carbon cycles in the temperate grasslands of China. Using meteorological data and GIMMS NDVI data during 1982–2015, this study analyzed the variations in end date of vegetation growing season (EOS) and their relationships with climate variations in temperate grasslands of China. The results showed that EOS was delayed by 1.62 days/decade across temperate grasslands of China. For different grassland vegetation types, the EOS was delayed by 1.65, 1.66, and 1.34 days/decade for temperate meadows, steppes, and desert steppes, respectively. In terms of climate change effects, increasing summer precipitation and autumn temperatures is crucial for delaying the EOS of temperate grasslands in China. The increase in summer precipitation could delay the EOS, especially for temperate desert steppes, whereas the increase in autumn temperatures could significantly delay the EOS, especially for temperate meadows. In addition, we found that the influences of nighttime and daytime warming on the EOS were asymmetric. Specifically, the increase in autumn maximum temperature could significantly delay the EOS of temperate meadows and steppes, and increasing minimum temperature in summer and autumn could significantly delay the EOS of temperate meadows. For temperate steppes and desert steppes, increasing maximum and minimum temperature during summer had a weakly advancing and delaying effect on the EOS, respectively. Our study highlights the distinct influences of monthly climatic change on the EOS of different grassland vegetation types and indicates that the impacts of nighttime and daytime temperature should be included in simulating the EOS of grassland ecosystems in arid/semi-arid regions.