Interpretation of vegetation phenology changes using daytime and night-time temperatures across the Yellow River Basin, China.

Vegetation phenology is highly sensitive to climate change and has a crucial effect on the carbon balance. Prior studies have mainly investigated the effects of mean temperature and precipitation on phenology. The asymmetric and opposing response of phenology to daytime and night-time temperature remains largely unknown. Using the satellite phenology derived from GIMMS NDVI3g datasets dating back to the 1980s, we show that significantly advanced start of the season (SOS), delayed end of the season (EOS) and prolonged length of growing season (LEN) (P < 0.05) has been taking place in the Yellow River Basin in China. The extension of LEN was more attributed to the advance of SOS than a delayed EOS. The daytime Tmax and night-time Tmin had opposite effects on the timing of SOS, MOS, and EOS in 63.1%,40.0%, and 53.5% of the pixels of the study area, respectively. If higher Tmax leads to an earlier or later transition date, an increase in Tmin systematically leads to an opposite effect. These opposite effects were obvious in SOS of 70.5%, 66.2%, and 70.6% of shrubland, grassland, and crop fields, respectively. For EOS, the opposing effects accounting for 58.2%,60.2%, and 54.5% of forest land, shrubland, and grassland, respectively. These results reveal different impacts of climate change on the daytime and night-time carbon cycle in terrestrial ecosystems, and such impacts vary with the land surface type. Knowledge of these opposing responses of phenology to daytime and night-time warming may help to understand the feedback of terrestrial ecosystem structure and function to climate change, thus to improve the existing terrestrial ecosystem carbon cycle model, which is of great significance for climate change and ecology research.