Response of water-use efficiency to phenology in the natural forest and grassland of the Loess Plateau in China

Ecosystem water use efficiency(WUE) is an integrated physiological metric for the coupling cycle between terrestrial carbon, water, and energy. How WUE responds to vegetation phenology(e.g., SOS, EOS-start, end of growing season,and GSL-growing season length) shifting in temperate semi-arid regions is a hot spot in relative research fields. Based on remotesensing products and in-situ measured climate data, this study discussed how gross primary productivity(GPP), evapotranspiration(ET), and WUE(quantified by GPP/ET) would change with the altering vegetation phenology and climate in the untouched semi-arid forests and grasslands of the Chinese Loess Plateau during 2001–2020. Our results show that vegetation tended to green-up earlier and brown-down later from 2001 to 2020, causing an extended GSL. The forests had an earlier SOS,later EOS, and longer GSL than the grasslands, but the latter had a bigger variation amplitude. The WUE in the study area decreased significantly during spring and summer, while the grassland WUE increased in autumn; the annual mean reduction rate in grassland WUE was approximately twice that of woodland. Earlier SOS could increase forest WUE but reduce grassland WUE in spring, mainly because leaf unfolding has a more pronounced limitation on soil evaporation beneath the forest canopy.EOS had less impact on WUE, and no apparent difference existed between these two ecosystems. Climate change could affect WUE directly by changing GPP and ET and indirectly by regulating vegetation phenology. Warming can increase GPP and ET,causing an earlier SOS, further promoting GPP and ET(except forest ET). Precipitation significantly affected forest GPP and ET in spring, grassland GPP and ET in summer, and grassland ET in autumn; precipitation affects spring grassland WUE mainly via regulating SOS. Enhanced solar radiation could suppress grassland GPP in spring, promote forest ET in autumn, and regulate grassland WUE by affecting phenology. This study is meaningful for improving the process-based vegetation model and studying arid and semi-arid ecosystems’ responses to a changing climate.