Abstract: Precipitation use efficiency (PUE) is an effective index to evaluate the relationship between grassland productivity and precipitation in arid and semi-arid regions. To explore the driving mechanism of climate change and human activities on grassland PUE, we used the improved CASA model to estimate net primary productivity (NPP) of grassland from 2001 to 2020 in Otog Banner, Inner Mongolia. The PUE was obtained combining with the spatial interpolation data of precipitation. The spatiotemporal evolution of PUE and its responses to the six climate factors were analyzed using simple and piecewise linear regression. A quantitative analysis method based on partial derivatives was used to quantitatively evaluate the relative contributions of climate change and human activities to PUE dynamics. The results showed that the annual average value of PUE was 0.748 g C·m^-2·mm^-1 in Otog Banner, and that the inter-annual fluctuation had a significant downward trend at a rate of 0.014 g C·m^-2·mm^-1·a^-1. Across the study area, PUE increased from west part to east part, and exhibited significant single-peak piecewise linear patterns along the growth gradients of temperature, precipitation, relative humidity, sunshine hours, and ET₀. There was a sustained and significant increase pattern of fast first and then slow along the wind speed gradient. 94.3% of the grassland in the study area showed a decrease trend in PUE, and 43.6% area showed severely decreased. This prominent decrease in PUE was co-driven by climate change and human activities, whose contributions were -1.162×10^-2 and -0.240×10^-2 g C·m^-2·mm^-1·a^-1, respectively. Climate change was the primary driving force and precipitation was the key climate driving factor for the decrease in PUE.

Key words: precipitation use efficiency, spatiotemporal evolution, climate change, human activity