Climate and human impact together drive changes in ecosystem multifunctionality in the drylands of China

China's drylands encompass an area of approximately 6.60 x 106 km2 and are home to around 580 million people, providing crucial ecosystem services. The escalating impacts of global climate change are expected to exacerbate aridity, thereby adversely affecting ecosystem services and human well-being. Hence, there is a crucial need to prioritize the study of ecosystem multifunctionality (EMF) in drylands and investigate the underlying factors driving EMF changes. We collected field sampling and indoor analysis data, as well as raster level sampling data to obtain indices related to soil nutrients, vegetation, human footprint index (HFI), soil biodiversity index (SBI), and climate (mean annual temperature [MAT], mean annual precipitation [MAP], potential evapotranspiration [PET]), and soil characteristics (soil sand content [Sand], soil moisture [SM], soil pH). EMF was calculated using the averaging approach, and the cluster-multi-threshold approach. We employed the mantel test, mixed-effects model, random forest algorithm, piecewise structural equation model, and variation partition analysis to explore the significance of factors on EMF changes. The findings revealed significant correlations between EMF and all factors except pH and Sand. All factors collectively accounted for 52.7 %-57.4 % of the variation in mono- (the effect on each of the carbon, nitrogen and phosphorus functions separately) and multifunctionality. The combined effects of all factors had the greatest contribution (8.1 %-16.3 %) variation in EMF. HFI consistently emerged as the most important factor in controlling EMF variation and proved to be the most significant predictor of EMF changes. All factors examined were significant predictors of EMF change. Our study demonstrated that HFI played the most direct influence on EMF changes, surpassing the impact of any single variable. However, the overall contribution of climate (integrating MAP, MAT, and PET) to EMF change was also extremely important. Therefore, future assessments and predictions of ecological functions in drylands should consider not only the direct effects of climate change but also the integrated analysis of human impact.