Divergent responses of ecosystem water use efficiency to environmental drivers in arid versus humid regions.
Ecosystem water use efficiency (EWUE), defined as carbon gain per unit water consumed, is increasingly used as an integrative indicator to support land and water management under climate change. However, whether long-term EWUE trends and their dominant drivers provide consistent guidance for ecosystem restoration and management across contrasting hydroclimatic regions remains unclear. Using global datasets from 2000 to 2023, we found that global EWUE increased by 13.3%, but with strong spatial heterogeneity. Hyper-arid regions exhibited a pronounced increase (15.9%), whereas humid regions showed negligible change (0.59%). Attribution analysis reveals that the global EWUE increase was mainly associated with improvements in soil water use efficiency (SWUE), defined here as the ratio of gross primary productivity to soil evaporation, which contributed 64.6% to the observed trend. Notably, arid ecosystems exhibited a temporal shift in EWUE sensitivity around 2015, characterized by weakened soil-moisture sensitivity and a marked increase in atmospheric CO2 sensitivity. This transition suggests that recent efficiency gains in water-limited regions increasingly rely on CO2 fertilization rather than hydrological regulation, potentially increasing the exposure of restoration investments to extreme drought risk. In contrast, humid ecosystems showed a "double suppression" pattern, in which limited GPP enhancement and persistent water losses from both soil evaporation and plant transpiration jointly constrained further EWUE increases. These findings demonstrate that EWUE trends cannot be directly translated into uniform management strategies. We suggest that effective management in arid regions should consider physical evaporation suppression (e.g., via ground cover), whereas humid forest strategies may benefit from stand structure optimization to mitigate saturation effects. Such differentiation provides a reliable basis for designing region-specific restoration policies and avoiding maladaptive decisions under increasing climate variability.