Nutrient enrichment shifts peak water-use efficiency to wetter conditions in an alpine meadow

Ecosystem water-use efficiency, the ratio of gross ecosystem productivity to plant transpiration, is found to show a hump-shaped response to extreme water variations (with a peak at moderate water level). This peak water-use efficiency represents maximum carbon gain per unit of water consumed, with its associated water threshold indicating the optimal water level to maximize carbon uptake. However, it remains unclear whether widespread nutrient enrichment alters this peak water-use efficiency and its associated water threshold. Here, we conducted a 6-year field experiment involving precipitation reduction and multiple nutrient additions in an alpine meadow on the Qinghai-Tibet Plateau. A wide water gradient is created by combining experimental precipitation manipulation and interannual precipitation fluctuations. Our results showed a consistent hump-shaped response of ecosystem water-use efficiency to water variations in control, N, NP, and NPK treatments. While nutrient addition had little effect on peak water-use efficiency, it significantly increased the associated water threshold by 16.9-25.2% (from 540 mm to 631-676 mm). This increase was primarily caused by a U-shaped transpiration response, but not by the positive responses of gross ecosystem productivity to precipitation across all treatments. Specifically, nutrient addition elevated the transpiration water threshold (the water level for minimum transpiration) by 10.1-16.0% (from 406 mm to 447-471 mm). This U-shaped transpiration response to precipitation was mainly regulated by soil temperature in control plots, but by vertical root distribution in nutrient treatments. Synthesis. These new findings show that nutrient enrichment does not alter peak water-use efficiency but elevates its associated water threshold. This pattern indicates that nutrient-enriched ecosystems require more water than nutrient-poor ones to reach maximum carbon uptake, with important implications for predicting the dynamics of grassland carbon sink under the combined effects of soil water and nutrients.