Unraveling the sensitivity and nonlinear response of water use efficiency to the water–energy balance and underlying surface condition in a semiarid basin

Ecological water use efficiency (WUE) is a crucial indicator of hydrological and vegetation carbon cycle coupling and can drastically affect water and energy exchanges. However, little is known about the sensitivity of WUE to water–energy exchange in semiarid regions. Moreover, few studies have considered the link between WUE and water–energy exchange models, i.e., the Budyko-type framework. Here, we investigated the relationships between WUE and a Budyko-type model in a semiarid basin. Seven primary indicators were selected to represent the water, energy, and vegetation variations in the ecosystem: gross primary productivity (GPP), air temperature (T), potential evapotranspiration (PET), evapotranspiration (ET), precipitation (P), soil water content (SWC), and snow water equivalent (SWE). Three similar ecological WUEs were defined as GPP/ET, GPP/P, and GPP/SWC to analyze the factors of the water–energy exchange model (i.e., ET/P, PET/P, SWC/P, and SWE/P). Based on the results, four linear Budyko-type models were proposed for the basin (GPP/P, ET/P, SWC/P, and SWE/P as a function of PET/P). The results suggested that both SWC and P influenced the trend of GPP variation, whereas P influenced the lower limit of GPP or GPP/P within the Budyko model for grassland areas. The results indicated that the rate of increase of SWC was lower than that of P in forest areas because of differences in canopy structure. The results also revealed a nonlinear (s-type) relationship between the WUEs and the underlying surface parameter m within the Budyko framework, suggesting that unit plant productivity consumes less water when the water–energy supply condition is invariant, if the variation of the underlying surface characteristics promotes the increase of m. Our research provides new insight regarding quantification of the sensitivity of WUE to the water–energy balance in a semiarid region.