Vapor pressure deficit (VPD) and soil moisture (SM) are the primary atmospheric and edaphic water stress indicators for terrestrial ecosystem gross primary productivity (GPP). While the individual effects of VPD and SM on GPP have been widely discussed, their respective contributions to GPP in the ecosystem remains debating due to their complicate interactions. Most of precious studies focus on coupling between VPD and SM and their linear effects on GPP, neglecting the interactions among VPD, SM and other environmental variables, such as air temperature (Ta) and photosynthetically active radiation (PAR), as well as the indirect and nonlinear effects through these variables. To investigate the independent effects and the relative roles of VPD and SM on GPP, approaches of quantile binning, ridge regression, and structural equation modeling (SEM) are applied to quantify the direct, indirect and total effects of VPD and SM on GPP, as well as their respective roles played in affecting ecosystem GPP. We have chosen the grasslands of the semi-arid region in northern China as the study area, characterized by strong land-atmosphere coupling and frequent droughts. The result indicates that VPD exerts a negative effect on GPP in the grassland ecosystem under all SM conditions, whereas SM suppressed GPP under high VPD conditions. The negative effect of SM is largely attributed to the indirect effect of SM through Ta. SM-Ta coupling and evapotranspiration (ET) play a significant role in GPP responses to SM. In the region where ET is low and water-limited, SM influences GPP positively; in the region where ET is relatively high and energy-limited, SM has a negative effect on GPP, due to the negative SM-Ta coupling. Ta and VPD dominate the variation of GPP in the semi-arid grassland ecosystem of northern China, while SM mainly attributed to the variability of GPP through land-atmospheric coupling.
Guo et al. (Sat,) studied this question.