Abstract
Ecosystem gross primary productivity (GPP) is the largest carbon flux between the atmosphere and biosphere and is strongly influenced by soil moisture. However, the response and acclimation of GPP to soil moisture remain poorly understood, leading to large uncertainties in characterizing the impact of soil moisture on GPP in Earth system models. Here we analyze the GPP-soil moisture response curves at 143 sites from the global FLUXNET. We find that GPP at 108 sites exhibits hump-shaped response curves with increasing soil moisture, and an apparent optimum soil moisture ( SMoptGPP , at which GPP reaches the maximum) exists widely with large variability among sites and biomes around the globe. Variation in SMoptGPP is mostly explained by local water availability, with drier ecosystems having lower SMoptGPP than wetter ecosystems, reflecting the water acclimation of SMoptGPP . This acclimation is further supported by a field experiment that only manipulates water and keeps other factors constant, which shows a downward shift in SMoptGPP after long-term water deficit, and thus a lower soil water requirement to maximize GPP. These results provide compelling evidence for the widespread SMoptGPP and its acclimation, shedding new light on understanding and predicting carbon-climate feedbacks.