Aim Non-structural carbohydrates (NSC), primary products of photosynthesis, play a key role in plant responses to environmental stress and, in turn, modulate photosynthesis under changing climates. This study aims to evaluate how land surface models represent the bi-directional relationship between NSC and photosynthesis, and to quantify the model-emergent causality between NSC and gross primary productivity (GPP) at the global scale.Location Globe.Time Period From 2000 to 2016.Major Taxa Studied Plants.Methods We assessed the causal interactions between NSC and GPP using the Convergent Cross-Mapping (CCM) technique applied to outputs from the latest Community Land Model versions (CLM5-BGC and CLM5-FATES).Results Our results reveal a pervasive bi-directional causal relationship between GPP and NSC, with significant coupling across 88.14% of vegetated land. Surprisingly, NSC exerts a more substantial influence on GPP than vice versa in more than 73.06% of the land area. This NSC-dominant control is particularly evident in arid and semi-arid ecosystems.Main Conclusions Our findings underscore the critical role of NSC dynamics in modulating photosynthesis in land surface models and highlight the need for improved representation and observational constraints of the multiple processes linked to NSC, such as carbon allocation, phenology, mortality and canopy structure.