Photosynthesis (GPP), the largest CO 2 flux from the land surface, is currently estimated with considerable un- certainty. More robust estimates of global GPP could be obtained from the atmospheric budgets of other tracers such as, the oxygen isotopic composition ( δ 18 O) of atmospheric CO 2 or carbonyl sulphide (COS). However, es- timating GPP using these tracers hinges on a better understanding of how soil microbes modify the atmospheric concentrations of CO 18 O and COS at large scales. In particular, understanding better the role and activity of the enzyme Carbonic Anhydrase (CA) in soil microbes is a critical factor underpinning the successful implementation of these tracers in global scale models. We addressed this knowledge gap by measuring the exchange of CO 18 O between soil microcosms and the atmosphere from sites covering a range of biomes. This novel dataset has led to the development of a new mechanistic framework that can be easily implemented in multi-tracer Earth system models to predict variations in soil CA activity across the terrestrial land surface. Using this multi-tracer approach we provide independent estimates of global GPP constrained by the atmospheric budgets of CO 2 and CO 18 O.