Predicting photosynthetic capacity from first principles

Background/Question/Methods

Photosynthesis by plants represents one of the largest uncertainties in projecting the rate and magnitude of future atmospheric carbon dioxide (CO₂) concentrations. While models of photosynthesis perform well at small temporal and spatial scales, there is a need for formulations that perform well at scales relevant for projecting future global changes. Current formulations for simulating large-scale photosynthetic processes such as acclimation are questionable due to the empirical nature of the underlying parameterizations. Here, we develop a theoretical model to predict photosynthetic capacity (i.e., the maximum rate of Rubisco carboxylation, V_cmax) and test it using the largest observational dataset of V_cmax ever assembled (>4500 datapoints at sites across most of the world’s major biomes). The theoretical model used climatic variables for each site to predict V_cmax rates using the coordination hypothesis for photosynthesis, which states that plants allocate resources such that light and CO₂ are equally limiting to photosynthesis. We first explored the sensitivity of the theoretical model to input data. Next, theoretical predictions were compared to observational data at each site. Finally, we explored a set of covariate data to better understand the limitations of the theoretical model and potential inferences regarding the mechanisms underlying photosynthetic allocation at large scales.

Results/Conclusions

The theoretical model used mean growing season CO₂, incident light, temperature, vapor pressure deficit, and elevation to predict V_cmax. As expected, the model predicted a decrease in V_cmax with CO₂ and an increase with light, temperature, vapor pressure deficit, and elevation. These trends suggest an allocation shift from CO₂- to light-capturing mechanisms under elevated CO₂ and a shift from light- to CO₂-capturing mechanisms under elevated temperature and vapor pressure deficit as well as at higher elevations. The observational data followed similar trends. As such, we found that the theoretical model could capture >50% of the variation in observed V_cmax values across all sites. We analyzed the residuals from our comparison against climatic, soil, and plant trait covariates at each site. These analyses revealed that the model tended to overpredict rates at sites and in individuals where nitrogen in leaves was low. This suggests that nitrogen limitation and/or allocation to other plant processes may be influencing allocation to photosynthetic processes. Our findings suggest that the long-term integrated photosynthetic capacity can be well predicted from first principles of photosynthetic theory. Our model should provide a more reliable alternative to the empirical models currently used for projecting future global photosynthetic rates.