How does increasing CO2 influence the land-atmosphere exchange of carbon and water in response to soil and air dryness?

Variability in the photosynthetic uptake of CO₂ by plants due to climate variability plays an essential role in modulating the growth rate of atmospheric CO₂. In particular water stress induced by the compounding effect of vapor pressure deficit (VPD) and soil moisture anomalies has a large bearing on photosynthetic CO₂ uptake, especially in semi-arid areas. The ongoing rise in atmospheric CO₂ concentration can influence the water-use efficiency of plants, their carbon assimilation rate, and consequently the global cycles of carbon and water.  However, the extent to which physiological effects of increasing CO₂ influence the coupling of VPD, soil moisture, and land-atmosphere CO₂ fluxes is currently poorly understood.

In this study we use the terrestrial ecosystem model QUINCY (QUantifying Interactions between terrestrial Nutrient CYcles and the climate system, Thum et al. 2019, GMD) to study CO₂-induced changes in the interaction of plant productivity and both soil moisture and VPD. With the sensitivity of stomatal conductance to atmospheric CO₂ concentration implemented in the model, we investigate century-long simulations across different climate regimes and biomes. In semi-arid regions we find a positive relationship between anomalies of VPD and gross primary productivity (GPP) at both start and end of summer months, and a negative relationship in the dry period (usually from June to August in boreal summer). This suggests there is a transition in the limiting factor of GPP from energy (compound effect of temperature and radiation) to water and back in the course of one year. The negative correlation between VPD and GPP during the dry period weakens over time with rising CO₂, while a stronger positive correlation between soil moisture and GPP becomes apparent. After quantifying and understanding the CO₂ effects in these model simulations, we apply our analysis framework to observational data from the FLUXNET site collection to analyze whether we can confirm the model-based findings despite shorter records.