Stomatal conductance limited the CO(2) response of grassland in the last century

BACKGROUND: The anthropogenic increase of atmospheric CO(2) concentration (c(a)) is impacting carbon (C), water, and nitrogen (N) cycles in grassland and other terrestrial biomes. Plant canopy stomatal conductance is a key player in these coupled cycles: it is a physiological control of vegetation water use efficiency (the ratio of C gain by photosynthesis to water loss by transpiration), and it responds to photosynthetic activity, which is influenced by vegetation N status. It is unknown if the c(a)-increase and climate change over the last century have already affected canopy stomatal conductance and its links with C and N processes in grassland. RESULTS: Here, we assessed two independent proxies of (growing season-integrating canopy-scale) stomatal conductance changes over the last century: trends of δ(18)O in cellulose (δ(18)O(cellulose)) in archived herbage from a wide range of grassland communities on the Park Grass Experiment at Rothamsted (U.K.) and changes of the ratio of yields to the CO(2) concentration gradient between the atmosphere and the leaf internal gas space (c(a) – c(i)). The two proxies correlated closely (R(2) = 0.70), in agreement with the hypothesis. In addition, the sensitivity of δ(18)O(cellulose) changes to estimated stomatal conductance changes agreed broadly with published sensitivities across a range of contemporary field and controlled environment studies, further supporting the utility of δ(18)O(cellulose) changes for historical reconstruction of stomatal conductance changes at Park Grass. Trends of δ(18)O(cellulose) differed strongly between plots and indicated much greater reductions of stomatal conductance in grass-rich than dicot-rich communities. Reductions of stomatal conductance were connected with reductions of yield trends, nitrogen acquisition, and nitrogen nutrition index. Although all plots were nitrogen-limited or phosphorus- and nitrogen-co-limited to different degrees, long-term reductions of stomatal conductance were largely independent of fertilizer regimes and soil pH, except for nitrogen fertilizer supply which promoted the abundance of grasses. CONCLUSIONS: Our data indicate that some types of temperate grassland may have attained saturation of C sink activity more than one century ago. Increasing N fertilizer supply may not be an effective climate change mitigation strategy in many grasslands, as it promotes the expansion of grasses at the disadvantage of the more CO(2) responsive forbs and N-fixing legumes. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12915-021-00988-4.