The Earth’s carbon and hydrologic cycles are intimately coupled by gas exchange through plant stomata. However, uncertainties in the magnitude and consequences of the physiological responses of plants to elevated CO2 in natural environments hinders modelling of terrestrial water cycling and carbon storage. Here we use annually resolved long-term 13C tree-ring measurements across a European forest network to reconstruct the physiologically driven response of intercellular CO2 (Ci) caused by atmospheric CO2 (Ca) trends. When removing meteorological signals from the 13C measurements, we find that trees across Europe regulated gas exchange so that for one ppmv atmospheric CO2 increase, Ci increased by 0.76 ppmv, most consistent with moderate control towards a constant Ci=Ca ratio. This response corresponds to twentieth-century intrinsic water-use efficiency (iWUE) increases of 14 ±10 and 22 ± 6% at broadleaf and coniferous sites, respectively. An ensemble of process-based global vegetation models shows similar CO2 effects on iWUE trends. Yet, when operating these models with climate drivers reintroduced, despite decreased stomatal opening, 5%increases in European forest transpiration are calculated over the twentieth century.This counterintuitive result arises from lengthened growing seasons, enhanced evaporative demand in a warming climate, and increased leaf area, which together oppose effects of CO2-induced stomatal closure. Our study questions changes to the hydrological cycle, such as reductions in transpiration and air humidity, hypothesized to result from plant responses to anthropogenic emissions

No abstract

Interlaboratory comparisons involving nine European stable isotope laboratories have shown that the routine methods of cellulose preparation resulted in data that generally agreed within the precision of the isotope ratio mass spectrometry (IRMS) method used: +/-0.2 per thousand for carbon and +/-0.3 per thousand for oxygen. For carbon, the results suggest that holocellulose is enriched up to 0.39 per thousand in 13C relative to the purified alpha-cellulose. The comparisons of IRMS measurements of carbon on cellulose, sugars, and starches showed low deviations from -0.23 to +0.23 per thousand between laboratories. For oxygen, IRMS measurements varied between means from -0.39 to 0.58 per thousand, -0.89 to 0.42 per thousand, and -1.30 to 1.16 per thousand for celluloses, sugars, and starches, respectively. This can be explained by different effects arising from the use of low- or high-temperature pyrolysis and by the variation between laboratories in the procedures used for drying and storage of samples. The results of analyses of nonexchangeable hydrogen are very similar in means with standard deviations between individual methods from +/-2.7 to +/-4.9 per thousand. The use of a one-point calibration (IAEA-CH7) gave significant positive offsets in delta2H values up to 6 per thousand. Detailed analysis of the results allows us to make the following recommendations in order to increase quality and compatibility of the common data bank: (1) removal of a pretreatment with organic solvents, (2) a purification step with 17% sodium hydroxide solution during cellulose preparation procedure, (3) measurements of oxygen isotopes under an argon hood, (4) use of calibration standard materials, which are of similar nature to that of the measured samples, and (5) using a two-point calibration method for reliable result calculation.