Increasing atmospheric CO₂ will likely affect both the hydrologic cycle and ecosystem productivity. Current assumptions that increasing CO₂ will lead to increased ecosystem productivity and plant water use efficiency (WUE) are driving optimistic predictions of higher crop yields as well as greater availability of freshwater resources due to a decrease in evapotranspiration. The plant physiological response that drives these effects is believed to be an increase in carbon uptake either by (a) stronger CO₂ gradient between the stomata and the atmosphere, or by (b) reduced CO₂ limitation of enzymatic carboxylation within the leaf. The (a) scenario will lead to increased water use efficiency (WUE) in plants. However, evidence for increased WUE is mostly based on modeling studies, and experiments producing a short duration or step-wise increase in CO₂ concentration (e.g. free-air CO₂ enrichment). We hypothesize that the increase in atmospheric CO₂ concentration is having a positive effect on ecosystem productivity and WUE. To investigate this hypothesis, we analyzed meteorological, ANPP, and soil CO₂ flux datasets together with carbon isotopic ratio (¹³C/¹²C) of archived plant samples from the long term ecological research (LTER) program at Kellogg Biological Station. The datasets were collected between 1989 and 2007 (corresponding to an increase in atmospheric CO₂ concentration of ~33 ppmv at Mauna Loa). Wheat (Triticum aestivum) samples taken from 1989 and 2007 shows significant decrease of discrimination factor (Δ) over period of 18 years. Because Δ is tightly connected to stomata conductance it reflects plant intrinsic WUE (iWUE). Historical changes in the ¹³C/¹²C ratio (δ¹³C) in plant material samples of perennial forbs, Canada goldenrod (Solidago canadensis), taken from adjacent successional fields, indicate changes in Δ upon uptake of CO₂ as well. The Δ of LTER plant samples show a positive feedback between the increasing CO₂ concentration in the atmosphere, air temperature, and plant iWUE. This positive feedback is expressed by (a) nonparallel changes of δ¹³C signal of air (δ_a) and plant samples (δ_p), (b) negative correlation between the Δ and average temperatures during the growth season, although only for temperatures up to 21 °C. The lack of effect at higher temperatures suggests a negative influence of growing season warming on the iWUE. These results suggest a complex feedback between atmospheric CO₂ increase, plant physiology, ecosystem productivity, and soil CO₂ fluxes. These complex effects support our hypothesis of a CO₂ fertilization effect on plant productivity, and they raise additional questions regarding adaptation of plants to changing atmospheric CO₂ and climate.