The feedback from forest growth to atmospheric CO₂ is a key issue in coupling carbon cycle models to climate models. A free-air CO₂ enrichment experiment to examine forest responses to rising CO₂ was established in 1997 in a 10-year-old deciduous forest in eastern Tennessee. The closed-canopy sweetgum stand has been exposed to current ambient atmospheric CO₂ (aCO₂) or air enriched with CO₂ to 550 ppm (eCO₂) since 1998 in 25-m diameter plots. Analysis of data from the first 6 years of the experiment indicated that net primary productivity (NPP) was significantly enhanced by eCO₂, and that this was a consistent and sustained response. Now, with 10 years of data, that conclusion must be revised.

Results/Conclusions

The response of NPP to eCO₂ has declined over time. The mean increase in NPP over 10 years of CO₂ enrichment was 21%, but the NPP response has steadily declined from 20% in 2004 to 10% in 2007. NPP in eCO₂ is closely correlated with NPP in aCO₂, suggesting that the trees are responding to similar environmental influences. The relative effect of CO₂ was greatest in years when NPP was greatest and declined as ambient conditions caused NPP to decline, as was the case from 2005-2007. We examined possible causes of the declining growth response under eCO₂. Leaf area index has been relatively constant and there is no correlation between interannual variation in NPP and LAI. Soil moisture can explain some of the variation in fine-root production, which is the most responsive NPP component to eCO₂, but it cannot explain the more gradual decline in NPP response to eCO2. The most likely explanation for that decline is reduced N availability, which is supported by a model of the stand's carbon-nitrogen-water economy. We have shown previously that tree growth at this site is N limited, and our 10-year record provides evidence that N limitation has been increasing throughout the duration of the experiment, based on the steady decline in foliar nitrogen concentration in both aCO₂ and eCO₂ plots. Increased fine-root proliferation under eCO₂, especially at depth in the soil, has thus far enabled sustained production responses and increased N uptake, but this has not benefited aboveground wood production. Wood production has declined, and the effect of CO₂ on wood production, which previously was small, has now disappeared. Plant and soil analyses from the remaining 2 years of the experiment will further elucidate the mechanism of NPP decline.