Xiahong Feng Long-term c i /c a response of trees in western North America to atmospheric CO 2 concentration derived from carbon isotope chronologies Received: 12 June 1997 / Accepted: 29 June 1998 Abstract To evaluate how the land carbon reservoir has been responding to the rising CO 2 concentration of the atmosphere, it is important to study how plants in nat- ural forests adjust physiologically to the changing at- mospheric conditions. Many experimental studies have addressed this issue, but it has been dicult to scale short-term experimental observations to long-term eco- system-level responses. This paper derives carbon-iso- tope-related variables for the past 100±200 years from measurements on trees from natural forests. Calcula- tions show that the c i /c a ratios [c i /c a is the ratio of the CO 2 concentration (lmol mol )1 ) in the intercellular space of leaves to that in the atmosphere] of the trees were constant or increased slightly before the 20th cen- tury, but changed more rapidly in the 20th century; some increased, some decreased, and some stayed con- stant. In contrast, the CO 2 concentration inside plant leaves increased monotonically for all trees. Key words Atmospheric CO 2 concentration á Tree rings á Long-term changes á Carbon isotopes á c i /c a ratios Introduction It is now generally accepted that land biota have been responsible for taking up the portion of anthropogenic CO 2 that is not accounted for by the ocean and atmo- sphere sinks (e.g., Bender 1996; Bender et al. 1996, and references therein; Keeling et al. 1996). However, much less agreement has been reached on the mechanism of the land biosphere carbon sink. Experimental studies have shown that terrestrial carbon storage can respond to elevated CO 2 concentrations in a number of ways, including changes in photosynthetic capacity through acclimation (Gunderson and Wullschleger 1994), en- hanced activity of the enzyme ribulose bisphosphate carboxylase/oxygenase (rubisco) (Luo et al. 1996; Sharkey 1988), inhibition of leaf respiration (Amthor et al. 1992), changes in soil nutrient dynamics (Johnson et al. 1996), changes in litter quality and decomposition rates of foliar litter (O'Neill and Norby 1996), more litter fall (Field et al. 1996), and improvements in plant water-use eciency (e.g., Polley et al. 1993). All of these experiments are short-term, and it is usually dicult to scale the experimental observations up to long-term, ecosystem-level responses. Searching for long-term forest responses to the rise of the CO 2 level in the atmosphere using tree-ring chro- nologies has provided only a few positive results (Graybill 1987; Graybill and Idso 1993; LaMarche et al. 1984;). One of the diculties is to separate the eects of climate from the eects of atmospheric CO 2 concentra- tion in any enhanced tree growth. This paper focuses on long-term changes in the c i /c a ratio and c i of plants in response to increasing atmospheric CO 2 concentration. The c i /c a ratio is the ratio of the CO 2 concentration (c i, lmol mol )1 ) in the intercellular space of leaves to that in the atmosphere (c a ), and long-term variations of this ratio have a bearing on plant intrinsic water-use e- ciency (Farquhar et al. 1989). My calculations of c i /c a ratios are based on carbon isotopic tree-ring chronolo- gies from 20 sites in North America. Eects of non- atmospheric conditions were removed from the original tree-ring series, and therefore the calculated values can be considered to be aected predominantly by the in- crease of CO 2 concentration in the atmosphere. Methods De®nitions Stable carbon isotopic composition of natural materials is expres- sed in the d notation: Oecologia (1998) 117:19±25 Ó Springer-Verlag 1998 X. Feng Department of Earth Sciences, Dartmouth College, 6105 Fairchild, Hanover, NH 03755, USA e-mail: xiahong.feng@dartmouth.edu