The effect of experimental ecosystem warming on CO2 fluxes in a montane meadow

Climatic change is predicted to alter rates of soil respiration and assimil ation of carbon by plants. Net loss of carbon from ecosystems would form a positive feedback enhancing anthropogenic global warming. We tested the eff ect of increased heat input, one of the most certain impacts of global warm ing, on net ecosystem carbon exchange in a Rocky Mountain montane meadow. O verhead heaters were used to increase the radiative heat flux into plots sp anning a moisture and vegetation gradient. We measured net whole-ecosystem CO2 fluxes using a closed-path chamber system, relatively nondisturbing bas es, and a simple model to compensate for both slow chamber leaks and the CO 2 concentration-dependence of photosynthetic uptake, in 1993 and 1994. In 1 994, we also measured soil respiration separately. The heating treatment al tered the timing and magnitude of net carbon fluxes into the dry zone of th e plots in 1993 (reducing uptake by approximate to 100 g carbon m(-2)), but had an undetectable effect on carbon fluxes into the moist zone. During a strong drought year (1994), heating altered the timing, but did not signifi cantly alter the cumulative magnitude, of net carbon uptake in the dry zone . Soil respiration measurements showed that when differences were detected in dry zone carbon Fluxes, they were caused by changes in carbon input from photosynthesis, not by temperature-driven changes in carbon output from so il respiration. When differences were detected in dry-zone carbon fluxes, t hey were caused by changes in carbon input from photosynthesis, not by a te mperature-driven changes in carbon output from soil respiration. Regression analysis suggested that the reduction in carbon inputs from plants was due to a combination of two soil moisture effects: a direct physiological resp onse to decreased soil moisture, and a shift in plant community composition from high-productivity species to low-productivity species that are more d rought tolerant. These results partially support predictions that warming m ay cause net carbon losses from some terrestrial ecosystems. They also sugg est, however, that changes in soil moisture caused by global warming may be as important in driving ecosystem response as the direct effects of increa sed soil temperature.