Soil and biomass carbon pools in model communities of tropical plants under elevated CO2.

The experimental data presented here relate to the question of whether terrestrial ecosystems will sequester more C in their soils, litter and biomass as atmospheric CO2 concentrations rise. Similar to our previous study with relatively fertile growth conditions (Körner and Arnone 1992), we constructed four rather nutrient-limited model communities of moist tropical plant species in greenhouses (approximately 7 m2 each). Plant communities were composed of seven species (77 individuals per community) representing major taxonomic groups and various life forms found in the moist tropics. Two ecosystems were exposed to 340 μl CO2 l-1 and two to 610 μl l-1 for 530 days of humid tropical growth conditions. In order to permit precise determination of C deposition in the soil, plant communities were initially established in C-free unwashed quartz sand. Soils were then amended with known amounts of organic matter (containing C and nutrients). Mineral nutrients were also supplied over the course of the experiment as timed-release full-balance fertilizer pellets. Soils represented by far the largest repositories for fixed C in all ecosystems. Almost 5 times more C (ca. 80% of net C fixation) was sequestered in the soil than in the biomass, but this did not differ between CO2 treatments. In addition, at the whole-ecosystem level we found a remarkably small and statistically non-significant increase in C sequestration (+4%; the sum of C accretion in the soil, biomass, litter and necromass). Total community biomass more than quadrupled during the experiment, but at harvest was, on average, only 8% greater (i.e. 6% per year; n.s.) under elevated CO2, mainly due to increased root biomass (+15%, P=0.12). Time courses of leaf area index of all ecosystems suggested that canopy expansion was approaching steady state by the time systems were harvested. Net primary productivity (NPP) of all ecosystems-i.e. annual accumulation of biomass, necromass, and leaf litter (but not plant-derived soil organic matter)-averaged 815 and 910 g m-2 year-1 at ambient and elevated CO2, respectively. These NPPs are remarkably similar to those of many natural moist tropical forested ecosystems. At the same time net productivity of soil organic matter reached 7000 g dry matter equivalent per m2 and year (i.e. 3500 g C m-2 year-1). Very slight yet statistically significant CO2-induced shifts in the abundance of groups of species occurred by the end of the experiment, with one group of species (Elettaria cardamomum, Ficus benjamina, F. pumila, Epipremnum pinnatum) gaining slightly, and another group (Ctenanthe lubbersiana, Heliconia humilis, Cecropia peltata) losing. Our results show that: (1) enormous amounts of C can be deposited in the ground which are normally not accounted for in estimates of NPP and net ecosystem productivity; (2) any enhancement of C sequestration under elevated atmospheric CO2 may be substantially smaller than is believed will occur (yet still very important), especially under growth conditions which permit close to natural NPP; and (3) species dominance in plant communities is likely to change under elevated CO2, but that changes may occur rather slowly.