The present work was part of the Giessen FACE experiment (FACE = free air carbon dioxide enrichment). The Giessen FACE contains 3 CO₂ enrichment and 3 control rings, which were installed in 1997 on the grassland investigation area of the "Environmental Monitoring and Climate Research Station Linden" near Giessen, Germany. Since May 1998, the CO₂ concentrations have been raised 20% during daylight hours. The species-rich, wet grassland has been mown twice per year for decades and does not seem to have a ploughing history. It has been fertilized with 40 kg N ha^-1 yr^-1 and continuously cut twice per year (with n = 75 samples per CO₂ treatment and harvest date). The trace gas fluxes were measured every 3 to 4 days with the closed chamber method (chamber diameter 1 m, n = 9 per CO₂ treatment, i.e. 7 m² of covered area). Additionally, several microbial parameters were measured during the first year of CO₂ enrichment in order to detect and understand possible changes in N₂O fluxes. The relative amount of nitrification and denitrification in total N₂O emission were determined by the application of the acetylene inhibition technique (5 Pa C₂H₂ applied) to soil cores. Denitrification enzyme activity (DEA) and the net nitrification rates were measured every 1 to 2 months.

Harvests of the aboveground biomass were done from 1993 onwards. Before the start of the CO₂ enrichment experiment there were no significant differences between the control and the enrichment plots. However from the second harvest in September 1999 onwards, the biomass of the enrichment rings was significantly greater that that of the control rings, and it continued to be significantly higher in 2000. In the third year of CO₂ enrichment, the biomass gain accounted for 10 % more than the control biomass. Compared to other studies were the CO₂ concentrations (the actual or preindustrial) usually have been doubled, the biomass gain was delayed. The magnitude of the biomass increase was, however, comparable to other experiments with (semi-)natural grasslands under higher CO₂ enrichment. The composition of the three functional groups grasses, legumes and non-leguminous herbs and the leaf area index did not change significantly during the three years of CO₂ enrichment.

The community CO₂ respiration was significantly higher under CO₂ enrichment. The rise was most likely attributable to an increased soil respiration that was the biggest part of the community respiration. In addition, the concentration of organic carbon (KCl extraction) was significantly higher. The community respiration increase was not constant during the 2,5 years of measurement: it was highest during the winter month, and it decreased slowly while the FACE experiment continued (but it was still higher under elevated CO₂). The delayed occurrence of the aboveground biomass increase, in combination with the receding respiration increase, point towards a long-term acclimation of the grassland ecosystem to the elevated CO₂.

Before starting the experiment, the N₂O flux rates were not significantly different from each other. With the beginning of the FACE experiment, however, the N₂O emissions during the summer-autumn period were considerably higher under elevated CO₂. The mean N₂O losses from the FACE rings amounted to about 290 % of the control value (=100%) during the enrichment years 1998 – 2000 (4,3 kg N₂O-N ha^-1 yr^-1 versus 1,5 kg N₂O-N ha^-1 yr^-1). Additional measurements failed to explain the lasting rise of the N₂O emission level. The driest enrichment plot, which showed the highest N₂O emission, had the lowest denitrification enzyme activity and the lowest net nitrification rates. Moreover, no increase in mineral or organic N concentrations occurred under CO₂ enrichment. Mineral nitrogen concentrations were very low and in general almost not detectable. In addition, neither soil moisture (TDR sensors, 0-15 cm) nor soil temperatures (5, 10 and 20 cm depth) changed significantly with CO₂ enrichment.

Methane oxidation rates were relatively high despite high soil moisture and despite fertilization (NH₄ NO₃). The maximum values were comparable to rates from neutral forest soils and were about 130 µg CH₄-C m^-2 h^-1 in summer. Before beginning the FACE experiment and during the first year of CO₂ enrichment, the CH₄ oxidation rates were almost identical on the control and enrichment plots. However, with the second year of enrichment, the CH₄ oxidation rates started to decline on the FACE plots. The decrease continued during the third year where the rates of the enrichment rings amounted to only 75% of the rates of the control plots during the summer period. During September 2000 and under oxic soil conditions, a CH₄ emission event with a maximum value of 870 µg CH₄-C m^-2 h^-1 was observed in one of the rings. It was great enough to influence the balance of the month. While the soil water content did not rise, alterations in methanotrophic communities might be the cause of the CH₄ oxidation decline, as well as a rise in methane production under oxic or micro-aerobic soil conditions (e.g. in the rhizosphere).

For both trace gases, N₂O and CH₄, a positive feedback of elevated CO₂ has been found on the biological processes that are involved in the rise of the atmospheric concentrations of these gases. A great discrepancy between hypothesis and the in-situ measurements of the trace gas fluxes under CO₂ enrichment demonstrates the necessity to evaluate 'laboratory' knowledge in field experiments. A number of observed acclimation effects showed that short-term experiments may strongly under- or overestimate processes and effects of elevated CO₂, especially if these are scaled to the globe.