The authors assessed the effect of changes in rainfall amount and distribution …

• Abstract
• Introduction
• Materials and Methods
• Results
• Discussion
• References
• Figures
• Tables

Biology Articles » Bioclimatology » A Climate Change Scenario for Carbon Dioxide and Dissolved Organic Carbon Fluxes from a Temperate Forest Soil » Results

Results

- A Climate Change Scenario for Carbon Dioxide and Dissolved Organic Carbon Fluxes from a Temperate Forest Soil

Results 

Carbon Dioxide Emissions in Ambient and Control Plots
Carbon dioxide emission (weekly average) (Fig. 1a) showed a clear seasonal pattern: it increased at the beginning of spring and decreased in fall, following the pattern in soil temperature (Fig. 1b). In 1993, the highest emissions (1.55 ± 0.23 g C m^-2 d^-1) occurred from August to early October. During this year no severe drought occurred except for a short dry period in July with the lowest matric potential of -61 kPa (Fig. 1c). This natural drought was accompanied by a slight drop of 16% in CO₂ emission rates compared with June, although soil temperature increased by 1.0°C. In 1994, the highest emissions were observed from July to early September (1.53 ± 0.22 g C m^-2 d^-1). Low throughfall occurred from July to August 1994 (Fig. 1d), resulting in a soil drought for several weeks with low matric potential below -120 kPa (Fig. 1c). Concurrently, CO₂ emissions (1.35 ± 0.21 g C m^-2 d^-1) decreased from the end of July to August when soil temperature reached its maximum of 15.7°C. More than 75% of the annual total emission occurred during April to October.

Cumulative CO₂ emission from the ambient plot was 3015 kg C ha^-1 in 1993 and 3192 kg C ha^-1 in 1994 (Table 2) . This difference may be explained by the difference in mean soil temperatures of 6.0°C in 1993 and 6.8°C in 1994. Cumulative CO₂ emission from the control plot (3205 kg C ha^-1) was not significantly different from the ambient plot in 1993. Although throughfall was reduced by 14% in 1993 in the control plot (Table 1), we observed no roof effect on soil respiration.

Drought Effects on Carbon Dioxide Emission
Compared with the annual amount of throughfall at the ambient plot, the amount of throughfall at the drought plot was reduced by 45% in 1993 and 14% in 1994 (Table 1). After the drought experiment started in April 1993, the soil matric potential at the 10-cm depth sharply decreased from -5 to -126 kPa and maintained at this level throughout the 172-d drought period (Fig. 2c) . Compared with the treatment period in 1993, the drought in 1994 was less prolonged and matric potential remained at a level below -100 kPa for only 4 wk. The summer drought experiments only slightly affected soil CO₂ emissions. Emission rates slowly decreased when the droughts started (Fig. 2a and 2c). During the 172-d drought period in 1993, the soil emitted 1639 kg C ha^-1(Table 2). This rate was 23.3% (P lower than that of the control plot and 12.4% lower (P > 0.09) than that of the ambient plot. During the less intense 108-d drought period in 1994, the emission rate was not reduced (-1.2%). The stronger reduction in 1993 was due to the long and severe drought induced at the drought plot compared with the ambient plot, which received a throughfall input of 483 L m^-2(Table 1).

Rewetting Effects on Carbon Dioxide Emission
Rewetting resulted in an immediate increase of soil emission in both treatment years. The emission rates continued to increase for 3 wk and reached their maximum in October 1993 (2.24 g C m^-2 d^-1) and August 1994 (4.94 g C m^-2 d^-1). During rewetting, mean soil temperature was 9.2°C in 1993 and 13.8°C in 1994. The drop in CO₂ emission after the peak flush was accompanied by a decrease in soil temperature.

Rewetting lasted for a shorter time in 1993 than in 1994, but the intensity of rewetting was higher in 1993 (193 L m^-2 within 19 d) than in 1994 (184 L m^-2 within 33 d, Table 1). During the first 30 d after the start of rewetting, CO₂ emission was 504 kg C ha^-1 in 1993 and 1088 kg C ha^-1 in 1994 (Table 2). This corresponded with about one-sixth of the annual C release in 1993 and one-fifth in 1994. Compared with the C release at the ambient plot during these 30 d, emission rates at the drought plot were 48% (P P 0.01) higher in 1994.

Both duration of the summer drought and soil temperature during the rewetting had an effect on the annual CO₂ release. The drought period and subsequent rewetting led to an insignificantly lower annual rate in 1993. By contrast, the drought period and rewetting event in 1994 increased the annual CO₂ emission rate significantly (P

Quantitative Analyses of Effects of Soil Temperature and Matric Potential on Carbon Dioxide Emission
Carbon dioxide emission rates increased exponentially with increasing temperature from -1 to 16°C during the untreated and rewetting periods (Fig. 3) . It is apparent that soil temperature was the dominant factor for soil CO₂ emissions when matric potential was higher than -20 kPa (mostly during the untreated and rewetting period, Fig. 2c). During the drought periods, CO₂ emission rates were below 50 mg C m^-2 h^-1. Apparently a moisture-dependent threshold value existed for soil respiration below which temperature effects become virtually zero.

The calculated CO₂ emission rates using Eq. [1] and [2] strongly correlated with those measured at both the ambient (r² = 0.85) and drought plots (r² = 0.95) (Fig. 1a and 2a). The parameters in Eq. [1] were fitted as follows: E = 85.0 kJ mol^-1, A = 2.57 x 10¹⁷, and a = 0.0029 for the control plot; and E = 109 kJ mol^-1, A = 1.02 x 10²², and a = 0.0049 for the drought plot. Calculated Q₁₀ values were 3.9 for the ambient plot and 5.7 for the drought plot. The high values for E and a indicate a much stronger influence of soil temperature and matric potential on soil respiration at the drought plot. Disregarding soil matric potential resulted in lower correlations at the ambient (r² = 0.77) and drought plots (r² = 0.63). The parameters were fitted as follows: E = 65.8 kJ mol^-1 and A = 6.83 x 10¹³ for the ambient plot, and E = 102 kJ mol^-1 and A = 4.05 x 10²⁰ for the drought plot. The calculated Q₁₀ values of 2.87 for the ambient plot and of 5.11 for the drought plot suggested a smaller temperature effect than when matric potential was included.

Dissolved Organic Carbon Fluxes
During the growing season, mean DOC concentrations in the 10- and 100-cm soil depths were higher than in winter at the ambient and drought plots (Fig. 4) . The DOC concentrations in the top soil of the ambient plot (21.5 ± 5.2 mg L^-1) and the drought plot (18.5 ± 4.1 mg L^-1) were significantly higher (P L^-1 at ambient plot and 5.9 ± 3.3 mg L^-1 at drought plot). In both years and at both depths, DOC concentrations peaked following rewetting both at the drought and ambient plots. The DOC concentrations in the drought plot were not significantly different from the ambient plot.

The annual rates of DOC input by throughfall at the ambient plot were 106 kg ha^-1 in 1993 and 93 kg ha^-1 in 1994, while the drought and control plots had considerably lower input rates (Table 3) . The large difference in DOC input to the soil was mainly caused by the lower water input at the drought and control plots. Using the water flow results modeled by Xu et al. (1998), we calculated DOC fluxes between 91 and 176 kg ha^-1 at the 10-cm depth (Table 3). This higher DOC flux at 10 cm compared with the input indicated a mobilization of DOC in the upper soil at all plots. The difference between DOC input and output soil flux at 10 cm was highest at the ambient plot and lowest at the control plot. Much lower rates of DOC leaching at the 100-cm soil depth were calculated for all plots (28–46 kg ha^-1).