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 » Discussion

Discussion

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

We found no significant differences in CO₂ emission between the ambient and control plots and therefore used the measurements of the ambient plot to evaluate our drought and rewetting effects. Dissolved organic C fluxes were different between ambient and control plots, but patterns were similar, which is enough to justify comparison of the drought with the ambient plot. Annual soil respiration rates were 3015 kg C ha^-1 yr^-1 in 1993 and 3192 kg C ha^-1 yr^-1 in 1994 at the ambient plot. These rates are low compared with the mean flux rate of 6810 ± 950 kg C ha^-1 yr^-1 in a wide range of temperate coniferous forests summarized by Raich and Schlesinger (1992). The prolonged summer drought and subsequent rewetting of the spruce forest soil at the drought plot did not affect the annual CO₂ emission rate in 1993, but caused an increase of 51% in 1994. We explain this discrepancy between the years by the duration of the drought combined with the level of soil temperature during rewetting.

Drought Effects on Carbon Dioxide Emission
Soil respiration was not significantly reduced by drought treatments. However, our study suggests that the cumulative respiration was lower when length of drought increased. Natural droughts at the ambient plot may have reduced the respiration rate of the forest floor, although matric potential at the 10-cm mineral soil depth indicated sufficient water availability. Carlyle and U Ba Than (1988) reported a strong decline in soil respiration of 70% with decreasing soil moisture in a Monterey pine (Pinus radiata D. Don) stand in southeastern Australia, but that drought was much more severe than in our study. In agreement with our results, they also reported no correlation with temperature during the drought (Fig. 3).

Several studies in temperate forests summarized by Singh and Gupta (1977) have shown a reduction of respiration in the litter layer during dry summer months. Particularly, soils with a thick forest floor are sensitive to changes in temperature and water availability because this layer contains a large pool of labile organic matter and a large contribution of living roots. We could not differentiate between microbial activity and root respiration, the two major sources for soil CO₂ emission. We assume that root respiration was not decreased in our drought treatments since fine root growth was not severely affected (Murach, 1999, personal communication). However, root respiration may have shifted to deeper soil layers to ensure the water supply of the trees (Feil et al., 1988). Consequently, the reduction in CO₂ emission was most likely caused by a decrease in heterotrophic respiration in the dried soil. The relatively small reduction in CO₂ emission caused by droughts points at a dominant role of fungi in acid forest soils instead of bacteria as primary decomposers (Alexander, 1980). Soil fungi are active down to a water potential of -15 MPa, whereas bacteria are already inactive below -1.0 to -1.5 MPa (Swift et al., 1979). However, bacteria maintain a basic metabolism at low moisture content and may also contribute to the almost constant CO₂ emissions during the drought treatments.

Rewetting Effects on Carbon Dioxide Emission
Rewetting generally caused a strong increase in soil CO₂ emission, agreeing well with the laboratory observations of Orchard and Cook (1983), Seneviratne and Wild (1985), and Degens and Sparling (1995). It is unclear whether this strong increase in soil CO₂ emission is caused by an increase in root or heterotrophic respiration. Murach (personal communication, 1999) found an increase of root growth in the topsoil at the drought plot with a delay of 4 to 6 wk after rewetting. The initial CO₂ flush following rewetting was therefore most likely caused by enhanced activity of the decomposer community. Several factors can contribute to this CO₂ pulse after a rewetting. A considerable proportion of soil microorganisms die during drought (van Gestel et al., 1991), and the dead cells can be decomposed quickly during a rewetting. In addition, availability of organic substrates can increase through desorption from the soil matrix (Seneviratne and Wild, 1985) and through increased exposure of organic surfaces to microorganisms (Birch, 1959). Compared with the sharp peak of CO₂ after rewetting in the laboratory studies, the peak of CO₂ emission of our field study was delayed and reached a maximum after 2 to 3 wk. This difference may be caused by the slower rewetting in our field study. In addition, soils in laboratory studies are disturbed (cutting roots, sieving soils, etc.) which may increase the availability of C.

The higher CO₂ release in 1994 than in 1993 is probably associated with the higher temperature level. In 1993, the rewetting was between September and October, a period where soil temperature dropped from 11.1 to 5.7°C (Fig. 2b). In 1994, soil temperature during the first 30 d following rewetting varied between 10.9 and 15.7°C. In this year, both soil temperature and moisture conditions were optimal, so that CO₂ release was 144% higher than the ambient plot. Obviously, the length of the drought period is less important for the CO₂ release than the moisture and temperature conditions during rewetting.

Q₁₀ Values
Q₁₀ values are used to describe the temperature dependence of soil respiration. Carlyle and U Ba Than (1988) showed that during a natural drought period the Q₁₀ value was low and increased from 0.77 to 2.60 with moisture availability in a soil temperature range of 5 to 20°C. An inclusion of a moisture-dependent Q₁₀ term in their model, FRESP, resulted in a strong agreement between measured and fitted values (r² = 0.85). We obtained higher Q₁₀ values of 3.9 for the ambient plot and 5.7 for the drought plot using the modified Arrhenius equation including the water term. These Q₁₀ values may represent temperature dependence of soil respiration under optimum moisture condition. Our Q₁₀ values are considerably higher than the median Q₁₀ value of 2.4 found by Raich and Schlesinger (1992) from seasonal changes in soil temperature and soil respiration rates for various soils under field condition. The Q₁₀ values in their literature review may include possible moisture limitations and therefore do not only represent a temperature but also a moisture dependency. In addition, Q₁₀ values from the literature may be underestimated because Kicklighter et al. (1994) obtained lower Q₁₀ values with air temperature (1.99) than with soil temperature (3.08). Generally, Q₁₀ values appear to be higher in cold regimes and lower under warm regimes. Kirschbaum (1995) found a temperature dependence for Q₁₀ values obtained from CO₂ emissions of soil and litter of various climate regions. In this study, the fitted Q₁₀ values decreased nonlinear from 8 at 0°C to 4.5 at 10°C and 2.5 at 20°C. In cold regions, microorganisms show a stronger temperature reaction compared with temperate or tropical regions because of a high substrate availability during the few summer months when the soils are thawed. For instance, tundra microorganisms are adapted to low temperatures, but respond to temperature increases like microorganisms elsewhere (Flanagan and Veum, 1974). The results of Kirschbaum (1995) are also in agreement with our model results because under field conditions water availability at low temperatures is normally higher than at high temperatures.

The high Q₁₀ values of 5.7 and 5.1 calculated for the drought plot a direct temperature dependence of soil respiration but might also be the result of increased C availability during rewetting. As a considerable part of the CO₂ emission is produced in the forest floor, extended summer droughts and rewetting events at high temperatures may increase soil CO₂ emission rates from forest soils, especially those with a thick O horizon.

Dissolved Organic Carbon Fluxes
The DOC input by throughfall and DOC fluxes at the 10- and 100-cm soil depth were reduced mainly by lower water input at the drought plot (Tables 1 and 3). Our drought and rewetting treatment only slightly reduced the mobilization and degradation of DOC in the upper soil as indicated by the mobilization rate calculated from the difference of DOC input by throughfall and DOC fluxes at the 10-cm soil depth. Natural drought and rewetting events at the ambient plot led to a similar pattern in DOC concentration at the 10-cm soil depth (Fig. 3). Although the DOC flux at the 10-cm soil depth at the ambient plot was higher than at the drought plot, the annual DOC outputs at 100 cm were similar. Dissolved organic C leaching from the 10- to 100-cm soil depth increased at the drought plot because of water flux in macropores during rewetting (Lamersdorf et al., 1998). However, the much lower rates of DOC output below the rhizosphere at both plots suggest that soil has the potential to accumulate C in its sublayers. There is some evidence that DOC is absorbed by sesquioxides in the B horizons and that DOC may play a significant role for carbon storage in the mineral soil of spruce forests (Zech et al., 1994). This is in agreement with the results of Qualls and Haines (1992) that showed that the capacity of microorganisms to degrade DOC is limited and temperature increase does not affect microbial degradation of DOC. Our results suggest that droughts and rewettings under changing climate will have only a limited impact on DOC concentrations in groundwater.

Potential Impacts on Carbon Storage
An inventory of tree and root growth at the site showed no clear difference in litterfall, fine root biomass, and tree diameter growth between the ambient plot and the drought plot for both treatment years (Bredemeier et al., 1998). However, a reduced height growth rate of trees was measured during the drought period in 1993 and 1994 (Dohrenbusch et al., 1999). Although litterfall was not reduced, a decrease of litterfall may be expected in the following years because of the long life span of Norway spruce needles. In the long run, a decrease of litter production and an increase of the decomposition rate may reduce the C storage of this forest ecosystem. On the other hand, elevated atmospheric CO₂ concentrations have direct fertilization effects on tree and root growth (Bazzaz et al., 1990; Johnson et al., 1996; Norby et al., 1992). All of these processes will influence the soil C balance, which may have a direct feedback to the atmospheric CO₂ concentration.

Presently, northern hemisphere temperate forests are considered to be a substantial sink of C (Kauppi et al., 1992; Sedjo, 1992; Birdsey et al., 1993; Nakane and Lee, 1995). Taylor and Lloyd (1992) estimated a net sink effect of 0.33 Pg C yr^-1, based on an inventory of temperate forests in 1985. In the long run, prolonged summer droughts may reduce the C storage in temperate, coniferous forest soils as a result of lower net primary production and a larger CO₂ release during subsequent rewetting.Carlyle Ba Than 1988

ACKNOWLEDGMENTS

We would like to thank Professor E. Veldkamp for invaluable discussions and helpful comments. We also thank the Institute of Bioclimatology of the University of Goettingen for meterological data we used for hydrological modeling of water fluxes. The study was a part of the Environmental Program Project CT91-0052 financially supported by the Commission of European Communities.

Received for publication July 29, 1998.