- Carbon dynamics under natural and manipulated meteorological boundary conditions in a forest and a fen ecosystem (2009)
- Current climate models predict changes in the amount, intensity, frequency and type of precipitation within this century. These changes are likely to result in an increasing frequency of severe drought periods in summer, causing irregular and extreme drought stress in well-drained soils or a lowering of the water table in water-logged soils. Due to rising temperatures precipitation is more likely to occur as rain rather than snow, resulting in reduced snowpacks in winter. In some regions, this can lead to an increasing frequency of soil frost. In summary, changes in the global water cycle can significantly affect boundary conditions within soils. This thesis investigated the impact of extreme meteorological boundary conditions on CO2 fluxes in two ecosystems in South-eastern Germany. Using a combination of field site manipulation and laboratory experiments we investigated the effects of prolonged summer drought and soil frost on soil C dynamics in a Norway spruce forest. In a minerotrophic fen located nearby, the effect of water table lowering (as a result of summer drought) on ecosystem C dynamics was quantified. Additionally, soil C dynamics at both sites were modeled under current meteorological conditions. For the Norway spruce forest, modeling indicated that soil C turnover predominantly occurred within the organic horizons. During the last decades, the soil has acted as a small sink. The possibility of altered C dynamics at the site due to undocumented liming has to be considered when comparing results presented here to results from other sites. For the fen, modeling revealed that soil C turnover was dominated by processes occurring within the uppermost 15 cm of the peat and that root biomass was a very important soil C stock. Most important, modeling indicated that the fen was turned into a net C source during the last decades, presumably because of disturbance of the hydrological conditions. Results from this fen cannot be regarded as representative for undisturbed peatlands. Soil frost was induced at the forest site by removing the snowpack in the winter of 2005/2006. On the snow removal plots, soil frost occurred down to a depth of at least 15 cm and for several weeks, in contrast to the snow-covered control plots where no soil frost occurred. Soil C losses were significantly reduced not only during the soil frost period but also in the summer of 2006. This phenomenon could be explained by changes in the composition of the microbial community due to soil frost, primarily a reduction of fungal biomass. To investigate the effect of drought on soil C dynamics we experimentally induced prolonged drought at the forest-site by excluding throughfall with a transparent roof during the summers of 2006-2008. Additionally, undisturbed soil columns from the site were subjected to drought in the laboratory. In both experiments, drought reduced total soil C losses in comparison to C losses from a control. This reduction was mainly owed to decreased soil respiration rates during the actual drought period, but water repellency also hindered rewetting of the dry soil, thus further prolonging the period of reduced soil respiration rates. In the past, mobilization of stabilized C due to drying-wetting has been repeatedly discussed as a possibility to actually enhance soil C losses. In the studies presented here, no evidence for this assumption was found. Soil C mineralization rates were reduced during drought and recovery was slow, possibly delayed by water repellency and preferential flow. At the fen site we used two approaches: (i) Experimental lowering of water tables to measure resulting C fluxes in comparison to C fluxes under natural conditions (i.e. control plots), and (ii) repeated measurements under varying natural conditions to be able to later statistically identify the main drivers of CO2 fluxes. We included measurements of C uptake and respiration by aboveground vegetation, thus being able to study ecosystem rather than soil C dynamics at the fen site. In summary, the impact of the water table on CO2 fluxes in and out of the fen ecosystem was minimal. Soil respiration was not affected at all by the manipulative lowering of the water table from ca. 15 cm down to more than 60 cm, most likely due to low substrate quality in deeper peat. Measurements of the natural C dynamics indicate that water table could have an impact on soil respiration within the uppermost 0-15 cm of the soil, but predominantly low water tables during summer under current boundary conditions make it unlikely that further lowered water tables due to climate change will markedly affect soil respiration rates at this site. In summary, CO2 fluxes at the site are presumably very resilient towards an increasing frequency of summer drought resulting in lowering of the water table.