Introduction 

Soil respiration is the largest component of ecosystem respiration ([13]) and therefore a key element in the carbon source/sink role especially for forest ecosystems. The ongoing increase in both atmospheric temperature and CO₂ content has an enhancing effect on the flux of CO₂ from soil respiration ([8]). In spite of the increased soil respiration flux with elevated air temperature and CO₂ level, the individual components of soil respiration [autotrophic root respiration, heterotrophic microbial respiration, and litter and soil organic matter (SOM) degradation] remain difficult to quantify. This is a challenging quest because all existing methods ([9]) bear their respective limitations. Promising new insights have been obtained from the use of stable carbon isotopes ([4], [11], [15]), where either: a) natural variability in the abundance of carbon isotopes in different compartments is examined; or b) a labelled ¹³CO₂ signal is applied to compartments of the soil carbon cycle.

The aims of this mini review are firstly to present the main idea behind selected stable isotope approaches, namely: a) the method of natural ¹³C abundance; b) the method of isotopic labelling with ¹³C-depleted CO₂; and (c) the method of isotopic labelling with ¹³C-enriched CO₂. The second aim is to evaluate the methods’ applicability for the separation of the autotrophic and heterotrophic components of soil respiration in field conditions.

Method of natural abundance

The heavier stable isotope ¹³C is much less abundant than the lighter ¹²C.¹³C ≈ 1.10% of the total carbon ([5]). The isotopic carbon signature of any material is reported in the delta notation (δ¹³C) with per mil (‰) units, which refers to the ratio between the isotopic ¹³C/¹²C ratio (R) of the sample to the international reference V-PDB standard (eqn. 1):

where a higher δ¹³C stands for more heavy stable isotopes than the standard (enrichment) and a lower δ¹³C means less heavy stable isotopes than the standard (depletion). It is this ratio that is of particular interest for plant ecology as it is altered / fractionated to a varying degree by the different photosynthetic metabolisms of C₃ and C₄ plants. Because of differences in CO₂ uptake and therefore different rates of diffusive fractionation, tissue of C₃ plants will be isotopically depleted by about -29‰ ([9]), while C₄ plant tissue shows a depletion of only -13‰ as compared to the standard. Numerous processes ([14]) including physiological status and meteorological conditions can influence these averaged values, but still these differences afford an excellent starting-point for separating soil respiration components by natural abundances of isotopes in the case where a C₃ canopy follows a C₄ crop (or vice versa) ([17]). Up to now, this is still an important limitation in the applicability of the natural abundance approach in forests, because C₃ and C₄ substitutions involving trees are rare. In the original forest without C₃ and C₄ substitutions, only small differences in the isotopic signal of soil respiration components were found in laboratory experiments. These differences might be diminished by the great spatial variability of soil efflux isotopic signal found in situ ([6]).

Formánek & Ambus ([6]) were able to distinguish δ¹³C of CO₂ efflux from root-free mineral horizons, root-free humus layer and roots. The δ¹³C signatures alone were not sufficient to separate contributions of root respiration and SOM decomposition in the forest soil CO₂ efflux measured in the field. The necessary estimate of root-free respiration profile δ¹³C signature was roughly calculated by bulk densities related contribution of each layer to the total CO₂ efflux. Therefore, the estimation of 43% of root from total soil respiration could be realistic, but needs to be considered with caution, although it falls into the 29-59% range as reported by other carbon isotopic studies of temperate mixed deciduous forests ([3], [7]).

Method of isotopic labelling

There are two main labelling processes available: (i) labelling with ¹³C-depleted CO₂ originating from fossil fuel combustion; (ii) labelling with ¹³C-enriched CO₂ produced by nuclear processes.

The isotopic signature of air CO₂ is -8.00‰ at present ([2]), but continuously moves towards more negative values because of fossil C combustion. CO₂ derived from fossil carbon sources, i.e., mineral oil and coal, shows a depleted carbon isotopic signature of about -35.00‰ ([10]), which turns it in an ideal and inexpensive substrate for isotopic labelling of the soil carbon cycle under both field and enclosed conditions. There are many ongoing and already conducted isotopic labelling studies (e.g., [1]) with very promising results, but in the present paper we focus on one particular work by Lin et al. ([10]), because this study achieved clear partitioning of three forest soil respiration components (i.e., CO₂ efflux from root, litter, and SOM). In this study, 4-year old Douglas-fir trees were grown under controlled conditions in terracosms with four different combinations of ambient and elevated CO₂ concentrations and temperature. SOM and soil litter originated not from the present trees but from an old growth Douglas-fir forest. In both the elevated and ambient groups, a constant CO₂ concentration was maintained by fumigating the terracosms with tank CO₂. The tank CO₂ fumigation resulted in newly formed plant material such as needles and roots with a δ¹³C of -35.77 ± 0.09‰ (± standard deviation) compared to an atmospheric CO₂ δ¹³C of -8.55 ± 0.13‰ in the area where SOM as well as the organic litter of the soil in use had formed. With this approach alone, it would have not been possible to separate the three main components of soil respiration and to understand their individual responses to elevated temperature and CO₂, because CO₂ both from SOM and litter degradation carrying the same carbon isotopic signal. Lin et al. ([10]) found an elegant solution to this problem: they took advantage of the information in the δ¹⁸O oxygen isotopic signature in the soil respired CO₂ molecule, which is in equilibrium with the respective signature in soil water. When soil water evaporation occurs, there is a diffusive fractionation of the water molecules carrying the two different stable oxygen isotopes ¹⁶O and ¹⁸O. This isotopic enrichment of ¹⁸O takes place only in the upper 10 cm of soil which allows a separation of CO₂ originating from litter degradation and CO₂ from SOM degradation according to their respective oxygen isotopic signature. The disadvantages of this particular approach apply as well for other labelling studies, i.e., FACE. In most cases, the task of maintaining constant conditions results in an elevated CO₂ concentration. This means a major disturbance to the ecosystem carbon flux, because in the supply to microbial communities, the intricate relation between the supply of fresh energy abundant carbon compounds (leaf litter, rhizo-exudates and rhizo-deposition) and old, more nutrient prone soil organic carbon is altered ([12]). The need to label some components limits the applicability of the method under field conditions, even if a main focus is put on maintaining ambient conditions concerning the CO₂ concentration.

A very interesting example of isotopic labelling with enriched ¹³CO₂ is the study of Subke et al. ([16]) conducted on stand scale in a naturally regenerated Pinus silvestris forest in northern Sweden. Subke et al. ([16]) pulse-labeled (3h) the trees in two chambers (each 200 m³) with highly enriched ¹³CO₂ and monitored the resulting ¹³CO₂ soil efflux with a combination of deep, root free and surface collars over a period of 6 days. In this setup, the isotopic ¹³C signature of deep collars excluded plant derived CO₂. The authors were able to separate the initial physical ¹³CO₂ flux caused by the direct tracer diffusion from the atmosphere and the later occurring biological ¹³CO₂ flux caused by root respiration of photosynthates (biotic label return). Approximately 2-3 days after pulse labelling, only in deep collars ¹³CO₂ returned from the initial peak to near natural abundance, while in surface collars, ¹³CO₂ likely from the respiration of labelled assimilates allocated below-ground, increased again and reached the maximum ca. 3.5-4 days after the pulse. In spite of the separation of autotrophic pulse-derived ¹³CO₂ tracer from the difference between surface collars and deep collars, the experiment was primarily designed to separate abiotic and biotic tracer returns. However, the method has the potential for separating the autotrophic and heterotrophic part of soil CO₂ efflux, if the experimental design allows plant photosynthesis to use only the labelled ¹³C-enriched CO₂ source.

  Conclusions 

In this mini review, we compared the use of natural abundance and labelled carbon isotope methods in partitioning forest soil autotrophic and heterotrophic respiration components. The natural abundance carbon isotope method has the potential of separating forest soil respiration components, although its applicability is limited to situations where overplanting from C₄ to C₃ plants occurred. On forest sites with an exclusive C₃ vegetation history, the individual soil components must be physically isolated (disturbing the carbon cycle). In addition, natural δ¹³C signatures show high spatial variation allowing only for a rough partitioning of soil respiration. The difficulty of relying on the existing C₄ SOM can be overcome by the tank fumigation with ¹³C-depleted or ¹³C-enriched CO_2. The newly formed organic material with depleted or enriched ¹³C signatures enables to distinguish the root autotrophic respiration from heterotrophic respiration of litter, roots and SOM formed before labelling with natural abundance ¹³CO₂. However, also long-term fumigation experiments require an exclusion of the roots for a separation of the components of autotrophic soil respiration. A step further to non-invasive separation of autotrophic and heterotrophic components was achieved by the combination of long-term fumigation with dual δ¹³C and δ¹⁸O stable isotope tracers. Isotopic methods advanced from the disturbance of carbon cycle to a less invasive approach achieved by the combination of the former accomplishments. If a virtually undisturbed partitioning of the forest soil respiration components into fast and slow carbon turnover would be possible, it would enable better opportunities to study the effects of both elevated temperature and CO₂ concentration onto the balance of long-term soil carbon storage.

  Acknowledgements 

We want to thank the whole team of the WSL summer school 2009, especially Dr. Marcus Schaub, not only for making this work possible but also for a warm welcome to Zurich and for an enlightening week. We are grateful to three reviewers for their valuable comments on an earlier version of this manuscript. We also want to express our gratitude to our supervisors and institutions for their support.

  References