Microbial processes and carbon-isotope fractionation in tropical and temperate grassland soils



1. The carbon content and δ13C value of soil organic carbon (SOC), microbial biomass (Cmic) and respired CO2 were measured in a range of grassland soils from tropical and temperate biomes to determine if isotope effect of microbial degradation can induce a shift in isotope composition of SOC and CO2. The soil from a depth of 0–2 cm was analysed. Cmic was measured using the chloroform fumigation extraction method, while CO2 was measured in a closed system after 3 and 10 days of incubation. Two soils, temperate and tropical, were used for a long-term experiment, in which measurements were performed after 3, 10 and 40 days of incubation.

2. SOC and Cmic decrease exponentially with increasing mean annual temperature. Cmic decreases more slowly than SOC, resulting in a higher proportion of Cmic in the SOC of tropical soils relative to temperate soils.

3. The δ13C value of Cmic and respired CO2 reflects gross changes in the δ13C value of SOC in the corresponding sample. On average, Cmic is 13C-enriched by c. 2‰ compared with SOC, while respired CO2 is 13C-depleted by c. 2·2‰ compared with Cmic. Thus, the observed 13C-enrichment in Cmic is balanced by a corresponding 13C-depletion in respired CO2 resulting in the δ13C value of respired CO2 being approximately similar to the δ13C of SOC.

4. The isotope effect of microbial degradation is of importance in soil. It can be induced by selective utilization of SOC and isotope discrimination during metabolism. Metabolic isotopic discrimination is dependent on the growth stage of the soil microbial population.


In the broadest terms, the isotopic composition of soil organic carbon (SOC) reflects that of the local plant cover (Deines 1980). However, the δ13C value of SOC differs by −6·1‰ to +4·4‰ compared to that of C derived from local vegetation (Rightmire & Hanshaw 1973; Stout, Goh & Rafter 1981; Nadelhoffer & Fry 1988; Mellilo et al. 1989; Nakamura, Takai & Wada 1990; von Fischer & Tieszen 1995). The difference between δ13C of SOC and that of soil CO2 was found to vary from −3·2‰ to +2·1‰ (Readon, Allison & Fritz 1979; Dörr & Münnich 1980; Parada, Long & Davis 1983; Mellilo et al. 1989; Nakamura, Takai & Wada 1990; Hesterberg & Siegenthaler 1991; von Fischer & Tieszen 1995; Dudziak & Halas 1996a,b). This wide range of isotopic shifts is difficult to reconcile with the commonly held view that little fractionation accompanies heterotrophic metabolism by soil micro-organisms, i.e. the isotope composition of the SOC should equal that from plant input (Fritz et al. 1978; Cerling 1984; Quade, Cerling & Bowman 1989). The variability in the composition of soil microbial biomass and the fact that micro-organisms use various chemical fractions at different rates suggest that the difference between the δ13C of plant material, SOC and the δ13C of soil microbial biomass can potentially be larger than the value of 1–2‰ established for many heterotrophs (DeNiro & Epstein 1978; Tieszen & Boutton 1988; Hullar et al. 1996).

One source of the observed difference between δ13C of SOC, CO2 and plant input could be isotope effects which occur during microbial degradation of plant material (Deines 1980; Fry & Sherr 1988; Balesdent, Girardin & Mariotti 1993; Ågren, Bosatta & Balesdent 1996). This can be induced by the selective use of chemical compounds having δ13C values deviating from that of plant biomass and by kinetic discrimination during microbial metabolism. The δ13C value of both heterotrophic and autotrophic biomass is the mass-averaged δ13C value of diverse biomolecules. Relative to biomass C, secondary metabolites (aromatics, proteins, isoprenoids) are usually 13C-depleted, while primary products are 13C-enriched (Deines 1980; Schmidt & Gleixner 1998). During metabolism, catabolic reactions prefer the molecules which have less δ13C, while those with more δ13C are involved in biomass production (Blair et al. 1985; Schmidt & Gleixner 1998). δ13C of consumed material would be equal to δ13C of microbial biomass and respired CO2, if isotope effect of microbial degradation is induced only by selective use. If discrimination during metabolism is important, then consumed material would have less δ13C related to microbial biomass δ13C, and more, as related to the δ13C of respired CO2.

The importance of the isotopic effect of microbial degradation was determined in 21 grassland soils, by analysing for the δ13C content of SOC, soil microbial biomass (Cmic), and CO2 released by aerobic microbial respiration.

Materials and methods


Soil was sampled from grasslands along a temperature gradient, from a mean annual temperature of 25 °C in northern Australia to 7 °C in southern Australia (Fig. 1). Grasslands with mean annual temperatures above 21 °C are tropical grasslands with a dominance of C4 plants, while decreasing temperatures result in an increasing dominance of C3 grasses in more temperate regions. Soil samples (Table 1) were collected from the top 2 cm of soil at each sample site three to five locations within a 10–20 m radius circle and stored field moist at 4 °C prior to analysis.

Figure 1.

Location of Australian grassland soil samples used in this study. Soils are marked Ausg 1 to Ausg 37, from sites in Northern Australia to sites in Southern Australia.

Table 1.  Physico-chemical and biological properties of the soils analysed for this study. Respiration rate after 3 (CO2I), and 10 days (CO2II) of incubation is expressed on an SOC basis (µg CO2.g Corg−1.h−1). Sites Ausg 1–Ausg 21 represent tropical biomes with C4 plant cover and sites Ausg 23–Ausg 37 temperate biomes with a predominance of C3 plants
       Respir. rate Δ
SiteRainfall mmTemp °CpHH2OC/NSOC %Cmicµg C g−1CO2ICO2IIδ13C of SOC ‰SOC/CmicCmic/CO2CO2I/CO2II
Ausg 1220257·515·1 0·34 156238100−14·6−3·13·27−0·53
Ausg 4a220257·212·5 1·22 390104 38−16·5−1·95·72−1·94
Ausg 6220257·120·2 0·35 149203 66−15·3−2·73·25 0·71
Ausg 8260257·312·6 0·29 163279 76−15·3−3·74·06 0·20
Ausg 14220246·613·6 0·33 183291103−15·2−1·43·29 1·70
Ausg 18290216·212·5 1·41 758123 67−16·4−1·31·83−0·81
Ausg 19290217·710·5 0·72 498143 63−15·5−0·80·38−0·51
Ausg 21290217·012·8 0·50 349206112−16 1·40·71 1·63
Ausg 23210207·79·2 0·90 512134 61−20·6−2·40·92−0·31
Ausg 25270185·313·4 1·20 290133 96−22·1 1·32·98−1·03
Ausg 48240165·714·8 4·08 427106 55−25−1·82·36 0·10
Ausg 50240167·912·7 1·63 485179100−26·8−2·62·57−0·10
Ausg 52240165·712·3 2·46 580 81 30−27·5−1·91·33−2·76
Ausg 46240156·912·6 4·56 676 64 29−26·6−2·11·03 0·31
Ausg 27500116·112·0 2·85 615102 61−24·8−0·91·33 0·10
Ausg 29500116·110·2 3·17 317113 92−24·8−3·23·18−0·21
Ausg 32640 85·719·5 4·811107 64 55−27·1−3·32·77−0·10
Ausg 33640 74·721·111·49 568 33 18−26·8−3·32·26−0·31
Ausg 35640 74·113·512·19 849 24 14−26·1−3·32·15−1·13
Ausg 36640 74·714·812·871554  8  7−26·1−3·02·77−0·62
Ausg 37640 77·617·0 2·81 910153 67−26·7−1·70·1−0·10

The soil was analysed for SOC and the δ13C value of SOC (Bird & Pousai 1997). Cmic and microbial respiration (CO2), and the δ13C value of each were analysed using the same samples. Before Cmic and CO2 were measured, the soil was sieved (2·5 mm) and moisture adjusted to 55–65% WHC (water holding capacity, ISO 11274 1994) when necessary. Two soils, Ausg 19 and Ausg 46, were kept at constant moisture and temperature and used for a long-term experiment. Measurements (Cmic, CO2) were performed after 3, 10 and 40 days of incubation.

Microbial respiration

Two replicated samples (50 g each) were incubated in closed vials (600 ml) at 20 °C for 10 days. CO2 production was determined in two periods; from 1 to 3 days and 7–10 days. CO2 was trapped in bicarbonate-free 1 N NaOH then released by acid addition and purified cryogenically and the amount of purified CO2 was determined manometrically. The vessels containing the NaOH solution were handled in a N2 atmosphere. Deviations between the replicates did not exceed a coefficient of variation of 8%. The isotopic composition of the CO2 was corrected for the isotopic composition of the ambient air, which was present in the incubation vials at the beginning of measurement, using a simple two-component mixture model (Hesterberg & Siegenthaler 1991).

The δ13C value measured by the absorption method might be affected by isotopic fractionation if the CO2 is not completely absorbed into the hydroxide solution. We determined the minimum incubation time necessary to avoid the kinetic discrimination effect using an addition of CO2 of known volume and δ13C. 94%, 97% and 100% of added CO2 was trapped into NaOH solution in 15, 30 and 60 min after CO2 addition, respectively. No fractionation was found after 60 min. The incubation period in the present experiments was at least 72 h. Therefore, the kinetic effect of absorption was negligible.

Microbial biomass (cmic)

Cmic was measured after 10 days of incubation using the chloroform fumigation extraction method (Vance, Brookes & Jenkinson 1987; Ryan & Aravena 1994). Soil from each incubation vial was divided into two portions. One part was exposed to ethanol-free CHCl3 for 24 h, after which the fumigant was removed and the soil was extracted with 0·5 m K2SO4 (1/4 w/v ratio, 30 min, end-over-end shaker). The second aliquot of the soil, a non-fumigated control, was extracted under the same conditions but without the fumigation step. One to two grams of the freeze-dried sulphate extract was combusted at 900 °C with CuO and silver wire in evacuated sealed quartz tubes (Boutton et al. 1983). The CO2 released by combustion was purified cryogenically. Deviations in the amount of carbon between duplicates did not exceed a coefficient of variation of 6%.

Isotopic composition

Isotopic composition of CO2 and Cmic was measured using a Finnigan MAT-251 mass spectrometer. All measurements were performed in duplicate and the results are reported as the difference in parts per thousand (per mil; ‰) from the defined international V-PDB standard. Precision of the 13C analyses of standards was ± 0·1‰ and the standard deviation of the replicate samples did not exceed 0·5‰. δ13C of Cmic13Cmic) was estimated as the δ13C of the C extracted from fumigated soil (δ13Cf) in excess of that extracted from the non-fumigated control sample (δ13Cnf):

image(eqn 1)

Estimation of the δ13C value of Cin (organic C consumed by Cmic) was based on the supposition that the δ13C value of Cmic is the result of a mass balance between inputs (min) and outputs (mout) to the microbial cell (Hayes 1993). For a microbial cell containing Cmic moles of C, it can be written that:


McGill et al. (1981) showed that c. 40% of Cin is converted to Cmic, meaning that c. 60% is released as CO2. Thus, the amount of δ13C in Cin, used in aerobic microbial metabolism in soil, was estimated from the δ13C values of Cmic and CO2 as:


Values of the isotopic composition of δ13Cin and δ13CCO2 depend on the isotopic effects accompanying biosynthesis and catabolism, respectively. δ13C does not change dramatically if the conversion efficiency of substrate C is higher or lower than the suggested value of 0·4, staying within a range of only 0·5‰. For example, if the efficiency of microbial conversion is decreased to a value of 0·3 or increased to 0·5 then the proportion of metabolized C/respired C is changed from 0·4/0·6 to 0·3/0·7 and to 0·5/0·5, respectively.

Difference between isotopic composition

The difference between two isotopic compositions was estimated using the fractionation factor, Δ (Farquhar & Richards 1984). Four such ‘fractionations’ were used in this study: (1) the difference between the δ13C value of Cin and the δ13C value of Cmic; ΔCin/Cmic; (2) the difference between the δ13C value of Cin and the δ13C value of the CO2 respired by Cmic; ΔCin/CO2; (3) the difference between the δ13C value of SOC and the δ13C value of Cin; ΔSOC/Cin; (4) the difference between the δ13C value of CO2 respired after 3 days (CO2I) and 10 days (CO2II) of incubation; ΔCO2(I)/CO2(II)


The results obtained from the 21 samples analysed for this study are provided in Table 1. SOC and Cmic in the soil decreased exponentially with increasing mean annual temperature (Fig. 2). However, Cmic decreased more slowly than that of SOC, suggesting a higher proportion of Cmic in the SOC in tropical compared to temperate grasslands (Fig. 3).

Figure 2.

Relationship between SOC, Cmic and mean annual temperature.

Figure 3.

Relationship between the ratio of Cmic-to-SOC and mean annual temperature.

The δ13C of Cmic reflected that of the corresponding SOC. On average, Cmic was 13C-enriched by −2‰ (Fig. 4). The difference between Cmic and SOC, ΔSOC/Cmic, ranged from −3·7‰ to +1·4‰ (Table 1) with no discernible relationship to SOC, Cmic, SOC-to-Cmic ratio or biome type (i.e. tropical C4 or temperate C3 grasslands).

Figure 4.

Relationship between the δ13C value of Cmic and respired CO2 to the δ13C value of SOC.

CO2 respired by Cmic after 10 days (CO2II) was 13C-depleted when compared to the corresponding δ13C-value of Cmic (Fig. 4). The mean value of the fractionation factor, ΔCmic/CO2, was +2·2‰ but ranged between +0·1‰ and +5·7‰.

Cmic was 13C-enriched relative to Cin, with ΔCin/Cmic ranging from −0·1‰ to −3·4‰. Respired CO2 was depleted, with ΔCin/CO2 ranging from + 0·04‰ to + 2·3‰. δ13C of Cin differed substantially from δ13C of SOC (Fig. 5a); the values of ΔSOC/Cin varied from + 3·1‰ to − 2·0‰.

Figure 5.

The magnitude of the difference in isotope composition between (a) C consumed by micro-organisms (Cin) and SOC (ΔSOC/Cmic, ‰), (b) respired CO2 and CinCin/CO2, ‰) and (c) Cmic and CinCin/Cmic, ‰). Soils are ordered from the highest to the lowest mean annual temperature (see text for details).

CO2II differed from CO2I (Table 1), with the values of ΔCO2(I)/CO2(II) ranging from +1·7‰ to −1·9‰. No relationship of δ13C of CO2 to respiration rate, SOC, Cmic, or to environmental factors was discernible.

Over a 40 day incubation period, Cmic in the Ausg 19 and Ausg 46 soils became 13C-depleted, CO2 become 13C-enriched, while Cin did not change. Cmic and respiration rate decreased over the same period (Table 2).

Table 2.  The amounts (mean ± standard deviation, n = 3) and δ13C values of microbial biomass (Cmic) and respired CO2 after 3, 10 and 40 days of incubation for a tropical (Ausg 19) and a temperate (Ausg 46) soil kept at constant moisture content and temperature. δ13C of actually consumed C (Cin) was estimated using equation 3 (see text for details): ND, not defined
  Days of incubation
Soil(δ13C of SOC) 31040
Ausg 19 (−16·4‰)Cmic (µg C g−1)ND498 ± 17450 ± 26
 δ13C − Cmic (‰)ND−15·4−18·7
 CO2(mg CO2 − C g Cmic−1 h−1)1·03 ± 0·040·45 ± 0·030·36 ± 0·03
 δ13C − CO2 (‰)−15·5−15·1−13·2
 δ13C − Cin (‰)ND−15·2−15·4
Ausg 46 (−26·6‰)Cmic (µg C g−1)ND676 ± 24580 ± 31
 δ13C − Cmic (‰)ND−24·6−27·5
 CO2(mg CO2 − C g Cmic−1 h−1)2·92 ± 0·10 1·31 ± 0·01 0·52 ± 0·03
 δ13C − CO2 (‰)−26·6−26·9−24·5
 δ13C − Cin (‰)ND−26·0−25·7


The proportion of microbial biomass is larger in tropical soils than in temperate soils, based on the observation that the total SOC pool decreased in size with increasing mean annual temperature at a faster rate than Cmic. Thus, the turnover rate of organic material may be quicker in tropical soils. Also SOC levels in tropical grassland soils will respond more quickly to changes in carbon input rates than temperate grassland soils. These results are consistent with those of other studies. Wedin et al. (1995) found more rapid turnover of SOC under tropical C4 grasses than under C3 grasses. The mean residence time of the labile SOC in temperate soils is considered to be in the range of 15–75 years, while it is only 4–45 years in tropical soils (Jenkinson & Rayner 1977; Bird, Chivas & Head 1996; Hsieh 1996).

The importance of the isotopic effect of microbial degradation of SOC cannot be assessed directly from the fractionation factor, ΔSOC/Cmic, because only a small part of the SOC can be used by micro-organisms as Cin. Therefore, Cin was estimated, as were the isotopic shifts resulting from isotopic discrimination during microbial metabolism by using the fractionation factors ΔCin/Cmic and ΔCin/CO2 (Fig. 5). Cmic was 13C-enriched related to Cin, indicating isotope discrimination during biosynthesis of new biomass. The preferential use of 13C for biosynthesis was approximately balanced by the depletion of CO2, confirming that catabolic reactions prefer ‘light’ isotopes (Blair et al. 1985; Schmidt & Gleixner 1998). Cin was enriched related to SOC in most cases, indicating that 13C-enriched compounds are preferentially used by Cmic. Such a selective use induces a more rapid loss of 13C than 12C during decomposition of plant detritus (Benner et al. 1987; Ågren, Bosatta & Balesdent 1996). In those cases when Cin was 13C-depleted relative to SOC, three of the four soils (AUSG 14, 21 and 25) were from mixed C3/C4 grasslands; under such circumstances C3 carbon may be preferentially used over C4 carbon. In addition, in these samples, the magnitude of ΔCin/Cmic and ΔCin/CO2 decreased slightly with decreasing temperature, possibly owing to the presence of both C3 and C4 derived SOC. However, the relationship is unclear at intermediate temperatures.

Many studies have demonstrated that the δ13C value of SOC increases with both depth in the soil profile and decreasing particle size (e.g. Bird & Pousai 1997). The results from this study suggest that at least part of this increase may be owing to the preservation/stabilization of a proportion of microbially processed carbon, often in association with the fine mineral fraction. This carbon has a δ13C value which is higher than the carbon input to the soil from local vegetation, owing to the selective utilization of δ13C-rich organic compounds and isotopic fractionation accompanying heterotrophic metabolism.

The δ13C value of respired CO2 differed after 3 and 10 days of incubation (ΔCO2I/CO2II; Table 1). In addition, the δ13C values of CO2 and Cmic changed during long-term incubation of soils (Table 2). In the early stage of incubation, when micro-organisms were growing and microbial activity was higher, Cmic was more 13C-enriched, suggesting the formation of 13C-rich proteinaceous material. However, Cmic and respiration rates decreased during prolonged incubation. As a result, Cmic became 13C-depleted relative to the earlier period of incubation. The shift in the δ13C of Cmic was balanced by an opposite shift in δ13C of respired CO2. Therefore, heterotrophic shifts in the δ13C values of Cmic and respired CO2 can be influenced also by the growth stage of the microbial population. Micro-organisms synthesize cell compounds with different δ13C values at various growth stages. Growing cells synthesize mainly proteinaceous compounds rich in 13C (Deines 1980) while growth-limited cells produce mainly storage material which is depleted in 13C. Coffin et al. (1989) observed that the δ13C value of bacterial cells grown in sea water decreased with corresponding increase in the C/N ratio of the cells and with a prolonged incubation. This implies that the δ13C value of the storage material which contributes to the increase in the C/N ratio of a bacterial cell is 13C-depleted.

The results presented in this study demonstrate that: (1) there is a faster turnover of SOC in tropical grassland soils, with a higher proportion of microbial biomass in the SOC of tropical grassland soils relative to temperate grassland soils; (2) on average, the isotope shift of Cmic with respect to SOC is balanced by an inverse isotope shift of the δ13C value of respired CO2 resulting in the δ13C value of respired CO2 being approximately similar to the δ13C of SOC; (3) an isotope effect of microbial degradation of organic material can induce a shift in isotopic composition of SOC. It can be induced by both the selective use of organic compounds and isotope discrimination during microbial metabolism. The degree of isotopic discrimination during metabolism is dependent on the growth stage of the microbial population.


The authors acknowledge the Australian Research Council for a Queen Elizabeth II Fellowship to M.I.B.; Grant Agency of Academy of Sciences of the Czech Republic (A6066901); Joan Cowley and Joe Cali for assistance with sample preparation and mass spectrometry measurements and Dr Keith Edwards for English revision.

Received 3 December 1998; rrevised 12 August 1999;accepted 18 August 1999