Carbon (δ¹³C) dynamics in agroecosystems under traditional and minimum tillage systems: a review

Keywords: tillage, no-till, zero till, C₃, C₄, soil microbial biomass, δ¹³C, C sequestration.

The preferential fixation of ¹²C over ¹³C during photosynthesis results in the δ¹³C signature of C₃ plants (Calvin cycle) being lower (more negative) than that of C₄ plants (Hatch-Slack pathway) (Smith and Epstein 1971; Farquhar et al. 1982, 1989; Lloyd and Farquhar 1994). In undisturbed systems, the soil mirrors the δ¹³C signature of the dominant vegetation. When a C₃ forest is clear-felled for continuous cultivation of C₄ maize (Zea mays), for example, the soil δ¹³C signature gradually changes towards that of the replacement C₄ vegetation.

The proportion of soil C derived from C inputs (crop or residue) can be calculated using an isotope end-member mixing model (Balesdent et al. 1987), with the assumptions of similar isotopic fractionation during humification of C₃ and C₄ plant residues (Dorodnikov et al. 2007), and a constant temporal ¹³C abundance of the input C. Growing C₄ plants on a soil that has previously been under C₃ vegetation, or vice versa, can be considered as an in-situ labelling of the organic matter incorporated into the soil and is best used in the case of several successive cultivations (Balesdent et al. 1987). Based on measured changes in C stocks and δ¹³C signatures, the turnover times of the original and replacement soil organic matter fractions can be estimated (Zach et al. 2006; Chalk et al. 2021).

Smith and Chalk (2020) reviewed the effect of tillage systems on nitrogen cycling by using ¹⁵N tracing techniques and concluded that no-till had little impact on ¹⁵N fertiliser recovery, or the accumulation of N released from the mineralisation of ‘native’ soil organic matter. However, the influence of tillage systems on the stability and turnover of soil organic matter in the soil profile was not considered. Additionally, Smith and Chalk (2020) found the effect of no-till was restricted to the surface soil layers (<0.05 m), but the ¹⁵N studies could not confirm or refute the widely-held view that soil organic N increased under no-till, and by inference, soil organic C sequestration. New insights with respect to the influence of tillage on the stability and turnover of soil organic C may be obtained by measuring changes in the soil δ¹³C signatures in systems that have undergone a C₃ to C₄ vegetation shift, where original vs replaced C pools can be identified in situ. We reviewed studies of the effects of tillage on shifts in ¹³C natural abundance (δ¹³C) implemented for varying lengths of time following the clearing of native vegetation.

In several of the papers reviewed, C in a soil layer was reported as a concentration rather than an amount, and the key soil property, bulk density, was not reported. Consequently, it was not possible to convert C concentrations into C stocks in the various soil layers. Furthermore, nominal sampling depths vary among studies, because of choice, as well as disturbance caused by tillage, and comparisons of tillage effects using mass-based concentrations and a material coordinate (defined by integrating the bulk density profile down from the soil surface; Smiles 2009) cannot be made. That is, comparisons should be made on values adjusted to an equivalent depth based on the weight of soil being the same in the comparisons (Resh et al. 2002; Zach et al. 2006). The objective of the review is to examine C dynamics under traditional and minimal tillage systems using the in situ natural ¹³C abundance technique, taking into account C stocks rather than C concentrations whenever possible. We do not review ¹³C enrichment studies because pulse labelling of plants with ¹³CO₂ results in a non-uniform label (Sangster et al. 2010). Furthermore, such studies are short-term by nature and consequently do not provide information on long-term soil C dynamics.

Plants preferentially fix ¹²C over ¹³C and accordingly have a lower δ¹³C value compared with atmospheric CO₂ (Farquhar et al. 1989; Lloyd and Farquhar 1994). The C₃ photosynthetic pathway (Calvin cycle) results in stored C depleted in δ¹³C of ~18‰ relative to the atmosphere (Farquhar et al. 1982). Since 1975, δ¹³C in the atmosphere is not constant and has steadily decreased (Francey et al. 1999; Keeling et al. 2017). Ice core records of atmospheric CO₂ show that it had a δ¹³C value of approximately –6.5‰ (1387–1900 years; Francey et al. 1999), which has declined to around –8‰ (Keeling et al. 2005, 2017; Dean et al. 2014). The decline in δ¹³CO₂ is linked to the burning of fossil fuels derived from the organic remains of organisms (mainly plants) that lived millions of years ago. The δ¹³C of oil or coal is like modern C₃ plants with values ranging from −32 to −21‰ for oil and −28 to −23‰ for coal relative to the Vienna Pee Dee Belemnite (VPDB) standard (Given 1984; Warwick and Ruppert 2016; Xu et al. 2017; Winiger et al. 2019). The annual trend in the decrease in atmospheric δ¹³C is explained by the addition of CO₂ to the atmosphere that must come from the terrestrial biosphere and/or fossil fuels. As a result of the change in the atmospheric δ¹³C, carbon (C) inputs from plants are likely to be more depleted than in the past, thus increasing the uncertainty in subsequent C cycling calculations.

In general, the average δ¹³C abundance for a range of organic materials of C₃ origin under a range of climates was –28.0 ± 1.9‰ (Balesdent et al. 1987, 1988, 1990, 1993; Natelhoffer and Fry 1988; Vitorello et al. 1989; Wanniarachchi et al. 1999; Desjardins et al. 1994; Cadisch et al. 1996; Andriulo et al. 1999; Gerzabek et al. 2001; Kaler at al. 2018), whereas the average value from C₄-derived organic sources was –12.5 ± 1.5‰ (Balesdent et al. 1987, 1990; Desjardins et al. 1994; Puget et al. 1995; Cadisch et al. 1996; Andriulo et al. 1999; Wanniarachchi et al. 1999; Hansen et al. 2004; Kristiansen et al. 2005; Fuentes et al. 2010). There are small differences in the isotopic composition of plants in different land use systems. For example, forest, –29.0 ± 1.8‰ (Natelhoffer and Fry 1988; Desjardins et al. 1994; Cadisch et al. 1996); grass and cereal, –28.4 ± 2.3‰ (Balesdent et al. 1990; Andriulo et al. 1999; Fuentes et al. 2010); organic wastes from animal housing such as manure, straw and slurries, –27.6 ± 1.5 (Gerzabek et al. 2001); and peat, –25.6‰ (Gerzabek et al. 2001).

Definitions of tillage systems are somewhat variable and were described by Smith and Chalk (2020). The systems and the degree of soil disturbance and mixing needs to be understood in order to make valid comparisons between different experiments. This is not the case in much of the literature. There is a need to use standard terminology in tillage experiments to quantify the impact of the tillage on incorporation of residues, tillage depth and soil bulk density. This will enable C and isotope stocks to be correctly determined on an area basis. Furthermore, soil C and ¹³C stocks of different treatments should be compared on an equivalent mass of soil, not on a fixed depth basis (Smiles 2009; Lam et al. 2013).

The influence of mouldboard ploughing (MP), chisel ploughing (CP), conventional tillage (CT) and no-tillage (NT), except for soil disturbance during the seeding operation are in Fig. 1. Where possible, C stocks in the surface 0.3 m are presented because the cultivation methods disturbed the soil to different depths. For example, mouldboard ploughing inverts the soil to a depth of 0.2 m and moves organic material from the surface to depth.

Fig. 1.  C stocks in the surface 0.3 m under various tillage systems and crop rotations. Data are from Vitorello et al. (1989), Desjardins et al. (1994), Balesdent et al. (1998), Andriulo et al. (1999), Collins et al. (1999), Clapp et al. (2000), Dolan et al. (2006), Jantalia et al. (2007), de Sant-Anna et al. (2017) and Fuentes et al. (2010).

There was a general decline in soil C concentrations and stocks with years of cultivation of virgin land (Lefroy et al. 1993; Jolivet et al. 1997; Andriulo et al. 1999), which can either be described by a single or double exponential decay model. Of the 10 studies that report C stocks (Vitorello et al. 1989; Desjardins et al. 1994; Balesdent et al. 1998; Andriulo et al. 1999; Collins et al. 1999; Clapp et al. 2000; Dolan et al. 2006; Jantalia et al. 2007; Fuentes et al. 2010; de Sant-Anna et al. 2017), time series data are lacking, especially for long-term cultivation studies. Carbon stocks showed a reduction in C when comparisons were made on the same mass of soil equal to that under 0.3 m of native vegetation (Jantalia et al. 2007). Whilst no-till systems reduced the C loss compared to conventional tillage, the largest losses occurred with mouldboard ploughing and/or twice-yearly tillage (Jantalia et al. 2007). Differences in the C stocks because of land-use change and no-tillage were predominately restricted to the surface soil, whereas mouldboard ploughing resulted in decreases in the C stock in the surface 0.25 m (Clapp et al. 2000; Layese et al. 2002; Huggins et al. 2007; Loss et al. 2012).

Ploughing and harrowing in annual conventional tillage systems promoted greater loss of soil C compared to no-till (Loss et al. 2016). In a 13-year continuous maize study, soil C stocks were found to be sensitive to tillage and harvest of the aboveground plant material (Clapp et al. 2000). Continuous maize grown after a low input mixture of lucerne (Medicago sativa) showed a small decrease in soil C in the annually tilled (MP, CP) treatments. In the no-till treatment, C stocks in the surface plough layer (0.15 m) remained unchanged when the stover was harvested but increased when the stover was returned (Clapp et al. 2000). In contrast, C stocks were unaffected by tillage (non-till compared to MP and CP) when summed to 0.5 m (Dolan et al. 2006), nor did tillage system affect the mineralisation of soil organic C (Liu et al. 2015).

Tillage and residue retention cause stratification in C stocks that are dependent on tillage type (depth) with the highest C concentrations and stocks found in the surface (<0.05 m) under NT (Balesdent et al. 1990; Andriulo et al. 1999; Machado et al. 2003; Allmaras et al. 2004; Dolan et al. 2006; Huggins et al. 2007). For comparison of the effect of tillage on soil C stocks, comparisons need to be made by sampling the tilled layer and ensuring that the cumulative mass of soil per unit area of cross-section is the same between the different treatments. Furthermore, soil C is responsive to the quantity of organic C applied, including the recycling of animal manures (Gerzabek et al. 2001), which in turn is influenced by climate, crop selection, agronomic management, fertilisation and residue management.

The δ¹³C signatures of soil organic C under C₃, C₄ and mixed plant systems show end members ranging from –28‰ under C₃ vegetation to –15‰ from C₄ vegetation (Fig. 2). Mixed plant communities have δ¹³C values that lie between these two end member signatures. In general, the δ¹³C composition of the soil C tended to decrease with soil depth suggesting that the soil C at depth was predominately derived from C₃ plant material. The δ¹³C values were lower than –12‰ in land use systems that supported C₄ plants, which is the value one would anticipate if the soil C was derived solely from C₄ plants. It should be noted that in all the papers reviewed, the C was determined on soil samples that had carbonates (CO₃) removed, thereby ensuring the δ¹³C signature is that of organic C. Systems with a mixture of C₃ and C₄ plants have intermediate δ¹³C values ranging from –17‰ to –22‰. The δ¹³C values decreased for higher proportions of C₄ crops in the rotation, and for the length of time the C₄ crops were grown.

Fig. 2.  Relative isotopic composition (δ¹³C) of the soil C under C₃, C₃/C₄ and C₄ plant systems. Data are from Balesdent et al. (1987, 1988, 1990, 1993), Natelhoffer and Fry (1988), Vitorello et al. (1989), Bonde et al. (1992), Desjardins et al. (1994), Gregorich et al. (1995), Cadisch et al. (1996), Andriulo et al. (1999), Wanniarachchi et al. (1999), Shang and Tiessen (2000), Gerzabek et al. (2001), Roscoe and Burrman (2003), Sisti et al. (2004), Antil et al. (2005), Jantalia et al. (2007), Fuentes et al. (2010), de Sant-Anna et al. (2017) and Muñoz-Romero et al. (2017).

Although not shown in Fig. 2, the soil organic C declined from the surface to 1.0 m depth, while the turnover period of soil C below the cultivation depth increased with tillage. The δ¹³C signature integrated for the profile reflects more closely the current vegetation under CT compared to NT (Machado et al. 2003), with significantly ‘newer’ C under CT compared to NT, which in CT extended below the cultivation depth. The difference between the two tillage systems reflects the retention of the fresh C in the surface layer (0.05 m), which declines rapidly with depth under NT. Under CT, the fresh C is distributed through the plough layer and declined below the plough layer. Tillage had no significant effect on the total amount of C present to 0.4m, but the effect was significant only when the surface soil (defined as 0.2 m) was considered (Machado et al. 2003). Furthermore, the result remained to be corrected after adjusting for different bulk densities and using the same mass of soil.

Stratification in soil C under the different tillage regimes is highlighted by δ¹³C values, with the least negative values in the very surface soil. δ¹³C values decreased under CP or MP, which mixes plant residues throughout the depth of cultivation (Table 1). The stratification was most pronounced in the studies of Balesdent et al. (1990) and Layese et al. (2002) and to a lesser extent in Wanniarachchi et al. (1999) as shown in Fig. 3. The highest δ¹³C values were reported by Balesdent et al. (1990) and the value (–16.4‰) was closer to that of the maize residue compared to the values under CT or superficial tillage. These results emphasise the need to sample to depth and make comparisons on adjusted soil C values based on the same mass of soil, if the effect of tillage practices on soil C stocks is to be accurately assessed, as discussed by Jantalia et al. (2007).

Table 1.  Effect of years of tillage and depth under maize rotation on δ¹³C signatures of the soil organic C

Fig. 3.  The isotopic stratification of δ¹³C (‰) in the surface 0.3 m of soil under no-till and different cultivation methods.

Whilst there was no large change in the C stocks when summed to a constant mass of soil (Balesdent et al. 1990; Andriulo et al. 1999; Clapp et al. 2000; Allmaras et al. 2004; Dolan et al. 2006), the shifts in the δ¹³C signature indicated significant exchange of ‘new’ C for the original (old) C (Fig. 3; Zang et al. 2018). The data highlight the dynamic nature of the soil C and enable the contribution of the recently derived C to the total C stock in a soil layer and profile to be estimated. Furthermore, assuming soil organic C decomposition follows first order kinetics, the mean residence time of the soil C can be estimated from the mass balance and isotopic composition data (Balesdent and Mariotti 1996; Zang et al. 2018; Chalk et al. 2021).

Mean residence time (MRT) is the average length of time that an element such as C spends in a given pool that is not uniform in composition. Each component or fraction of the pool changes at a different rate. The turnover of C is calculated using either a first order decay model (Six et al. 1998; Six and Jastrow 2002) or a double exponential decay model representing a rapid and then a slow change in the δ¹³C signature of soil organic matter pools (Arrouays et al. 1995; Jolivet et al. 1997). The MRT is calculated as the reciprocal of the decomposition rate (Six and Jastrow 2002).

The effect of tillage on the MRT of C at soil depth intervals is summarised in Table 2. In general, MRTs increase with depth. In the surface soil, MRT values are quite variable, ranging from <10 to 100 years under cultivation systems. Conventional tillage and MP tended to have shorter MRT values compared to CP. NT systems had a similar spread in MRT values with the average value of 53 years in the surface 0.3 m of soil. The MRTs under NT were larger than those under MP and CT. Apparent difference in MRT among tillage systems integrate several variables, including intrinsic soil properties (e.g. pH and texture), the amounts of residues produced and incorporated into the soil and the method of calculation (Table 2).

Table 2.  Effect of tillage method on mean residence times of C in soil depth intervals (blank cells denote no data)
MP, mouldboard ploughing; CT, conventional tillage; CP, Chisel ploughing; NT, no-tillage; MRT, mean residence time; 1^st, first order exponential decay; GCT, gross C turnover , which is soil organic C stock expressed as equivalent soil mass (g m^–2)/annual C input (g m^–2)

Change in microbial biomass is regarded as an early indicator of tillage and crop rotation effects on soil organic matter compared with total organic C or N measurements and is a potential index of C sequestration (Powlson and Jenkinson 1981; Carter 1986; Powlson et al. 1987; Balota et al. 1998; Sá et al. 2001; Potthoff et al. 2003). Furthermore, microbial biomass can be a valuable tool for understanding changes in soil carbon, nutrient availability, aggregate stability, and quality (Smith and Paul 1990; Doran and Parkin 1994; Brookes 1995; Sparling 1997; Dalal 1998).

It might be anticipated that tillage treatments such as MP or CT that incorporate residues may have a greater effect on MB than NT, where residues are largely left on the soil surface. However, the limited data available allow few definitive comparisons (Table 3). Generally, there were only small differences in MB-C under maize rotation between NT and MP treatments (Rochette et al. 1999), between NT and CT (Ryan and Aravena 1994, Ryan et al. 1995) and between ST, RT and MP (Angers et al. 1995).

Table 3.  Effect of tillage regime on soil and microbial biomass (MB) C concentrations and their δ¹³C signatures (blank cells denote no data)
ST, surface tillage; RT, ridge tillage; MP, mouldboard ploughing; NT, no-tillage; CT, conventional tillage; control, repacked soil cylinders without maize, or with maize after 35 or 70 days

However, after 4 to 6 years of continuous maize, a larger percentage of the total microbial biomass was derived from C₄-C under CT compared to NT management (Ryan and Aravena 1994; Ryan et al. 1995) (Table 3), which may reflect the degree of mixing of residues and soil under different tillage regimes. Microbial biomass varies seasonally according to crop growth, with an annual minimum before planting and a maximum during the growing season (Ryan et al. 1995). We therefore conclude that MB-C per se is a poor indicator of long-term changes in soil C under different tillage regimes, representing only 1–2% of soil organic C.

Tillage intensity generally increase the turnover rate of soil organic matter and decreases soil aggregate stability. The reduced aggregate stability and increased soil organic C decomposition is a function of the physical disturbance (Six et al. 1998). The turnover rate of the soil organic matter in the different particle size fractions has the potential to provide a basis for the partitioning the larger soil organic C pool into smaller functional pools. The changes in the soil δ¹³C signature before and after cultivation of either C₄ or C₃ plants (Δ¹³C) for different particle size fractions are in Fig. 4, where plant sequences of C₄ to C₃ or C₃ to C₄ are shown in the upper and lower panels, respectively. Changing from C₃ to C₄ vegetation cover, in general resulted in a decrease δ¹³C values (positive shift in Δ¹³C) and vice versa for a C₄ to C₃ vegetation change. In most studies, the largest change in the Δ¹³C values was observed in the coarse sand fraction (200–2000 µm), whereas the smallest change occurred in the clay fraction. Collectively, these data confirm the slower turnover rate of soil organic C in clay (as evident by the lower replacement of the original C), the fastest turnover associated with coarse and fine sand, and intermediate values for silt fractions. Carbon in macroaggregates in general had a mean residence time 42 ± 18 years, whereas microaggregate-associated C has a mean residence time of 206 ± 95 years (Six and Jastrow 2002).

Fig. 4.  Effect of particle size fractions on the change in the δ¹³C signatures of organic C for different soil depths. Data are from Balesdent et al. (1988), Bonde et al. (1992), Desjardins et al. (1994), Gregorich et al. (1995), Puget et al. (1995), Jastrow et al. (1996), Six et al. (1998, 1999, 2002), Shang and Tiessen (2000), Gerzabek et al. (2001), Roscoe et al. (2001), Sá et al. (2001), Six and Jastrow (2002), Antil et al. (2005), Zach et al. (2006), Murage et al. (2007), Liang et al. (2014) and de Sant-Anna et al. (2017).

Comparison of CT (including mouldboard ploughing) to NT showed inconsistent results of the effect of tillage on C in the different particle size fractions (Fig. 5). In contrast, Six et al. (1999) reported greater C accumulation in NT compared to CT because of the slower turnover of C in macroaggregates under NT. According to Denef et al. (2007), the formation of macroaggregates is important to the long-term C stabilisation in NT systems. Formation of macroaggregates serves as an early indicator for potential total SOC sequestration caused by the greater stability and slowed turnover of C in macroaggregates. More frequent tillage may lead to the formation of small aggregates rich in labile C (Urbanek et al. 2011). The δ¹³C signature in the surface soil (<0.01 m) suggests that C₄-C derived from maize is the dominant fraction of C in agroecosystems, accounting for up to 80% of the total soil organic C concentration (Jha et al. 2017).

Fig. 5.  δ¹³C (‰) signatures of particle size fractions with soil depth for conventional (CT) and no-till (NT) systems (derived from a subset of data shown in Fig. 4).

The light fraction of soil organic C is dynamic and responds to short-term shifts in organic C storage. This pool has been shown to contain charcoal or a char component formed by the incomplete burning of organic residues (Skjemstad et al. 1994; Murage et al. 2007). Jantalia et al. (2007) estimated that 40% of soil C was charcoal in a soybean (Glycine max) rotation on an Oxisol in the Cerrado region of Brazil. The char or charcoal is generally composed of C from the original vegetation, and consequently may lead to misinterpretation of C turnover. For example, Murage et al. (2007) showed that the turnover of the light fraction of organic matter appears to be 2.5 times slower compared to the same pool without charcoal. Given that most of the cropping systems involve burning of residues, the char component should be removed from the light fraction when determining the changes to this pool. In a comparison of a soil under forest and under maize cultivation, the light fraction of soil organic matter had a shift in the δ¹³C value (Δ¹³C) of +10‰, indicating rapid mineralisation of C in the light fraction compared to the other particle size fractions (Gregorich et al. 1995). However, sample char or charcoal was not determined, and thus the turnover of the light fraction may have been overestimated.

The soil particulate organic matter fraction (POM) has been reported to be affected by cultivation of a forest soil (Balesdent et al. 1998), with the change being reached within the first few years after forest clearing. Six et al. (1999) suggest that free POM is mostly affected by plant C input rates. Generally, coarse and fine POM increased under NT, with C derived from C₃ plants being mainly associated with the finer compared to the coarse particles (Liang et al. 2014). The turnover time of POM increased with NT compared to CT, especially in the finer POM (53–250 µm) fraction, which largely controls aggregate stability (Six et al. 1998, 2002).

The potential of C sequestration in soil to off-set the anthropogenic increase in atmospheric CO₂ has been discussed by several authors (e.g. Luo et al. 2011; Lam et al. 2013; Amundson and Biardeau 2018; Sun et al. 2020). The studies reported the impact of different tillage practices and rotations on changes in soil C stocks in different soil layers. Generally, the accumulation of C is the largest in the surface layer and declines exponentially with depth. Whilst the studies all use total C stocks, where possible, there is an opportunity to use knowledge of changes in the δ¹³C isotopic signatures of soil C to confirm estimates of net and gross C changes that are likely to occur over longer times scales. The shifts in δ¹³C signatures provide valuable information on the residence times of C in soil and assist in confirming that predictions of tillage and rotations on C storage are reasonable. Acton et al. (2013) reported a framework for soil C and isotopic mass balance modelling for forests that includes multiple C pools, including the vertical distribution and turnover of soil C. There is a need to include similar routines in agricultural system models used to evaluate the long-term impact of reduced tillage on soil C sequestration.

Generally, the literature shows that tillage has little impact on the accumulation on soil organic C when the comparisons are made on the same volume of soil that the different tillage systems influence. Ideally the comparisons should be made using mass of soil as the x coordinate to allow for differences in bulk density. Tillage, however, does cause stratification in soil C with the accumulation of the largest amounts of C in the surface layers under no-till systems. In contrast, the use of mouldboard ploughing that inverts the soil to depths of 20–30 cm distributes the new C throughout this layer. The relative ¹³C signature of the carbon is an ideal natural in situ labelling technique that confirms the long-term fate of the C when it is incorporated into the soil. The technique is limited to systems where there is a change in the δ¹³C signature of the newly applied land use system from that developed under the previous system. The shift in the δ ¹³C signature is a powerful tool for quantifying the fate and turnover times of the different sources of C in the soil. These findings are consistent with the previous review by Smith and Chalk (2020) that showed tillage having little effect on soil N dynamics.

Soil C stocks are the balance between new C inputs in the form of crop residues and root biomass and the decomposition of the new and older soil C. Net primary productivity of the plant system regulates potential C input which in turn is constrained by climate, nutrients and agronomic management.

The authors declare no conflict of interest.

This research did not receive any specific funding.