The effect of grazing management and legumes on soil organic carbon stocks in pastoral systems relevant to the Australian Carbon Credit Unit Scheme: a critical review

Search terms and article selection approach

The search terms ‘rotational grazing’, ‘grazing management’, and ‘legumes’ were used in combination with other search terms to find research articles on the impact of grazing management and legume inclusion in Australian pastoral systems. The terms were selected because they are both eligible management activities under the 2021 Estimating Soil Organic Carbon Sequestration Using Measurement and Models method, and to establish the level of evidence available, especially evidence that was conducted in the field and established SOC stocks according to current method guidelines (Federal Register of Legislation 2021).

The literature critique was undertaken in two stages. The first stage used the Scopus database to search for relevant literature. The first search string query entered into the ‘Search documents’ bar was ‘soil organic carbon’ AND ‘rotational grazing’ OR ‘grazing management’. The second search string query was ‘soil organic carbon’ AND ‘legumes’ AND NOT ‘crop’. The scope of the literature search included no date or country restrictions and was limited to published journal articles. These two searches returned 27 and 30 articles, respectively. The second stage of the review was to undertake a manual review of the 57 articles, and apply inclusion or exclusion criteria to decide on the final set of eligible articles. The inclusion criteria for the review were: the study was conducted in Australia; under a pastoral land use with a comparison of soil carbon stocks to at least 0.3 m; between set stocking and at least one other type of grazing management; or comparing pasture management with and without legumes. The exclusion criteria were: the study was conducted outside Australia; using pot trials (in Australia); soil carbon stocks not calculated; soil sampled at shallow depth (<0.1 m); and in cropping or mixed livestock and cropping systems. The inclusion and exclusion criteria for studies were directly related to their relevance to the eligible management activities, and calculation of SOC stocks in the 2021 Estimating Soil Organic Carbon Sequestration Using Measurement and Models method (Federal Register of Legislation 2021). The method is specific to Australia, and ACCUs can be issued using measured and modelled changes in soil carbon stocks. Eligible management activities must meet the ‘newness requirement’, hence the comparisons between the eligible management activities selected for this review, rather than articles that explore other eligible management activities that appeared in the search strings, for example switching from cropping to pasture or including a legume in a cropping rotation.

Following the manual review, 10 articles on grazing management and four articles on legumes (Table 1), one of which had already appeared in the rotational grazing search query, were selected for closer examination. In addition, if articles were identified outside the search terms, then their eligibility was examined using the stated inclusion/exclusion criteria.

Analysis

The 14 articles that met the inclusion criteria were examined to extract data on the following parameters. The information from each article included: SOC accumulation (tonnes of carbon per hectare per year); soil sampling depth (m); measured or modelled; experimental design and sampling regime; length of time over which the study was conducted (years) and eligible activity applied; type of management activity (grazing management or legumes); type of grazing or type of legume; location, climate, and average annual rainfall (AAR); and soil type. In addition, for each article the strength of evidence was assessed based on Reinhart et al. (2022) who developed three standards: before–after randomised controlled trial (Gold standard); randomised controlled trial (Silver standard); and control-impact trial (Bronze standard). Some articles conformed to higher levels, but also may have lacked other qualities, such as calculation of SOC stocks based on equivalent soil mass, which placed them in the lower category.

Distribution of the literature

Experimental sites and duration of study

Each climate zone (Fig. 1) was represented on at least one experimental site of the dataset studied, with some articles having multiple experimental sites across up to three different climate zones (Fig. 2).

Fig. 2.

Climate zones (according to Koppen Climate classes) in which the article’s study site(s) is/are located. Coloured cells indicate presence in a climate zone and white cells indicate there is no study in that climate zone (source: Table 1).

The majority of the articles (72%) were replicated paired sites, but with no baseline measure of SOC stocks. Therefore, SOC stocks were a relative comparison between treatments at one point in time, and were not a reflection of changes in SOC stocks before imposition of management activity. These comparative observational studies did not have a baseline measurement or a subsequent measurement of SOC at a second point in time. This presents significant limitations to showing whether the management activities result in accumulation of soil carbon stocks because it assumes each sampling area had a similar starting point (Table 1).

For the remaining 28% of studies, which were also replicated, there was a baseline of SOC stocks, with at least one subsequent set of measurements. The length of time the studies were active for was highly variable, ranging from the shortest, at 5 years, and the longest, at 40 years (Table 1). All these studies were on a research station where conditions can be monitored consistently, but still consisted of one sampling period as using a space for time substitution experimental design. The length of study is important due to the other factors that influence SOC stocks, particularly drought, so that climate variability and atypical years can be accounted for in a longer period of study.

The majority of study sites were situated on commercial farms (72% of articles), while the remaining study sites were situated on research stations. Most of the study sites that were on research stations were those that had measurements over at least two points in time to allow for temporal comparison. There are pros and cons of commercial farms over a research station that relate to greater research relevance and credibility with on-farm research. Overall, the proportion of studies that were classified as Gold standard in terms of experimental design and explanatory power were low (14%), followed by 36% of studies at Silver standard, and 50% of studies at Bronze standard (Table 1). Reinhart et al. (2022) in their global review on grazing management and SOC stocks reported a lower proportion of studies at Gold Standard (4%), and 68% of studies at Silver standard, and the remainder of studies at Bronze standard. Cowie et al. (2013) explained some of the experimental problems that arise from comparative studies, including the limitations of the survey approach that was used as it had limited capacity to detect differences in SOC stocks that were attributable to specific management activities because of the wide variation in management histories and strong influence of environmental factors.

Soil measurement depth

Depths to which soil measurements were taken were from 0.15 to 1 m (Fig. 3), typically divided into smaller intervals closer to the surface soil, which were 0–0.1 m, 0.1–0.2 m and 0.2–0.3 m. Approximately half of the papers measure SOC stocks to 0.3 m (n = 7). There was a lack of measurement of SOC at depths greater than 0.3 m (n = 5), despite anecdotal evidence that influencing SOC at depth presents an opportunity for increased ACCUs, which can be beyond 0.3–0.5 m into the subsoil. However, soil carbon sequestration beyond 0.3 m is not reported to the United Nations as part of the national GHG inventory, even though soil carbon sequestration could be occurring at deeper depths (Simmons et al. 2023). In addition, only 50% of the studies examined SOC stocks between the grazing treatments or baseline measurement using an equivalent soil mass (Fig. 3).

Fig. 3.

Maximum sampling depth of soil cores (n = 14 articles), G - Grazing management, L - Legume, ESM - Equivalent Soil Mass (source: Table 1).

Effect of grazing management and legumes on SOC stock

The articles included in this review were from a peak period of publicly funded research and were published from 2010 to 2020. Since the end of the project funding, there has not been any follow-up research on those same experimental study sites. The scale of the research often included other explanatory variables of soil carbon sequestration, such as soil type, climate, and land use type, which confounded the effect of the management practice (Amin et al. 2020). Most articles that appeared in both search strings were ultimately excluded, due to not conforming to all inclusion criteria, and others examining the conversion of crop to pasture, which reflects a potential research bias. Another recent review on grazing management for soil carbon in Australia identified 30 published articles, but 17 of these articles were excluded from this literature critique (McDonald et al. 2023). There were 12 articles common to both reviews.

The variation in experimental design, and in the measurement of SOC stocks under the management activities, made it difficult to compare findings between studies. The variation in how SOC stocks were reported also made it difficult to compare the various studies. The reported impact of the grazing treatments were tabulated and were generally not significant (Table 1). One grazing management study suggested that 1.46 Mg C ha⁻¹ year⁻¹ to a depth of 0.3 m over 5 years could accrue, but this was also based on an assumption that SOC levels were similar at the start for all treatments (Orgill et al. 2018) (Table 1). Less than a third of the articles (28%) found that SOC stocks were positively influenced by either grazing management (n = 1) or legume inclusion (n = 3); and for the legume studies, there was baseline measurement of SOC stocks and the work was conducted on a research station under long-term trials (Table 1). The lack of significance in findings derived from observational studies means a level of caution needs to be applied to eligible activities for soil carbon projects, when the only significant change is recorded on a research station. Often the geographical spread of sites on commercial farms studied led to a highly variable dataset where only a relative comparison is useful. Dean et al. (2012) noted that substituting space for time can be inferior to long-term experimental sites, since the former approach needs to take into account spatial heterogeneity, appropriate site-pairing, and interconnections for SOC in time.

Additionally, apart from one article, the results were not also presented in SOC%, which would aid in communication of the research findings to landowners, who are more likely to measure SOC in percentage than stocks (Ingram et al. 2016). The separate sources of funding for the identified studies meant they have operated largely independently of each other, without coordination of influential factors, climate, soil, land use practice, or grass type. In addition, competitive funding has restricted the type of studies that have taken place and the length of study and ability to undertake long-term work, which measures temporal effects over time (Lobry de Bruyn et al. 2022). Soil bulk density used to measure SOC stocks would also change over time with grazing treatment, and hence sampling depth would need to be adjusted or soil mass kept constant between treatments for equivalent soil mass calculations, which only occurred in 50% of studies. In a global review of grazing and its effects of SOC, most studies (n = 71) had reported SOC stocks for a specific depth interval (0–30 cm), and grazing is known to alter bulk density (Reinhart et al. 2022). Also, this meta-analysis reported that only three of the 71 studies (<4%) had tested effects of grazing treatments on changes in SOC variables (Reinhart et al. 2022). Furthermore, the ambiguity of the impact of eligible activities on SOC stocks means that for carbon project developers it could result in a lack of legitimacy or uncertainty in soil carbon project deliverables. For example, when determining the profitability of a soil carbon project by predicting sequestration rate over a project crediting period (25 years for soil carbon projects registered under the ACCU Scheme), a project developer may observe a dataset that indicates carbon sequestration at a rate of 0.5 t CO₂e ha⁻¹ year⁻¹ but it is difficult to interpret in that dataset how long that sequestration rate could be maintained for, and the developer cannot assume that this sequestration rate is possible for the entirety of the crediting period.

Interaction between the two management activities

The interaction between the inclusion of legumes and different grazing management activities and their combined impact on SOC stocks was not addressed in any of the articles. The best potential for SOC sequestration is likely a combination of these management activities, rather than one or the other. However, this is only relevant to a point, that is soils can only sequester up to as much carbon as they can hold, which is dependent on inherent soil characteristics such as clay percentage, referred to as the ‘soil carbon saturation concept’ (Hassink 1997; Six et al. 2002). Orgill et al. (2018) was the only article in which the saturation capacity was mentioned at the end of the discussion. A recent review questioned the level of carbon saturation, and that soils under pasture are probably already close to the ceiling of SOC saturation with little potential for more soil carbon sequestration (Moinet et al. 2023). This presents a limitation of the data available, because if some of the soils were close to their saturation point while others were not, it is less likely a meaningful comparison can be made.

Exploration of other facets of SOC accumulation

Many of the articles were also investigating total nitrogen and presenting the different pools of carbon (particulate, humic, and recalcitrant pools) under different scenarios. Co-benefits such as increased livestock carrying capacity, above and below ground biomass, plant available water (except for Radrizzani et al. 2011), drought resilience, and on-farm profitability due to increased soil carbon stocks were not explored.

Legume inclusion in pastoral systems has been found to have several co-benefits due to their biological nitrogen fixation capacities, as well as additional ecological services such as improved soil structure, erosion protection, and increased biodiversity. However, there has been a decline in the inclusion of leguminous plants in farming systems with readily available synthetic nitrogen fertiliser (Jensen and Hauggaard-Nielsen 2003). Legume-induced soil acidification has been observed (Haynes 1983), and a long-term fertility trial in the USA found that ammoniacal nitrogen fertiliser was accompanied by increased soil acidity, and decreased cation exchange capacity, exchangeable calcium, and magnesium ions (Barak et al. 1997). It has also been argued that legume-derived nitrogen has similarly negative environmental impacts (Crews and Peoples 2004). Further research is required to establish the net GHG emissions of legume inclusion versus the use of synthetic nitrogen fertiliser, as well as the other environmental impacts of each of these practices. It could be argued that sourcing nitrogen for plant growth from either legumes or fertiliser is practical for Australian pastoral systems, noting that in some systems, particularly intensively farmed dairies, legume establishment and persistence is rare.

There are different ways to manage grazing in farming systems, and there is a broad spectrum of systems and practices that can be described by the term ‘rotational grazing’ with more common terms including deferred grazing, rotational grazing (e.g. intensive, simple, and slow), time-controlled grazing, cell grazing, and set stocking. In a recent survey of Australian farmers, 62% of those surveyed were undertaking a form of rotational grazing, and most for more than 3 years (Coelli 2021). Under grazing management, finding the optimal grazing pressure can also be included. Nie et al. (2015) found that deferred grazing treatments resulted in a 200% perennial grass germinable seed pool increase and a 50% germinable seed pool increase for annual grasses. Sanjari et al. (2008) found that time-controlled grazing compared with continuous grazing reduced nitrate and extractable phosphorus concentrates due to increased grass growth under the time-controlled grazed areas, which decreased contamination potential for downstream water bodies. One of the limitations of comparing the identified articles is that different grazing management activities were imposed in the examined studies or it was an observational study collecting data from operating commercial farms, and therefore comparison is difficult as it is not a fully factorial, replicated, and controlled experimental design with baseline measures of SOC stocks (Table 1).

Grazing management and legume inclusion in Australian pastoral systems are management activities that can increase farm profitability, provided they are implemented strategically. The bio-economic analysis of grazing regimes has been studied in the south-west of Western Australia and it was demonstrated that rotational grazing and continuous grazing were both profitable but the more robust regime was rotational grazing (Sounness 2004). In both systems, perennial pastures were also superior to annual pastures. For example, the capital and labour costs of a rotational or cell grazing system must be considered prior to implementation of such a strategy, as this system requires relatively small paddock size and frequent movement of stock. Legume inclusion is likely to result in decreased fertiliser use; however, establishing legumes in some systems may be challenging and seeding multiple times may be required. Another significant challenge to legume establishment and persistence, particularly lucerne (Dear et al. 2003), is the prevalence of acid soils in Australia. This is especially prevalent in the more intensively grazed systems, of which it is thought approximately 21% have a high risk of acidification. Acid soils increase the risk of topsoil loss through erosion because they have decreased ability to support productive vegetation cover, thereby reducing potential inputs to SOC stocks (Barson et al. 2011).

Additionally, there is a social dimension to adoption of new management activities. Examination of long-term practitioners of rotational grazing in the New England Tablelands of New South Wales, Australia, showed that soil carbon management was focused more on the co-benefits of managing soil carbon for plant production, soil health, and overall farm sustainability than the opportunity to earn ACCUs (Amin et al. 2023a). While the two management activities assessed in this paper are already at reasonable levels of adoption Australia-wide, especially grazing management, ‘carbon farming’ is reported at 12% of Australian farmers in a recent survey of farm practices (Coelli 2021), with the adoption of a carbon soil project closer to 0.4% of Australian farmers (Fitch et al. 2022). This low level of adoption in a soil carbon project further highlights the need for research on the co-benefits, including on-farm profitability in a commercial farming system and soil stewardship if adoption is to scale-up (Amin et al. 2023b). Pannell and Crawford (2022) argue that paying farmers for carbon that they were going to sequester anyway, despite the policy measures around additionality that intend to prevent this from happening, makes no ‘new’ contribution to mitigating climate change.